1commit 622152ee42a310876f10602601206954b8d0613e
2Author: Matt Turner <mattst88@gmail.com>
3Date:   Fri Apr 12 12:35:55 2024 -0400
4
5    libxcb 1.17.0
6
7commit 4574ab269b082bf412e1b6fb5f09daf8134acd0b
8Author: Matt Turner <mattst88@gmail.com>
9Date:   Fri Apr 12 12:35:32 2024 -0400
10
11    configure.ac: Require xcb-proto >= 1.17.0
12    
13    Signed-off-by: Matt Turner <mattst88@gmail.com>
14
15commit b78d304531bfe2359baf40db06099d9158d398d4
16Author: Uli Schlachter <psychon@znc.in>
17Date:   Sun Sep 17 12:43:43 2023 +0200
18
19    Always write C code in UTF-8
20    
21    Some people apparently use non-utf8 locales and this caused errors when
22    xcb-proto started using some "fancy" quote marks. Fix this by always
23    using utf8 encoding.
24    
25    Fixes: https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/72
26    Signed-off-by: Uli Schlachter <psychon@znc.in>
27
28commit 389f22d1cb53e47337e96d9f78a3ef0001e8e588
29Author: Erik Kurzinger <ekurzinger@nvidia.com>
30Date:   Wed Mar 20 07:06:01 2024 -0700
31
32    Add xcb-dri3 dependency to xcb-present.pc.in
33    
34    Present version 1.4 introduces a dependency on DRI3 for the DRI3Syncobj
35    protocol type.
36    
37    Signed-off-by: Erik Kurzinger <ekurzinger@nvidia.com>
38
39commit 86a478032ba93f30adbc0ce96eecd3420fdf7ed1
40Author: Alan Coopersmith <alan.coopersmith@oracle.com>
41Date:   Sat Mar 2 12:01:54 2024 -0800
42
43    xcb_popcount: Use __builtin_popcount if compiler supports it
44    
45    If the compiler knows of a better implementation for counting the number
46    of bits set in a word for the target CPU, let it use that, instead of the
47    classic algorithm optimized for PDP-6.
48    
49    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
50
51commit c268499c305317d2e9a67cc590c9147e11438fc7
52Author: Alan Coopersmith <alan.coopersmith@oracle.com>
53Date:   Sat Mar 2 11:47:47 2024 -0800
54
55    tests: fix -Werror=discarded-qualifiers errors in check_public.c
56    
57    check_public.c: In function ‘parse_display_pass’:
58    check_public.c:32:32: error: passing argument 1 of ‘putenv’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
59       32 |                         putenv("DISPLAY=");
60          |                                ^~~~~~~~~~
61    In file included from check_public.c:4:
62    /usr/include/stdlib.h:148:19: note: expected ‘char *’ but argument is of type ‘const char *’
63      148 | extern int putenv(char *);
64          |                   ^~~~~~
65    check_public.c:57:16: error: passing argument 1 of ‘putenv’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
66       57 |         putenv("DISPLAY=");
67          |                ^~~~~~~~~~
68    /usr/include/stdlib.h:148:19: note: expected ‘char *’ but argument is of type ‘const char *’
69      148 | extern int putenv(char *);
70          |                   ^~~~~~
71    check_public.c: In function ‘parse_display_fail’:
72    check_public.c:73:32: error: passing argument 1 of ‘putenv’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
73       73 |                         putenv("DISPLAY=");
74          |                                ^~~~~~~~~~
75    /usr/include/stdlib.h:148:19: note: expected ‘char *’ but argument is of type ‘const char *’
76      148 | extern int putenv(char *);
77          |                   ^~~~~~
78    check_public.c:99:16: error: passing argument 1 of ‘putenv’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
79       99 |         putenv("DISPLAY=");
80          |                ^~~~~~~~~~
81    /usr/include/stdlib.h:148:19: note: expected ‘char *’ but argument is of type ‘const char *’
82      148 | extern int putenv(char *);
83          |                   ^~~~~~
84    check_public.c: In function ‘public_suite’:
85    check_public.c:244:16: error: passing argument 1 of ‘putenv’ discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
86      244 |         putenv("DISPLAY=");
87          |                ^~~~~~~~~~
88    /usr/include/stdlib.h:148:19: note: expected ‘char *’ but argument is of type ‘const char *’
89      148 | extern int putenv(char *);
90          |                   ^~~~~~
91    cc1: all warnings being treated as errors
92    
93    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
94
95commit c671b9b30aecada4e43cb48e0dee46d19fb5cb9c
96Author: Alan Coopersmith <alan.coopersmith@oracle.com>
97Date:   Sat Mar 2 11:13:37 2024 -0800
98
99    libxcb 1.16.1
100    
101    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
102
103commit 3c946010c8521497b0fba2c8bc9bde184622345a
104Author: Fergus Dall <sidereal@google.com>
105Date:   Tue Nov 7 17:42:42 2023 +1100
106
107    Enable large file support
108    
109    Signed-off-by: Fergus Dall <sidereal@google.com>
110
111commit 02a7bbed391859c79864b9aacf040d84f103d38a
112Author: Peter Williams <peter@newton.cx>
113Date:   Thu Aug 24 21:50:29 2023 -0400
114
115    Fix compilation on Windows
116    
117    Merge request !30 broke compilation on Windows by introducing unbalanced
118    curly braces. This change restores the previous behavior on Windows:
119    only TCP is supported.
120
121commit 18c3f055e7c02db6f15ae56f6995639635d1b055
122Author: Alan Coopersmith <alan.coopersmith@oracle.com>
123Date:   Sun Aug 20 11:34:31 2023 -0700
124
125    gitlab CI: use latest xcbproto from git to build against
126    
127    Closes issue #70
128    
129    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
130
131commit 453115f7eeb694de9f41ea842a29bbb31b90c8dd
132Author: Thomas Klausner <wiz@gatalith.at>
133Date:   Sat Aug 19 19:34:51 2023 +0200
134
135    configure.ac: drop `pthread-stubs` dependency on NetBSD
136
137commit f11691a09893d2329948521e32a18c13b0b03d01
138Author: Alan Coopersmith <alan.coopersmith@oracle.com>
139Date:   Sun Aug 20 10:22:19 2023 -0700
140
141    gitlab CI: regenerate container used for builds
142    
143    This updates the included xcb-proto to the newly required 1.16.
144    As explained in issue #70, this is a quick workaround, not the
145    long-term fix for this problem.
146    
147    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
148
149commit cc4b93c9cd93bad15b7106747b0213e4b9c53a1c
150Author: Matt Turner <mattst88@gmail.com>
151Date:   Wed Aug 16 16:09:50 2023 -0400
152
153    libxcb 1.16
154    
155    Signed-off-by: Matt Turner <mattst88@gmail.com>
156
157commit 1519334652837cb7b2a8f8117628ea3cc744f7cb
158Author: Matt Turner <mattst88@gmail.com>
159Date:   Wed Aug 16 16:08:46 2023 -0400
160
161    configure.ac: Require xcb-proto >= 1.16.0
162    
163    Signed-off-by: Matt Turner <mattst88@gmail.com>
164
165commit 038636786ad1914f3daf3503ae9611f40dffbb8f
166Author: Ilya Pominov <ipominov@astralinux.ru>
167Date:   Sun Apr 2 11:35:16 2023 +0300
168
169    c_client: Fix crash in xcb_randr_set_monitor
170    
171    Take into account c_need_sizeof when evaluating
172    xcb_protocol_request_t.count
173    
174    Incorrect xcb_protocol_request_t.count causes a segmentation fault when
175    calling functions:
176    - xcb_randr_set_monitor{_checked}
177    - xcb_input_change_feedback_control{_checked}
178    - xcb_input_change_device_control{_unchecked}
179    
180    Steps to reproduce:
181    Call xcb_randr_set_monitor() with valid arguments
182    
183    OBSERVED RESULT
184    Segmentation fault
185    ```
186    Process terminating with default action of signal 11 (SIGSEGV)
187      Access not within mapped region at address 0x780
188        at 0x4852925: memmove (in /usr/libexec/valgrind/vgpreload_memcheck-
189    amd64-linux.so)
190        by 0x486967C: memcpy (string_fortified.h:29)
191        by 0x486967C: send_request (xcb_out.c:59)
192        by 0x486967C: send_request (xcb_out.c:46)
193        by 0x486967C: xcb_send_request_with_fds64 (xcb_out.c:338)
194        by 0x48699FC: xcb_send_request (xcb_out.c:359)
195        by 0x4891F11: xcb_randr_set_monitor_checked (randr.c:5350)
196    ```
197    
198    EXPECTED RESULT
199    Function returns cookie
200    
201    Amend: 77b594f9583ea0247ff27130316d8e045da7f921
202    
203    Signed-off-by: Ilya Pominov <ipominov@astralinux.ru>
204
205commit 8935793f1f3751a6aa9d78955c7d6236177986de
206Author: Demi Marie Obenour <demiobenour@gmail.com>
207Date:   Sat Mar 25 15:27:33 2023 -0400
208
209    Add tests for unix socket parsing
210    
211    Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
212
213commit 095255531b90f0b442e6ca41fb3752a058562d07
214Author: Demi Marie Obenour <demiobenour@gmail.com>
215Date:   Fri Aug 12 22:02:02 2022 -0400
216
217    DISPLAY starting with / or unix: is always a socket path
218    
219    If DISPLAY starts with / or unix:, do not check for anything but a full
220    filesystem socket path.  In particular, abstract AF_UNIX sockets and TCP
221    sockets will not be used in this case.   Also be stricter about parsing
222    the screen part of /path.screen displays, and bail out after all stat()
223    errors other than ENOENT.
224    
225    Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
226
227commit ccdef1a8a5b6956a3fb7115d634db213433dca75
228Author: Demi Marie Obenour <demiobenour@gmail.com>
229Date:   Sun Jul 3 17:57:11 2022 -0400
230
231    Allow full paths to sockets on non-macOS
232    
233    When combined with xorg/lib/libxtrans!7, this allows CVE-2020-25697 to
234    be mitigated by placing the AF_UNIX socket in a secure directory on the
235    filesystem.
236    
237    This enables HAVE_LAUNCHD unconditionally and deletes the configure
238    switch.
239    
240    Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
241
242commit 18e109d755c5ce18157fdabb6de8ee6845b348ff
243Author: Matthieu Herrb <matthieu@herrb.eu>
244Date:   Tue Feb 28 14:19:21 2023 +0100
245
246    Fix xcb_connect() call with a custom xcb_auth_info_t
247    
248    If the function implementing xcb_connect is called directly with a
249    custom xcb_auth_info_t then checking that the screen in $DISPLAY
250    is valid is skipped.
251    
252    Reported by chohag AT jtan DOT com
253    
254    Signed-off-by: Matthieu Herrb <matthieu@herrb.eu>
255
256commit fd04ab24a5e99d53874789439d3ffb0eb82574f7
257Author: Jan Beich <jbeich@FreeBSD.org>
258Date:   Fri Feb 10 14:16:20 2023 +0000
259
260    configure.ac: drop `pthread-stubs` dependency on DragonFly and FreeBSD
261    
262    All `pthread_*` symbols used by libxcb have stubs in libc. So, stop
263    linking against libpthread.
264
265commit 4d1a578dd5348909ade2a853d806272326d228d7
266Author: Alan Coopersmith <alan.coopersmith@oracle.com>
267Date:   Thu Feb 9 14:56:38 2023 -0800
268
269    Raise minimum required version of Python from 2.6 to 3.0
270    
271    Trying to build with Python 2.7.14 fails with:
272      File "./c_client.py", line 2270
273        key = (*self.name[:-1], field.enum)
274               ^
275    SyntaxError: invalid syntax
276    
277    Fixes: 33f3dbe ("Fix handling of documented enum parameters")
278    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
279
280commit e2a3e80eb837eaaa834a01a2741d4c249c984442
281Author: Alan Coopersmith <alan.coopersmith@oracle.com>
282Date:   Mon Feb 6 15:48:49 2023 -0800
283
284    configure.ac: Remove obsolete AC_HEADER_STDC
285    
286    Nothing checks for the STDC_HEADERS flag this set, and all supported
287    systems have C89 compatible headers now.
288    
289    Clears autoconf warnings of:
290    configure.ac:93: warning: The macro `AC_HEADER_STDC' is obsolete.
291    configure.ac:93: You should run autoupdate.
292    ./lib/autoconf/headers.m4:704: AC_HEADER_STDC is expanded from...
293    configure.ac:93: the top level
294    
295    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
296
297commit 3333d5bde8baaca8981e806ef86b74f4adb77ba7
298Author: Daniel G <giri@nwrk.biz>
299Date:   Mon Jan 9 16:21:35 2023 +0000
300
301    Fix indentation.
302
303commit 973b510e954f6989ed773f8b357ca50fbc99bced
304Author: Daniel G <giri@nwrk.biz>
305Date:   Sat Jan 7 19:35:47 2023 +0000
306
307    Fix windows build.
308
309commit cb8c70f5a65b4bd68b449dcaa637c3c4753e2f81
310Author: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
311Date:   Sat Nov 26 14:55:07 2022 -0800
312
313    xcb_conn: Add a check for NULL to silence a UBSan runtime error
314    
315    xcb_conn.c:314:60: runtime error: applying zero offset to null pointer
316    
317    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
318
319commit 33f3dbe3699a92e8ca18f470adac456e0b935e75
320Author: Demi Marie Obenour <demiobenour@gmail.com>
321Date:   Sun Oct 16 01:50:53 2022 -0400
322
323    Fix handling of documented enum parameters
324    
325    Previously this would crash the code generator.
326    
327    Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
328
329commit c9513aac2d78808f63b83d71072c2c483b20162d
330Author: Demi Marie Obenour <demi@invisiblethingslab.com>
331Date:   Sat Aug 28 12:25:05 2021 -0400
332
333    Fix a compiler warning
334    
335    The warning is harmless but annoying.
336    
337    Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
338
339commit 06e1ef43bbb8c84fc243fa6a53c81147a3a24273
340Author: Alan Coopersmith <alan.coopersmith@oracle.com>
341Date:   Tue Oct 4 11:30:09 2022 -0700
342
343    configure.ac: drop dependency `pthread-stubs` on Solaris
344    
345    On Solaris 10 and later, the pthread functions are directly in libc,
346    and libpthread only has metadata to redirect calls from it to the
347    libc functions.
348    
349    On Solaris 9 and older (no longer supported), libc contained its own
350    thread stubs that libpthread then overrode.
351    
352    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
353
354commit d53ad60d774176c5b057e2ed4d03a4658ee13690
355Author: PaperChalice <lgamma@163.com>
356Date:   Sat Aug 13 13:11:10 2022 +0800
357
358    configure.ac: drop dependency `pthread-stubs` on macOS
359    
360    `libpthread.dylib` on macOS is an alias of `libSystem.B.dylib`,
361    every program should link against `libSystem.B.dylib`.
362
363commit e2ee5aabe9f731b7231b0d4a5a21367d1aced6ed
364Author: Uli Schlachter <psychon@znc.in>
365Date:   Wed Sep 14 14:53:35 2022 +0200
366
367    Improve tutorial example
368    
369    This tutorial example only handles XCB_EXPOSURE and XCB_KEY_RELEASE
370    events and ignores everything else. Thus, there is no point in asking
371    for more kinds of events.
372    
373    A while ago, I ported this tutorial to x11rb [1]. Recently, I received a
374    pull request [2] removing these unnecessary event masks. This commit is
375    thus only partially by me and the 'issue' was originally found by the
376    author of [2].
377    
378    [1]: https://github.com/psychon/x11rb/blob/master/x11rb/examples/tutorial.rs
379    [2]: https://github.com/psychon/x11rb/pull/754
380    
381    Signed-off-by: Uli Schlachter <psychon@znc.in>
382
383commit 9dcb08170898d3bd5ddb3bc26c514a6bfccf3fae
384Author: Mihail Konev <k.mvc@ya.ru>
385Date:   Thu Jan 26 13:52:49 2017 +1000
386
387    autogen: add default patch prefix
388    
389    Signed-off-by: Mihail Konev <k.mvc@ya.ru>
390    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
391
392commit 7d798d3ccbe537f2f102665447ec3e15a8cdb1a1
393Author: Emil Velikov <emil.l.velikov@gmail.com>
394Date:   Mon Mar 9 12:00:52 2015 +0000
395
396    autogen.sh: use quoted string variables
397    
398    Place quotes around the $srcdir, $ORIGDIR and $0 variables to prevent
399    fall-outs, when they contain space.
400    
401    Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
402    Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
403    Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
404    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
405
406commit 7071b4a13c1e4df13b138748feebadc17fe2522b
407Author: Peter Hutterer <peter.hutterer@who-t.net>
408Date:   Tue Jan 24 10:32:07 2017 +1000
409
410    autogen.sh: use exec instead of waiting for configure to finish
411    
412    Syncs the invocation of configure with the one from the server.
413    
414    Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
415    Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
416    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
417
418commit ee60239b7308f98a7a32e2a567681e00c6b652f5
419Author: Alan Coopersmith <alan.coopersmith@oracle.com>
420Date:   Mon Jul 18 17:55:43 2022 -0700
421
422    gitlab CI: add a basic build test
423    
424    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
425
426commit 3123dfe723648c5abeab429e7cb4bd194ceeb8f6
427Author: Mike Sharov <msharov@users.sourceforge.net>
428Date:   Wed Jun 22 10:56:37 2022 -0400
429
430    Require xcb-proto >= 1.15.1, which has Dbe
431    
432    Signed-off-by: Mike Sharov <msharov@users.sourceforge.net>
433
434commit 816407655f75cf8ae505432cb743c316e009679d
435Author: Mike Sharov <msharov@users.sourceforge.net>
436Date:   Thu Jun 2 13:27:48 2022 -0400
437
438    Build DOUBLE-BUFFER extension.
439    
440    Signed-off-by: Mike Sharov <msharov@users.sourceforge.net>
441
442commit c2c4a2cd1947e559718acdba19ef6e7db731dbeb
443Author: Matt Turner <mattst88@gmail.com>
444Date:   Tue May 3 15:09:54 2022 -0700
445
446    libxcb 1.15
447    
448    Signed-off-by: Matt Turner <mattst88@gmail.com>
449
450commit ddafdba11f6919e6fcf977c09c78b06f94de47aa
451Author: Hodong <hodong@yozmos.com>
452Date:   Sat Jan 15 02:32:04 2022 +0900
453
454    Fix a memory leak
455    
456    Signed-off-by: Hodong <hodong@yozmos.com>
457
458commit 43fbf03e549bf6da8d1d8522e0ceddc4d49c37c6
459Author: Demi Marie Obenour <demiobenour@gmail.com>
460Date:   Thu Sep 23 16:46:32 2021 -0400
461
462    Fix integer overflows in xcb_in.c
463    
464    This fixes an integer overflow security vulnerability in xcb_in.c, which
465    may allow for memory corruption.
466
467commit 233d7b7f1f03ef18bf3955eb1f20421e745d22f0
468Author: Thomas Anderson <thomasanderson@google.com>
469Date:   Wed Dec 2 00:25:42 2020 +0000
470
471    Fix hang in xcb_request_check()
472    
473    This fixes https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/53
474    
475    The issue was that libxcb expected to get a reply based on the request_expected
476    variable, but a reply would never arrive because the request was never actually
477    written.  To resolve this, a separate request_expected_written variable is
478    added.
479
480commit dc2811874729ee83fa2aef110f60808c450f9a5a
481Author: Ran Benita <ran@unusedvar.com>
482Date:   Tue Nov 17 23:18:53 2020 +0200
483
484    Avoid request counter truncation in replies map after 2**32 requests
485    
486    The c->in request counters are uint64_t, and can realistically go over
487    2**32 over a lifetime of a client. The c->in->replies map however uses
488    unsigned int keys and the passed request numbers are silently truncated.
489    
490    I haven't analyzed in depth what happens what it wraps around but it's
491    probably nothing good.
492    
493    The only user of the xcb_list.c map code is c->in->replies, so just
494    change it to use uint64_t keys.
495    
496    Reviewed-by: Uli Schlachter <psychon@znc.in>
497    Signed-off-by: Ran Benita <ran@unusedvar.com>
498
499commit 26396bf156cfa00ecd655ec6a5f2f46626d674c0
500Author: Julien Cristau <jcristau@debian.org>
501Date:   Tue Feb 2 12:15:10 2021 +0100
502
503    Add newline when printing auth/connection failure string to stderr
504    
505    The reason strings returned by the server don't all include a newline,
506    so make sure we add one to avoid confusing clients.  Xlib used to do
507    this before it delegated that work to libxcb.
508    
509    Fixes #34
510    
511    Signed-off-by: Julien Cristau <jcristau@debian.org>
512
513commit a503167f751ba77e6819df568b7f5042d2baa9c9
514Author: Uli Schlachter <psychon@znc.in>
515Date:   Sat May 22 21:43:22 2021 +0200
516
517    Improve/fix docs for reply fds functions
518    
519    Fixes: https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/56
520    Signed-off-by: Uli Schlachter <psychon@znc.in>
521
522commit 3c76c0579ffa521af41f78ee937b1337581da985
523Author: Povilas Kanapickas <povilas@radix.lt>
524Date:   Fri Jul 30 22:58:47 2021 +0300
525
526    c_client.py: Implement handling of <length> element
527    
528    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
529
530commit bdc3f21a5205293852bbaa173f43389cbf66f52b
531Author: Povilas Kanapickas <povilas@radix.lt>
532Date:   Fri Jul 30 22:58:46 2021 +0300
533
534    c_client: Extract _c_get_field_mapping_for_expr()
535
536commit 068af21cb376eb824fa1dee29a6539feadb93587
537Author: Povilas Kanapickas <povilas@radix.lt>
538Date:   Fri Jul 30 22:58:45 2021 +0300
539
540    c_client.py: Use get_expr_field_names directly to resolve list fields
541    
542    Using get_expr_fields() is only needed in case we are doing things that
543    can span multiple types easily, e.g. when deciding what data to pass via
544    function parameters and so on.
545    
546    In _c_serialize_helper_list_field() we are building function body, so
547    acquiring field names via get_expr_field_names() is enough.
548    
549    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
550
551commit 4d678b162bf8a3b10e5bdf76df2be63d33c23381
552Author: Povilas Kanapickas <povilas@radix.lt>
553Date:   Fri Jul 30 22:58:44 2021 +0300
554
555    c_client.py: Extract get_expr_field_names()
556    
557    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
558
559commit 21414e7c447f18224c577ed5e32bd5d6e45c44f9
560Author: Peter Harris <pharris@opentext.com>
561Date:   Mon Feb 1 19:45:28 2021 -0500
562
563    Fix writev emulation on Windows
564    
565    There are at least two bugs in the previous implementation:
566    
567    - If an early iovec is partially written, there can be a gap of missing
568      data (as a later iovec will be started before the early iovec is
569      completed).
570    - If a late iovec returns WSAEWOULDBLOCK, *vector and *count are not
571      updated, leading to a re-send of the entire request.
572    
573    Move the *vector update into the send() loop to update piecemeal as
574    individual iovecs are sent.
575    
576    Example program that demonstrates the issue (this program should run
577    forever after these bugs have been fixed):
578    
579    #include <stdio.h>
580    #include <stdlib.h>
581    #include "xcb.h"
582    
583    // Non-cryptographic random number generator from http://burtleburtle.net/bob/rand/smallprng.html
584    // because Microsoft's random number generators either have a too small RAND_MAX or are too slow
585    typedef struct ranctx { uint32_t a; uint32_t b; uint32_t c; uint32_t d; } ranctx;
586    
587    static uint32_t ranval(ranctx *x);
588    static void raninit(ranctx *x, uint32_t seed);
589    
590    
591    #define MAX_PROP_LEN (128 * 1024)
592    
593    int main(int argc, char *argv[]) {
594        uint32_t seed = 0x12345678;
595        if (argc > 1) {
596            seed = strtoul(argv[1], NULL, 0);
597        }
598        ranctx ran;
599        raninit(&ran, seed);
600    
601        xcb_connection_t *c = xcb_connect(NULL, NULL);
602        if (!c || xcb_connection_has_error(c)) {
603            printf("Cannot connect to $DISPLAY\n");
604            return 1;
605        }
606        const xcb_setup_t *setup = xcb_get_setup(c);
607        char *buf = malloc(MAX_PROP_LEN + 8); // plus a bit of slack so we can run random values off the end
608        if (!buf) {
609            printf("oom\n");
610            return 1;
611        }
612        for (uint32_t i=0; i < (MAX_PROP_LEN + 3) / 4; i++) {
613            ((uint32_t *)buf)[i] = ranval(&ran);
614        }
615    
616        xcb_window_t win = xcb_generate_id(c);
617        xcb_create_window(c, 0, win, xcb_setup_roots_iterator(setup).data[0].root, 0, 0, 1, 1, 0,
618                XCB_WINDOW_CLASS_INPUT_ONLY, 0, 0, NULL);
619        printf("Created window 0x%X\n", win);
620    
621        for (;;) {
622            xcb_flush(c);
623            xcb_generic_event_t *ev = xcb_poll_for_event(c);
624            if (ev) {
625                if (ev->response_type == 0) {
626                    xcb_generic_error_t *err = (xcb_generic_error_t *)ev;
627                    printf("Unexpected X Error %d\n", err->error_code);
628                    printf("   Sequence %d\n", err->sequence);
629                    printf("   Resource ID 0x%X\n", err->resource_id);
630                    printf("   Opcode: %d.%d\n", err->major_code, err->minor_code);
631                    return 1;
632                }
633                printf("Unexpected X Event %d\n", ev->response_type);
634                return 1;
635            }
636    
637            uint32_t siz = ranval(&ran) % MAX_PROP_LEN + 1;
638            xcb_change_property(c, XCB_PROP_MODE_REPLACE, win, XCB_ATOM_STRING, XCB_ATOM_STRING, 8, siz, buf);
639        }
640    
641        return 0;
642    }
643    
644    
645    #define rot(x,k) (((x)<<(k))|((x)>>(32-(k))))
646    static uint32_t ranval(ranctx *x) {
647        uint32_t e = x->a - rot(x->b, 27);
648        x->a = x->b ^ rot(x->c, 17);
649        x->b = x->c + x->d;
650        x->c = x->d + e;
651        x->d = e + x->a;
652        return x->d;
653    }
654    
655    static void raninit(ranctx *x, uint32_t seed) {
656        uint32_t i;
657        x->a = 0xf1ea5eed, x->b = x->c = x->d = seed;
658        for (i = 0; i<20; ++i) {
659            (void)ranval(x);
660        }
661    }
662    
663    Signed-off-by: Peter Harris <pharris@opentext.com>
664
665commit 4b0d9d3868aad8d5f4266821e9eda586e6c2bfa7
666Author: Peter Harris <pharris@opentext.com>
667Date:   Mon Feb 1 17:43:52 2021 -0500
668
669    Fix build on Windows
670    
671    Notable changes: Protect include of unistd.h (and other POSIX headers).
672    Use SOCKET (which is larger than int) and closesocket (because close is
673    not compatible) for sockets. Use <stdint.h>'s intptr_t instead of the
674    non-portable ssize_t.
675    
676    Signed-off-by: Peter Harris <pharris@opentext.com>
677
678commit cd0fba98a2d0867d505ff1a7ca8d7a7c757acfa2
679Author: Alan Coopersmith <alan.coopersmith@oracle.com>
680Date:   Tue Jun 1 18:36:17 2021 -0700
681
682    xcb_auth: Quiet -Wimplicit-fallthrough warning in get_authptr()
683    
684    xcb_auth.c:135:14: warning: this statement may fall through [-Wimplicit-fallthrough=]
685             addr += 12;
686             ~~~~~^~~~~
687    xcb_auth.c:138:5: note: here
688         case AF_INET:
689         ^~~~
690    
691    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
692
693commit 2ef8655987b9862cdddb72223c9f259a860d5777
694Author: Julien Cristau <jcristau@debian.org>
695Date:   Mon Feb 1 19:23:02 2021 +0100
696
697    Increment libtool version info for libxcb-dri3
698    
699    Somewhat belatedly given the last update was in xcb-proto 1.13 in 2017...
700    
701    Quoting @smcv from https://bugs.debian.org/921069:
702    >>>
703    libxcb-dri3 version 1.13 appears to have added new symbols without increasing
704    the minor ABI version in its -version-info. This will break anything that
705    compares libraries by their version info to decide which one is newer.
706    
707    The Steam Runtime uses libraries' major/minor/micro ABI version info (in this
708    case 0.0.0) to decide whether to use the system copy of a library or the copy
709    in the Steam Runtime, depending on which one is newer (#921026). We can
710    work around this by adding a versioned dependency on libxcb-dri3-0 and
711    deleting the copy from the Steam Runtime, but this isn't a particularly
712    scalable solution.
713    >>>
714
715commit 4cbcac4eca967abfbff7cf1ea473777c5e8c375c
716Author: Ran Benita <ran@unusedvar.com>
717Date:   Tue Nov 17 23:45:14 2020 +0200
718
719    gitignore: add files generated by make check
720    
721    Signed-off-by: Ran Benita <ran@unusedvar.com>
722
723commit f01f3c378eb0168fbb055c7be1c2d08a7acd3752
724Author: Ran Benita <ran@unusedvar.com>
725Date:   Tue Nov 17 23:43:06 2020 +0200
726
727    tests: don't use deprecated fail_unless check API
728    
729    It causes errors like this when running make check:
730    
731    check_public.c:40:24: error: too many arguments for format [-Werror=format-extra-args]
732       40 |   fail_unless(success, "unexpected parse failure %sfor '%s'", test_string[test_type], name);
733    
734    Closes: https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/49
735    Tested-by: Matt Turner <mattst88@gmail.com>
736    Signed-off-by: Ran Benita <ran@unusedvar.com>
737
738commit 704e0a91b1dece9a4cce7cfe8beaeb86a5e14267
739Author: Eduardo Sánchez Muñoz <esm@eduardosm.net>
740Date:   Mon Mar 2 18:59:45 2020 +0100
741
742    Use the 'present' field to properly check that the XC-MISC
743    extension is available in xcb_generate_id.
744    
745    Also document the returned value when xcb_generate_id fails.
746
747commit 4b40b44cb6d088b6ffa2fb5cf3ad8f12da588cef
748Author: Matt Turner <mattst88@gmail.com>
749Date:   Sat Feb 22 12:10:53 2020 -0800
750
751    Release libxcb 1.14
752    
753    Signed-off-by: Matt Turner <mattst88@gmail.com>
754
755commit 78c492deaa7ae3aad0b0eeb4b126bb291cc637a7
756Author: Matt Turner <mattst88@gmail.com>
757Date:   Sat Feb 22 12:19:58 2020 -0800
758
759    Build xz tarballs instead of bzip2
760    
761    Signed-off-by: Matt Turner <mattst88@gmail.com>
762
763commit 8f7e4c4e9fc6ee6b5b5498a7d8af79150be0d26a
764Author: Matt Turner <mattst88@gmail.com>
765Date:   Sat Feb 22 11:24:11 2020 -0800
766
767    configure.ac: Depend on pthread-stubs only on not-Linux
768    
769    Signed-off-by: Matt Turner <mattst88@gmail.com>
770
771commit f9f4b00aad69ff36e81c63089b1b16660eaca900
772Author: Sam Varshavchik <mrsam@courier-mta.com>
773Date:   Sat Jan 4 10:43:59 2020 -0500
774
775    Implement xcb_total_read() and xcb_total_written().
776    
777    Returns raw byte counts that have been read or written to the
778    xcb_connection_t.
779    
780    I found it very useful when developing a high level widget toolkit, to
781    track down inefficient/sub-optimum code that generates a lot of X
782    protocol traffic.
783    
784    Signed-off-by: Sam Varshavchik <mrsam@courier-mta.com>
785
786commit 59e271e15bcecf0c461cd5c6c59081fb86b96c22
787Author: A. Wilcox <AWilcox@Wilcox-Tech.com>
788Date:   Mon Dec 23 21:49:29 2019 -0600
789
790    tests: Support Check 0.13.0 API
791    
792    [mattst88]: Keep compatibility  with old API via preprocessor
793    
794    Fixes: #43
795
796commit 21324989b7e121c008a2c4fdf98547541cbf7b83
797Author: Martin Dørum <martid0311@gmail.com>
798Date:   Sun May 19 16:05:08 2019 +0200
799
800    Handle EINTR from recvmsg in _xcb_in_read
801    
802    I have a GTK application which occasionally crashes with an "interrupted
803    system call" g_message from gdk. After a lot of debugging, I've found
804    that the call to recvmsg in _xcb_in_read occasionally fails with EINTR,
805    and instead of retrying the system call, xcb would just shut down the
806    connection.
807    
808    This change makes _xcb_in_read treat EINTR the same as it would treat
809    EAGAIN; it returns 1 and libX11 ends up calling xcb_poll_for_event
810    again (from what I have understood).
811    
812    I have spoken with a few people who think recvmsg failing with EINTR in
813    this case shouldn't ever happen, and I don't know enough to agree or
814    disagree with that. In case anyone wants to dig further and try to
815    figure out why the recvmsg call sometimes fails with EINTR, here's the
816    backtrace from inside of _xcb_in_read where that happened:
817    
818    Thread 1 "beanbar" hit Breakpoint 1, _xcb_in_read (c=c@entry=0x55ecbe4aba80) at xcb_in.c:1059
819    1059                fprintf(stderr, "Hello World am %s:%i, errno is %s\n", __FILE__, __LINE__, strerror(errno));
820    (gdb) bt
821    0  0x00007fa48fa48639 in _xcb_in_read (c=c@entry=0x55ecbe4aba80) at xcb_in.c:1059
822    1  0x00007fa48fa489d8 in poll_for_next_event (c=0x55ecbe4aba80, queued=queued@entry=0) at xcb_in.c:352
823    2  0x00007fa48fa48a3d in poll_for_next_event (queued=0, c=<optimized out>) at xcb_in.c:722
824    3  0x00007fa48fa48a3d in xcb_poll_for_event (c=<optimized out>) at xcb_in.c:722
825    4  0x00007fa4908d1b7e in poll_for_event (dpy=dpy@entry=0x55ecbe4a9730, queued_only=queued_only@entry=0) at xcb_io.c:245
826    5  0x00007fa4908d1cf0 in poll_for_response (dpy=dpy@entry=0x55ecbe4a9730) at xcb_io.c:303
827    6  0x00007fa4908d1fed in _XEventsQueued (mode=2, dpy=0x55ecbe4a9730) at xcb_io.c:363
828    7  0x00007fa4908d1fed in _XEventsQueued (dpy=dpy@entry=0x55ecbe4a9730, mode=mode@entry=2) at xcb_io.c:344
829    8  0x00007fa4908c3d47 in XPending (dpy=0x55ecbe4a9730) at Pending.c:55
830    9  0x00007fa493cadbc7 in  () at /usr/lib/libgdk-3.so.0
831    10 0x00007fa49234d08a in g_main_context_prepare () at /usr/lib/libglib-2.0.so.0
832    11 0x00007fa49234d6e6 in  () at /usr/lib/libglib-2.0.so.0
833    12 0x00007fa49234d8ae in g_main_context_iteration () at /usr/lib/libglib-2.0.so.0
834    13 0x00007fa4938b920e in g_application_run () at /usr/lib/libgio-2.0.so.0
835    14 0x000055ecbc820af4 in main (argc=1, argv=0x7ffd06238098) at src/main.c:190
836    
837    Signed-off-by: Martin Dørum <martid0311@gmail.com>
838
839commit 656c08c5429a3cf53e7abd7fc56cd3c3a79c0f64
840Author: Jon Turney <jon.turney@dronecode.org.uk>
841Date:   Sun Sep 14 00:05:27 2014 +0100
842
843    Include time.h before using time()
844    
845    Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
846
847commit be1745c8eb00defcb31d336ccc142de056e92bd8
848Author: Alan Coopersmith <alan.coopersmith@oracle.com>
849Date:   Sun Feb 17 12:06:10 2019 -0800
850
851    Add README.md to EXTRA_DIST
852    
853    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
854
855commit 58f37377c851960cbf21fac1caf38ab1e66c7d6c
856Author: Eduardo Sánchez Muñoz <esm@eduardosm.net>
857Date:   Sun Feb 17 13:33:12 2019 +0100
858
859    Add "ge.*" to src/.gitignore
860
861commit 7bac366953005233fc3c8736e6f1cd5798e65e28
862Author: Alan Coopersmith <alan.coopersmith@oracle.com>
863Date:   Sat Feb 16 13:41:28 2019 -0800
864
865    Update README for gitlab migration
866    
867    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
868
869commit 02ff3eadf48e6affe3b59ef688312b4dab538e8b
870Author: Alan Coopersmith <alan.coopersmith@oracle.com>
871Date:   Sat Feb 16 13:20:45 2019 -0800
872
873    Update configure.ac bug URL for gitlab migration
874    
875    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
876
877commit 542befe40a3a7c6a5d1dcb7f38fb9eb261b96b24
878Author: Alan Coopersmith <alan.coopersmith@oracle.com>
879Date:   Mon Jan 7 14:42:53 2019 -0800
880
881    c_client: fix "adress" typo
882    
883    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
884
885commit 8287ebd7b752c33b0cabc4982606fe4831106f7e
886Author: Uli Schlachter <psychon@znc.in>
887Date:   Thu Sep 27 14:04:17 2018 +0200
888
889    Release libxcb 1.13.1
890
891commit bbda345a718ff73086437e51f03fcbb73e4365b9
892Author: Erik Kurzinger <ekurzinger@nvidia.com>
893Date:   Mon Aug 20 12:06:25 2018 -0700
894
895    don't flag extra reply in xcb_take_socket
896    
897    If any flags are specified in a call to xcb_take_socket,
898    they should only be applied to replies for requests sent
899    after that function returns (and until the socket is
900    re-acquired by XCB).
901    
902    Previously, they would also be incorrectly applied to the
903    reply for the last request sent before the socket was taken.
904    For instance, in this example program the reply for the
905    GetInputFocus request gets discarded, even though it was
906    sent before the socket was taken. This results in the
907    call to retrieve the reply hanging indefinitely.
908    
909    static void return_socket(void *closure) {}
910    
911    int main(void)
912    {
913        Display *dpy = XOpenDisplay(NULL);
914        xcb_connection_t *c = XGetXCBConnection(dpy);
915    
916        xcb_get_input_focus_cookie_t cookie = xcb_get_input_focus_unchecked(c);
917        xcb_flush(c);
918    
919        uint64_t seq;
920        xcb_take_socket(c, return_socket, dpy, XCB_REQUEST_DISCARD_REPLY, &seq);
921    
922        xcb_generic_error_t *err;
923        xcb_get_input_focus_reply(c, cookie, &err);
924    }
925    
926    In practice, this has been causing intermittent KWin crashes when
927    used in combination with the proprietary NVIDIA driver such as
928    https://bugs.kde.org/show_bug.cgi?id=386370 since when Xlib fails to
929    retrieve one of these incorrectly discarded replies it triggers
930    an IO error.
931    
932    Signed-off-by: Erik Kurzinger <ekurzinger@nvidia.com>
933    Signed-off-by: Uli Schlachter <psychon@znc.in>
934
935commit 7e0f166579672d71efd819c81f0c932b0acd542c
936Author: Daniel Stone <daniels@collabora.com>
937Date:   Wed Feb 28 01:26:55 2018 +0000
938
939    Release libxcb 1.13
940    
941    Signed-off-by: Daniel Stone <daniels@collabora.com>
942
943commit a3e9821bec08a69b355b46b0655562b9df61bb21
944Author: Daniel Stone <daniels@collabora.com>
945Date:   Wed Apr 26 17:55:54 2017 +0200
946
947    c_client: Add support for lists of FDs
948    
949    Matching xcbgen changes, add support having a ListType which contains
950    file descriptors. Use this to send a variable number of FDs to the
951    server, including when the list size is not fixed.
952    
953    Signed-off-by: Daniel Stone <daniels@collabora.com>
954
955commit c7aa4e682fdc7f0035f928af6eafd052e38cc15a
956Author: Daniel Stone <daniels@collabora.com>
957Date:   Wed Apr 26 17:54:01 2017 +0200
958
959    c_client: Don't serialise non-wire fields
960    
961    For when we have a variable-sized field followed by a fixed field, make
962    sure we do not serialise non-wire fields.
963    
964    Signed-off-by: Daniel Stone <daniels@collabora.com>
965
966commit d10194a321c3db851b3ede9a98cdc95e951943aa
967Author: Christian Linhart <chris@demorecorder.com>
968Date:   Sat Mar 11 22:03:34 2017 +0100
969
970    enable xinput by default
971    
972    Support for the xinput extension is complete now,
973    as far as I can tell.
974    
975    According to our discussion on the list, we enable it now.
976    
977    Signed-off-by: Christian Linhart <chris@demorecorder.com>
978
979commit fad81b63422105f9345215ab2716c4b804ec7986
980Author: David McFarland <corngood@gmail.com>
981Date:   Tue Apr 18 23:58:59 2017 -0300
982
983    read from connection when polling special events and replies
984    
985    Using the mesa vulkan driver, if you acquire an image from a
986    swapchain using a finite timeout (x11_acquire_next_image_poll_x11),
987    it will occasionally lock, calling xcb_poll_for_special_event in
988    a loop until the timeout expires.
989    
990    Call _xcb_in_read() once from the polling functions for special
991    events and replies, in the same way as xcb_poll_for_event.
992    
993    Signed-off-by: David McFarland <corngood@gmail.com>
994    Signed-off-by: Uli Schlachter <psychon@znc.in>
995
996commit f830eb93c9c38b2c6c7ea2971af3bc6a61e92277
997Author: Tobias Stoeckmann <tobias@stoeckmann.org>
998Date:   Sun Mar 26 15:41:12 2017 +0200
999
1000    Check strdup for NULL return value.
1001    
1002    _xcb_open does not check strdup's return value for NULL if launchd suport
1003    was configured.
1004    
1005    Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
1006    Signed-off-by: Uli Schlachter <psychon@znc.in>
1007
1008commit ee9dfc9a7658e7fe75d27483bb5ed1ba4d1e2c86
1009Author: Christian Linhart <chris@demorecorder.com>
1010Date:   Wed Jan 25 10:21:05 2017 +0100
1011
1012    add support for eventstruct
1013    
1014    eventstruct allows to use events as part of requests.
1015    This is, e.g., needed by xcb_input_send_extension_event.
1016    
1017    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1018
1019commit 0c2c5d50f8670da3e7601feb6a29b53509513da5
1020Author: Christian Linhart <chris@demorecorder.com>
1021Date:   Fri Jan 20 20:14:57 2017 +0100
1022
1023    optionally build the GE extension
1024    
1025    xcb contains an xml-definition for the GenericEvent extension
1026    but this extension was neither generated nor built.
1027    
1028    This patch enables optional building of the GenericEvent extension
1029    with configure option --enable-ge
1030    
1031    By default, the GenericEvent extension is not built.
1032    Normally this is not needed by application programs
1033    because there is implicit support for the GE-extension
1034    for the specific events built with this extension.
1035    
1036    But it may be useful for X-protocol analyzers and stuff like that.
1037    
1038    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1039
1040commit 9bce1f72e329cb407b7a95589b9675a08129b65d
1041Author: Christian Linhart <chris@demorecorder.com>
1042Date:   Fri Jan 20 14:40:25 2017 +0100
1043
1044    move symbol lookup of sumof expr to the parser
1045    
1046    replace the complicated symboltable lookup for sumof expr
1047    by accessing the lenfield of the expr-object.
1048    
1049    This requires the corresponding patch for xcb/proto
1050    which sets the lenfield accordingly.
1051    
1052    This should be OK because for official releases we define
1053    that dependency in the build system.
1054    
1055    For getting versions off the HEAD of the git repo, it should
1056    be obvious that xcb/proto and xcb/libxcb have to be updated together.
1057    
1058    I have tested this patch and it generates exactly the same code
1059    as before.
1060    
1061    Tested-by: Christian Linhart <chris@demorecorder.com>
1062    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1063
1064commit 65b298c7ca317d7e4316aa2b9e0499e13047c65c
1065Author: Alan Coopersmith <alan.coopersmith@oracle.com>
1066Date:   Sat Feb 6 12:26:21 2016 -0800
1067
1068    Correct @param "e" to "error" in xcb_poll_for_reply*()
1069    
1070    Found by clang -Wdocumentation:
1071    
1072    ./xcbext.h:271:11: warning: parameter 'e' not found in the function
1073          declaration [-Wdocumentation]
1074     * @param e Location to store errors in, or NULL. Ignored for un...
1075              ^
1076    ./xcbext.h:271:11: note: did you mean 'error'?
1077     * @param e Location to store errors in, or NULL. Ignored for un...
1078              ^
1079              error
1080    
1081    ./xcbext.h:283:11: warning: parameter 'e' not found in the function
1082          declaration [-Wdocumentation]
1083     * @param e Location to store errors in, or NULL. Ignored for un...
1084              ^
1085    ./xcbext.h:283:11: note: did you mean 'error'?
1086     * @param e Location to store errors in, or NULL. Ignored for un...
1087              ^
1088              error
1089    
1090    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
1091    Signed-off-by: Uli Schlachter <psychon@znc.in>
1092
1093commit 32a9084546add979115a686f9a167e70b8967149
1094Author: Alan Coopersmith <alan.coopersmith@oracle.com>
1095Date:   Sat Feb 6 12:26:20 2016 -0800
1096
1097    Remove : from @param names in manually written headers
1098    
1099    Makes style match the @param names in autogenerated headers and makes
1100    clang -Wdocumentation stop complaining about all of them:
1101    
1102    ./xcb.h:523:11: warning: parameter 'display:' not found in the function
1103          declaration [-Wdocumentation]
1104     * @param display: A pointer to the display number.
1105              ^~~~~~~~
1106    ./xcb.h:523:11: note: did you mean 'display'?
1107     * @param display: A pointer to the display number.
1108              ^~~~~~~~
1109              display
1110    
1111    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
1112    Signed-off-by: Uli Schlachter <psychon@znc.in>
1113
1114commit 8740a288ca468433141341347aa115b9544891d3
1115Author: Thomas Klausner <wiz@NetBSD.org>
1116Date:   Thu May 19 17:31:18 2016 +0200
1117
1118    Fix inconsistent use of tabs vs. space.
1119    
1120    Needed for at least python-3.5.x.
1121    
1122    Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
1123    Signed-off-by: Uli Schlachter <psychon@znc.in>
1124
1125commit d34785a34f28fa6a00f8ce00d87e3132ff0f6467
1126Author: Uli Schlachter <psychon@znc.in>
1127Date:   Sat May 14 10:36:54 2016 +0200
1128
1129    Release libxcb 1.12
1130    
1131    Signed-off-by: Uli Schlachter <psychon@znc.in>
1132
1133commit b11fca06f75b26b94e3f1d3e3c3954d365f80759
1134Author: Uli Schlachter <psychon@znc.in>
1135Date:   Sat May 14 10:33:54 2016 +0200
1136
1137    Bump xcb-proto requirement to 1.12
1138    
1139    This is needed due to various changes that were done to the XML schema.
1140    
1141    Signed-off-by: Uli Schlachter <psychon@znc.in>
1142
1143commit 095353ff1a4f611922dfc4c98b0c4bd55d9f6d4f
1144Author: Mark Kettenis <kettenis@openbsd.org>
1145Date:   Sat Jan 23 17:29:32 2016 +0100
1146
1147    Increase unix socket send buffer to at least 64KB
1148    
1149    Some systems (e.g. OpenBSD) have a rather small default socket send buffer
1150    size of 4KB.  The result is that sending requests with a largish payload
1151    requires serveral writev(2) system calls.  Make sure the socket send buffer
1152    is at least 64KB such that we're likely to succeed with a single system
1153    call for most requests.  A similar change was made to the xtrans code
1154    some time ago.
1155    
1156    Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
1157    Reviewed-by: Matthieu Herrb <matthieu@herrb.eu>
1158
1159commit b3516102b4469df99db39e0e354deae496867f35
1160Author: Christian Linhart <chris@demorecorder.com>
1161Date:   Mon Jan 18 06:56:39 2016 +0100
1162
1163    do not serialize pads by default anymore
1164    
1165    Pads should not be serialized/deserialized to maintain
1166    ABI compatibility when adding explicit align pads.
1167    
1168    Therefore this pad switches off serialization of pads
1169    unless it is enforced by serialize=true in the xml-definition
1170    of that pad
1171    
1172    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1173
1174commit c03388ff9e253b86e4b41fb77034ca6007fe47e9
1175Author: Jaya Tiwari <tiwari.jaya18@gmail.com>
1176Date:   Wed Nov 11 01:02:09 2015 +0100
1177
1178    calculate lengthless list
1179    
1180    Some rework done by Christian Linhart
1181    
1182    Signed-off-by: Jaya Tiwari <tiwari.jaya18@gmail.com>
1183    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1184
1185commit 775825756714eb6b8467e099cda73a03b782ea0e
1186Author: Christian Linhart <chris@demorecorder.com>
1187Date:   Tue Nov 10 12:53:04 2015 +0100
1188
1189    Fix handling of align-pads in end-iterators
1190    
1191    If a list is preceded by an align-pad, then
1192    accessor for the end-iterator returned a wrong
1193    value.
1194    
1195    Reason: the length of the align-iterator was added
1196    to a pointer of list-member type. Therefore, the length
1197    was multiplied by the size of the list-member type,
1198    due to C pointer arithmetic rules.
1199    
1200    This has looked like the following, e.g., in
1201    xcb_randr_get_crtc_transform_pending_params_end:
1202    
1203    i.data = ((xcb_render_fixed_t *) prev.data) + ((-prev.index) & (4 - 1)) + (R->pending_nparams);
1204    
1205    This bug was introduced with the following commit:
1206    http://cgit.freedesktop.org/xcb/libxcb/commit/?id=4033d39d4da21842bb1396a419dfc299591c3b1f
1207    
1208    The fix handles this by casting to char* before adding the align,
1209    and then casting the result to the member type.
1210    
1211    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1212
1213commit 32a2189183696e942b002efcbca823a416fe5f6a
1214Author: Christian Linhart <chris@demorecorder.com>
1215Date:   Sun Nov 1 18:35:35 2015 +0100
1216
1217    set the align-offset as provided by proto
1218    
1219    instead of using the lower bits of the pointer address.
1220    This fixes a bug reported by Peter Hutterer in off-list communication
1221    back in June 2015.
1222    
1223    This requires the alignment-checker patches in xcb/proto.
1224    
1225    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1226
1227commit 6e0378ebbf5f842992e37c93ad60b960427849ac
1228Author: Adam Jackson <ajax@redhat.com>
1229Date:   Mon Sep 21 15:27:52 2015 -0400
1230
1231    Bump version to 1.11.90
1232    
1233    We've released 1.11.1 and new libX11 wants that or better.  git master
1234    will suffice, so bump the version number ahead of 1.11 branch.
1235    
1236    Signed-off-by: Adam Jackson <ajax@redhat.com>
1237
1238commit 4033d39d4da21842bb1396a419dfc299591c3b1f
1239Author: Christian Linhart <chris@demorecorder.com>
1240Date:   Fri Sep 19 13:44:38 2014 +0200
1241
1242    make lists after align-pads work
1243    
1244    Handle align-pads when generating an end-function
1245    in the same way as handling them when generating
1246    an accessor or iterator function.
1247    
1248    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1249
1250commit b15c96f9507119e5d38a61d92b4dbcd479ea2099
1251Author: Christian Linhart <chris@demorecorder.com>
1252Date:   Thu Jun 11 18:58:38 2015 +0200
1253
1254    make support for server side stuff optional
1255    
1256    and make it disabled by default with an EXPERIMENTAL warning
1257    
1258    reason: this feature is unfinished and we want to have flexibility for
1259    ABI/API changes, while still being able to make a release soon
1260    
1261    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1262
1263commit c5d923d8ff4a9b9fc1aef1c6c6918bab15098d34
1264Author: Jon TURNEY <jon.turney@dronecode.org.uk>
1265Date:   Tue Mar 17 17:49:14 2015 +0000
1266
1267    Link with winsock library for socket functions on MinGW
1268    
1269    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
1270
1271commit 5b40681c887192307f3ae147d2158870aa79c05f
1272Author: Uli Schlachter <psychon@znc.in>
1273Date:   Fri Jun 12 15:13:05 2015 +0200
1274
1275    Fix a thread hang with xcb_wait_for_special_event()
1276    
1277    Consider the following:
1278    
1279    - Two threads are calling xcb_wait_for_special_event() and xcb_wait_for_reply()
1280      concurrently.
1281    - The thread doing xcb_wait_for_reply() wins the race and poll()s the socket for
1282      readability.
1283    - The other thread will be put to sleep on the special_event_cond of the special
1284      event (this is done in _xcb_conn_wait() via the argument
1285      xcb_wait_for_special_event() gives it).
1286    - The first thread gets its reply, but does not yet receive any special event.
1287    
1288    In this case, the first thread will return to its caller. On its way out, it
1289    will call _xcb_in_wake_up_next_reader(), but that function cannot wake up
1290    anything since so far it did not handle xcb_wait_for_special_event().
1291    
1292    Thus, the first thread stays blocked on the condition variable and no thread
1293    tries to read from the socket.
1294    
1295    A test case demonstrating this problem is available at the bug report.
1296    
1297    Fix this similar to how we handle this with xcb_wait_for_reply():
1298    
1299    The function wait_for_reply() adds an entry into a linked list of threads that
1300    wait for a reply. Via this list, _xcb_in_wake_up_next_reader() can wake up this
1301    thread so that it can call _xcb_conn_wait() again and then poll()s the socket.
1302    
1303    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84252
1304    Signed-off-by: Uli Schlachter <psychon@znc.in>
1305    Tested-by: Michel Dänzer <michel.daenzer@amd.com>
1306
1307commit f85661c3bca97faa72431df92a3867be39a74e23
1308Author: Michel Dänzer <michel.daenzer@amd.com>
1309Date:   Mon Jun 1 11:04:18 2015 +0900
1310
1311    Call _xcb_wake_up_next_reader from xcb_wait_for_special_event
1312    
1313    All functions calling _xcb_conn_wait() must make sure that waiting
1314    readers are woken up when we read a reply or event that they are waiting
1315    for. xcb_wait_for_special_event() did not do so. This adds the missing
1316    call to_xcb_in_wake_up_next_reader().
1317    
1318    Fixes deadlock when waiting for a special event and concurrently
1319    processing the display connection queue in another thread.
1320    
1321    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84252
1322    Tested-by: Thomas Daede <bztdlinux@gmail.com>
1323    Tested-by: Clément Guérin <geecko.dev@free.fr>
1324    Reviewed-by: Uli Schlachter <psychon@znc.in>
1325    Signed-off-by: Michel Dänzer <michel@daenzer.net>
1326    Signed-off-by: Uli Schlachter <psychon@znc.in>
1327
1328commit 8584c0e09573a29d8ba7050e3d5afd925b4d8d80
1329Author: Uli Schlachter <psychon@znc.in>
1330Date:   Thu May 14 09:44:05 2015 +0200
1331
1332    send_fds(): Handle too many outstanding FDs to send
1333    
1334    Before this patch, the following code caused an endless loop in send_fds(),
1335    because the queue of FDs to send was eventually full, but _xcb_out_flush_to()
1336    didn't make any progress, since there was no request to send:
1337    
1338       while (1) { xcb_send_fd(conn, dup(some_fd)); }
1339    
1340    Fix this by sending a sync when flushing didn't make any progress. That way we
1341    actually have something to send and can attach the pending FDs.
1342    
1343    Because send_fds() can now send requests, the code in
1344    xcb_send_request_with_fds64() has to be changed. It has to call send_fds()
1345    before it establishes a good sequence number for the request it wants to send.
1346    
1347    Signed-off-by: Uli Schlachter <psychon@znc.in>
1348
1349commit 658fb4a5f0050db68fdf092936afe596412ef5f7
1350Author: Uli Schlachter <psychon@znc.in>
1351Date:   Wed Apr 22 09:26:05 2015 +0200
1352
1353    Code generator: Use xcb_send_request_with_fds()
1354    
1355    Signed-off-by: Uli Schlachter <psychon@znc.in>
1356
1357commit b15aa6bd4efde784e546d168bb23b8a8e816e85b
1358Author: Uli Schlachter <psychon@znc.in>
1359Date:   Wed Apr 22 09:23:47 2015 +0200
1360
1361    Add xcb_send_request_with_fds() and *_with_fds64()
1362    
1363    Doing xcb_send_fd(), xcb_send_request() is racy. If two threads do this at the
1364    same time, they could mix up their file descriptors. This commit makes it
1365    possibly to fix this race by providing a single function which does everything
1366    that is needed.
1367    
1368    Signed-off-by: Uli Schlachter <psychon@znc.in>
1369
1370commit cc04cfb41bece6ec239f57d83822286b507f4482
1371Author: Uli Schlachter <psychon@znc.in>
1372Date:   Mon May 18 21:40:34 2015 +0200
1373
1374    send_fds(): Make sure no other thread interrupts us
1375    
1376    Two threads trying to send fds at the same time could interfere. To guarantee a
1377    correct ordering, we have to use correct locking. The code in send_fds() missed
1378    one case: If there was another thread already writing requests, we slept on the
1379    "done with writing" condition variable (c->out.cond). This would allow other
1380    threads to re-acquire the iolock before us and could cause fds to be sent out of
1381    order.
1382    
1383    To fix this, at the beginning of send_fds() we now make sure that no other
1384    thread is already writing requests. This is what prepare_socket_request() does.
1385    Additionally, it gets the socket back in case xcb_take_socket() was called,
1386    which is a good thing, too, since fds are only sent with corresponding requests.
1387
1388commit 25f9e7e45a7652b35b71c7941beef774a39f0d86
1389Author: Uli Schlachter <psychon@znc.in>
1390Date:   Wed Apr 22 09:20:38 2015 +0200
1391
1392    xcb_send_fd(): Always close fds
1393    
1394    The API docs for xcb_send_fd() says "After this function returns, the file
1395    descriptor given is owned by xcb and will be closed eventually".
1396    
1397    Let the implementation live up to its documentation. We now also close fds if fd
1398    passing is unavailable (!HAVE_SENDMSG) and when the connection is in an error
1399    state.
1400    
1401    (This also does sneak in some preparatory functions for follow-up commits and
1402    thus does things in a more complicated way than really necessary.)
1403    
1404    Signed-off-by: Uli Schlachter <psychon@znc.in>
1405
1406commit bbdf1d133f7bd979c6ff3bf44ec3d0c2d2b9dbfe
1407Author: Ran Benita <ran234@gmail.com>
1408Date:   Wed Mar 18 12:27:32 2015 +0200
1409
1410    c_client.py: don't generate useless empty /** < */ comments
1411    
1412    (This does not change doxygen's output or warnings).
1413    
1414    Signed-off-by: Ran Benita <ran234@gmail.com>
1415    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1416
1417commit ff6cb3913b64a4aa29cb5e65168ea49d77195296
1418Author: Ran Benita <ran234@gmail.com>
1419Date:   Wed Mar 18 12:27:31 2015 +0200
1420
1421    c_client.py: use pattern matching with enumerate()
1422    
1423    Signed-off-by: Ran Benita <ran234@gmail.com>
1424
1425commit cb621341a62e6d2233db3e337611f6fdd4f675a6
1426Author: Christian Linhart <chris@demorecorder.com>
1427Date:   Wed Apr 29 09:11:37 2015 +0200
1428
1429    expose 64-bit sequence numbers for XLib
1430    
1431    While XCB uses 64-bit sequence number internally, it only exposes
1432    "unsigned int" so that, on 32-bit architecture, Xlib based applications
1433    may see their sequence number wrap which causes the connection to the X
1434    server to be lost.
1435    
1436    Expose 64-bit sequence number from XCB API so that Xlib and others can
1437    use it even on 32-bit environment.
1438    
1439    This implies the following API addition:
1440    
1441      xcb_send_request64()
1442      xcb_discard_reply64()
1443      xcb_wait_for_reply64()
1444      xcb_poll_for_reply64()
1445    
1446    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71338
1447    
1448    Reviewed-by: Uli Schlachter <psychon@znc.in>
1449    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1450    Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
1451
1452commit c49aa985941112be05599032b9bb45b2652301ce
1453Author: Alan Coopersmith <alan.coopersmith@oracle.com>
1454Date:   Wed Apr 29 23:23:05 2015 -0700
1455
1456    Escape \n to display properly in xcb-requests man page
1457    
1458    In nroff, \n is a macro that "Interpolates number register x" (where x
1459    is the character following the \n sequence), thus the man page currently
1460    prints 0 instead of \n" in several lines, leading to output such as:
1461    
1462     printf("The _NET_WM_NAME atom has ID %u0, reply-⁠>atom);
1463    
1464    It needs to be escaped here, as \\n, as is done in other examples in
1465    this man page already.
1466    
1467    https://bugs.freedesktop.org/show_bug.cgi?id=90231
1468    
1469    Reported-by: Stefan Merettig
1470    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
1471
1472commit a90be9955d2c5a635f791d44db1154633b9d3322
1473Author: Ran Benita <ran234@gmail.com>
1474Date:   Sun Oct 12 21:58:20 2014 +0300
1475
1476    c_client.py: make condition easier to follow in _c_complex()
1477    
1478    Signed-off-by: Ran Benita <ran234@gmail.com>
1479    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1480
1481commit f9f925107e37e0c9a0ed8220ee3a23f584e3b2ec
1482Author: Ran Benita <ran234@gmail.com>
1483Date:   Sun Oct 12 21:58:14 2014 +0300
1484
1485    c_client.py: don't add /* <name> */ before references to 'S'
1486    
1487    The name can be understood from the type of S already.
1488    
1489    For examples, look for 'S->' in xkb.c or xinput.c.
1490    
1491    Signed-off-by: Ran Benita <ran234@gmail.com>
1492    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1493    Reviewed-by: Rémi Cardona <remi@gentoo.org>
1494
1495commit 17f9bda6c291a8b9ccc98b22c241b64880b80621
1496Author: Ran Benita <ran234@gmail.com>
1497Date:   Sun Oct 12 21:58:19 2014 +0300
1498
1499    c_client.py: remove duplicated `cookie_type` argument for requests
1500    
1501    It is implied already inside the function by the `void` argument.
1502    
1503    Signed-off-by: Ran Benita <ran234@gmail.com>
1504    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1505
1506commit c65005e9d0ce60524d2e883c13c027a5f1f7914c
1507Author: Ran Benita <ran234@gmail.com>
1508Date:   Sun Oct 12 21:58:18 2014 +0300
1509
1510    c_client.py: spell out keyword arguments in c_request() for clarity
1511    
1512    Signed-off-by: Ran Benita <ran234@gmail.com>
1513    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1514
1515commit 6872e925828e615e2c1d4a0bbcc6be6a32fa4e7a
1516Author: Ran Benita <ran234@gmail.com>
1517Date:   Sun Oct 12 21:58:17 2014 +0300
1518
1519    c_client.py: simplify _c_reply_has_fds()
1520    
1521    Signed-off-by: Ran Benita <ran234@gmail.com>
1522    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1523
1524commit 8bf8b62316a3066b61243d797d22c2100fd173f5
1525Author: Ran Benita <ran234@gmail.com>
1526Date:   Sun Oct 12 21:58:16 2014 +0300
1527
1528    c_client.py: remove commented debug statements
1529    
1530    Signed-off-by: Ran Benita <ran234@gmail.com>
1531    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1532
1533commit ec435aebd6ed0523de2d341e6aff5ae66f230b10
1534Author: Ran Benita <ran234@gmail.com>
1535Date:   Sun Oct 12 21:58:15 2014 +0300
1536
1537    c_client.py: use C99 initializers instead of comments
1538    
1539    Signed-off-by: Ran Benita <ran234@gmail.com>
1540    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1541
1542commit 89498d1d450d0e6e476f6b24908ecae857863386
1543Author: Ran Benita <ran234@gmail.com>
1544Date:   Sun Oct 12 21:58:11 2014 +0300
1545
1546    c_client.py: remove end-of-function comments
1547    
1548    Signed-off-by: Ran Benita <ran234@gmail.com>
1549    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1550
1551commit 2871d4b1b8736044ba50df2f7dacfcc65d820fa9
1552Author: Ran Benita <ran234@gmail.com>
1553Date:   Sun Oct 12 21:58:10 2014 +0300
1554
1555    c_client.py: no need to compare bools to True/False
1556    
1557    Signed-off-by: Ran Benita <ran234@gmail.com>
1558    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1559
1560commit 30976e5255a25117d17d4d8162e29b84913953cd
1561Author: Ran Benita <ran234@gmail.com>
1562Date:   Sun Oct 12 21:58:09 2014 +0300
1563
1564    c_client.py: use "foo".join() instead of reduce
1565    
1566    Signed-off-by: Ran Benita <ran234@gmail.com>
1567    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1568
1569commit 0ab52cbcc630ff8ddfa6f6b245c6b3071867b291
1570Author: Ran Benita <ran234@gmail.com>
1571Date:   Sun Oct 12 21:58:08 2014 +0300
1572
1573    c_client.py: fix indentation
1574    
1575    (Also remove unnecessary parens around the condition).
1576    
1577    Signed-off-by: Ran Benita <ran234@gmail.com>
1578    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1579
1580commit 80341d5df3e30e2cfc3066d05256af9e513e4500
1581Author: Ran Benita <ran234@gmail.com>
1582Date:   Sun Oct 12 21:58:07 2014 +0300
1583
1584    c_client.py: use comprehensions instead of map/filter
1585    
1586    Signed-off-by: Ran Benita <ran234@gmail.com>
1587    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1588
1589commit 86ea6645d9e388112670cf88da1fe3b350953789
1590Author: Ran Benita <ran234@gmail.com>
1591Date:   Sun Oct 12 21:58:06 2014 +0300
1592
1593    c_client.py: use print as a function for python3 compatibility
1594    
1595    This works for all python>=2.6, which is what configure requires.
1596    
1597    Signed-off-by: Ran Benita <ran234@gmail.com>
1598    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1599
1600commit 70d32ce7d8ef3ce9490d9a51b189829c8bcdb5c4
1601Author: Ran Benita <ran234@gmail.com>
1602Date:   Sun Oct 12 21:58:05 2014 +0300
1603
1604    c_client.py: fix pyflakes errors
1605    
1606    c_client.py:2: 'from xml.etree.cElementTree import *' used; unable to detect undefined names
1607    c_client.py:3: 'basename' imported but unused
1608    c_client.py:9: 'time' imported but unused
1609    c_client.py:1437: local variable 'list_obj' is assigned to but never used
1610    c_client.py:1745: local variable 'varfield' is assigned to but never used
1611    c_client.py:2050: local variable 'length' is assigned to but never used
1612    c_client.py:2416: local variable 'R_obj' is assigned to but never used
1613    c_client.py:2441: local variable 'S_obj' is assigned to but never used
1614    
1615    Signed-off-by: Ran Benita <ran234@gmail.com>
1616    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1617
1618commit 12d23b934f189607ad77667834205dfcf89a41f0
1619Author: Ran Benita <ran234@gmail.com>
1620Date:   Sun Oct 12 21:58:04 2014 +0300
1621
1622    c_client.py: simplify maximum expression
1623    
1624    Signed-off-by: Ran Benita <ran234@gmail.com>
1625    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1626
1627commit 1b37d6ad3a37a9d0201e3b2caee6cd53b7b4b3dc
1628Author: Ran Benita <ran234@gmail.com>
1629Date:   Sun Oct 12 21:58:03 2014 +0300
1630
1631    c_client.py: remove unneeded call to get_serialize_params()
1632    
1633    The results are not used, and the function doesn't have side effects.
1634    
1635    Signed-off-by: Ran Benita <ran234@gmail.com>
1636    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1637
1638commit e3ec1f74637237ce500dfd0ca59f2e422da4e019
1639Author: Jaya Tiwari <tiwari.jaya18@gmail.com>
1640Date:   Thu Jan 22 12:16:33 2015 -0500
1641
1642    Adding accessors for requests
1643    
1644    Added accessor functions for requests the same way they were added for
1645    structs,events and replies.
1646    Lists for replies have accessor functions now.
1647    
1648    Signed-off-by: Jaya Tiwari <tiwari.jaya18@gmail.com>
1649    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1650    
1651    Comment from the Reviewer Christian Linhart:
1652    I have tested your patch after fixing the issues with the patch-format.
1653    It looks good:
1654    * only adds new functions, and does not modify existing functions.
1655      Therefore it is API and ABI compatible.
1656    
1657    * adds accessors for varsized-stuff in requests.
1658      This is needed for server-side XCB and may be useful for implementing X11-protocol proxies.
1659
1660commit 5353c0216e091b64d01a43e6580e6d69b2ac16c7
1661Merge: bbca7b8 c6f3fb2
1662Author: Christian Linhart <chris@demorecorder.com>
1663Date:   Tue Feb 10 10:13:04 2015 +0100
1664
1665    Merge http://git.demorecorder.com/git/free-sw/xcb/libxcb
1666    branch 'ParametrizedStruct-V7'
1667
1668commit c6f3fb2529a6211221e8254f58c85fd67c1d8844
1669Author: Christian Linhart <chris@DemoRecorder.com>
1670Date:   Mon Nov 3 09:57:59 2014 +0100
1671
1672    generator: support parametrized structs
1673    
1674    Parametrized structs contain paramref expressions which
1675    refer to the value of a field defined in the context
1676    where the struct is used.
1677    
1678    Implementing the parametrized structs turned out
1679    to be somewhat easier than previously thought
1680    because the generator already had some support for type-parametrization
1681    because this is needed when case or bitcase refers to fields outside
1682    of the switch.
1683    
1684    So I decided to go with the flow and to implement the solution
1685    which best fits the current implementation.
1686    
1687    I did the following:
1688    * I provided a way to specify fieldref with an explicitely given type:
1689      This resulted in <paramref type="CARD8>fieldname</paramref>
1690      A paramref is just a fieldref with an explicit type.
1691      The type is necessary because there is no local field of that
1692      name where the type can be derived from.
1693    
1694    * then I tested it and made several changes in the generator
1695      such that it really works.
1696    
1697    Basically the generated code is as follows:
1698    * The parameter appears on the parameter list of the
1699      sizeof-function of the parametrized struct.
1700      When that function gets called, an appropriate argument is supplied.
1701    
1702    * The parameter also appears as an additional member of the iterator-struct
1703      for the iterator of lists of that parametrized struct.
1704      This way, the next-function can get the value of that parameter from the iterator.
1705      When the iterator is created, this iterator-member is set accordingly.
1706    
1707    * When the paramref appears in the length-expression of a list, then
1708      the parameter appears on the parameterlist of the "length" and "end" functions.
1709      When these functions get called, an appropriate argument is supplied.
1710    
1711    Some comments:
1712    * I did not implement inline structs.
1713      This would probably have been more complicated, and at least some additional effort.
1714      But that can be implemented later if needed.
1715      (Inline structs could probably use some code from switch-case/bitcase which is already kind of
1716      an inlined struct but one has to be careful not to break the functionality
1717      of switch-case/bitcase. Support for inline structs inside lists must probably
1718      be implemented from scratch...)
1719    
1720    * The paramref expression refers to a field of the same name in the struct/request/...
1721      where it is used.
1722      So it is not possible to pass the value of arbitrary fields or even expressions
1723      to the parametrized struct.
1724      This would have been possible with the previously discussed <typearg>.
1725      That can be added later, if needed.
1726      ( Wont be too complicated )
1727    
1728    * So this is pretty much like the proposal from Ran Benita.
1729    
1730    changes for V2 of this patch, according to suggestions from Ran Benita:
1731    * replace map with list comprehension
1732      because map returns an iterator instead of a list from Python 3 on,
1733      so it cannot be added to a list anymore.
1734    
1735    * removed "self" parameter of function additional_params_to_str
1736      and accessed the variable additional_params from the outer
1737      function directly.
1738    
1739    changes for V2 of this patch:
1740    * adapt to revision 2 of patchset ListInputDevices
1741    * style fixes for similar things that Ran Benita has found in my previous patches
1742    
1743    Message-ID: <54574397.4060000@DemoRecorder.com>
1744    Patch-Thread-Subject: [Xcb] parametrized structs implemented
1745    Patch-Set: ParametrizedStruct
1746    Patch-Number: libxcb 1/1
1747    Patch-Version: V3
1748    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1749
1750commit 912cd97a6dd019e9e7ecf09c82c4577dd2ad7529
1751Author: Christian Linhart <chris@DemoRecorder.com>
1752Date:   Mon Nov 3 09:58:08 2014 +0100
1753
1754    generator: support listelement-ref
1755    
1756    Support for listelement-ref needs the following three changes
1757    (in the order as they appear in the patch):
1758    
1759    * making the current list-element accessible with the variable
1760      xcb_listelement which is a pointer to the list-element
1761    
1762    * supporting lists of simple-type for sumof with a nested expression
1763    
1764    * using the variable for resolving a listelement-ref expression
1765    
1766    Changes for V2 of this patch:
1767    - adapt to removal of patch "libxcb 2/6" from patchset "ListInputDevices".
1768    
1769    Changes for V3 of this patch:
1770    - adapt to V2 of patch "libxcb 5/6" from patchset "ListInputDevices"
1771    
1772    Changes for V4 of this patch:
1773    - adapt to revision 2 of the patchset "ListInputDevices"
1774    
1775    Message-ID: <545743A0.50907@DemoRecorder.com>
1776    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1777    Patch-Set: PopcountList
1778    Patch-Number: libxcb 4/4
1779    Patch-Version: V4
1780    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1781
1782commit 422458b66380e4103c4937f0e2e8bb93e31f273a
1783Author: Christian Linhart <chris@demorecorder.com>
1784Date:   Thu Sep 4 17:50:50 2014 +0200
1785
1786    generator: _c_accessor_get_length: remove buggy special case
1787    
1788    The function _c_accessor_get_length had a special case handling
1789    for intermixed var and fixed size fields.
1790    
1791    However:
1792    * The implementation of that special case was buggy:
1793      It tried to call a python-dict as a function which causes
1794      Python to abort the program with a stacktrace and error message.
1795      So this code was never used.
1796    
1797    * The case it tried to handle is handeled elsewhere in the
1798      meantime: in _c_helper_absolute_name by previous patches
1799      made by me.
1800    
1801    Message-ID: <1409845851-38950-3-git-send-email-chris@demorecorder.com>
1802    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1803    Patch-Set: PopcountList
1804    Patch-Number: libxcb 3/4
1805    Patch-Version: V1
1806    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1807
1808commit b1e4a3bbd8194d12d7fcd9705fcbbe0deb59bcba
1809Author: Christian Linhart <chris@demorecorder.com>
1810Date:   Thu Sep 4 17:50:49 2014 +0200
1811
1812    generator: generate accessors for events, too
1813    
1814    Accessors are generally needed for var-sized fields
1815    and fields after var-sized fields.
1816    
1817    Generic events can have ver-sized fields.
1818    Therefore they need accessors.
1819    
1820    Message-ID: <1409845851-38950-2-git-send-email-chris@demorecorder.com>
1821    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1822    Patch-Set: PopcountList
1823    Patch-Number: libxcb 2/4
1824    Patch-Version: V1
1825    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1826
1827commit 6234225b4be862c5882bf547d066c5a3885906dd
1828Author: Christian Linhart <chris@demorecorder.com>
1829Date:   Thu Sep 4 17:50:48 2014 +0200
1830
1831    generator: no type-setup for eventcopies anymore
1832    
1833    _c_type_setup is not called for eventcopies anymore:
1834    Reasons:
1835    * the type-setup of an eventcopy would overwrite members of the original
1836      event object such as c_type, ...
1837    * it is needed for the next patch, i.e., generating accessors:
1838      type_setup would create sizeof-etc funtions which called
1839      undefined accessor functions.
1840    
1841    Sizeof-functions are generated for compatibility:
1842    Reason:
1843    * Type-setup of eventcopies has previously generated
1844      sizeof-functions for eventcopies.
1845      So, we still need to generate these functions.
1846      These new sizeof-functions simply call the sizeof-function
1847      of the defining event of the eventcopy.
1848    
1849    Message-ID: <1409845851-38950-1-git-send-email-chris@demorecorder.com>
1850    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1851    Patch-Set: PopcountList
1852    Patch-Number: libxcb 1/4
1853    Patch-Version: V1
1854    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1855
1856commit 18ff453edd42712ea4d1e7218bbe8829f9f4caba
1857Author: Christian Linhart <chris@DemoRecorder.com>
1858Date:   Sun Nov 2 13:46:58 2014 +0100
1859
1860    _c_helper_fieldaccess_expr: remove handling for empty sep
1861    
1862    The loop-variable "sep" is never empty in function
1863    "_c_helper_fieldaccess_expr", after a fix elsewhere.
1864    Therefore I removed the handling of the case of "sep" being empty.
1865    
1866    Thanks to Ran Benita for the hint that this can be removed.
1867    
1868    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1869    Reviewed-by: Ran Benita <ran234@gmail.com>
1870    
1871    Message-ID: <545627C2.3050608@DemoRecorder.com>
1872    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1873    Patch-Set: ListInputDevices
1874    Patch-Number: libxcb 9/9
1875    Patch-Version: V1
1876
1877commit d905b886185fd8e0d9d7f70e0d112cb58ab86f98
1878Author: Christian Linhart <chris@DemoRecorder.com>
1879Date:   Sun Nov 2 13:46:50 2014 +0100
1880
1881    function _c_helper_fieldaccess_expr: improve description
1882    
1883    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1884    Reviewed-by: Ran Benita <ran234@gmail.com>
1885    
1886    Message-ID: <545627BA.1000909@DemoRecorder.com>
1887    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1888    Patch-Set: ListInputDevices
1889    Patch-Number: libxcb 8/9
1890    Patch-Version: V1
1891
1892commit 17f6e04493c93014beaf3a704c8cc4d050cb13bf
1893Author: Christian Linhart <chris@DemoRecorder.com>
1894Date:   Sun Nov 2 13:46:38 2014 +0100
1895
1896    rename _c_helper_absolute_name to _c_helper_fieldaccess_expr
1897    
1898    The function _c_helper_absolute_name was named in
1899    a misleading way.
1900    It computes a C-expression for accessing a field of an xcb-type.
1901    
1902    Therefore the name _c_helper_fieldaccess_expr is more appropriate.
1903    
1904    Note: Patch 6 of this series has been removed during the review process.
1905    
1906    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1907    Reviewed-by: Ran Benita <ran234@gmail.com>
1908    
1909    Message-ID: <545627AE.2040200@DemoRecorder.com>
1910    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1911    Patch-Set: ListInputDevices
1912    Patch-Number: libxcb 7/9
1913    Patch-Version: V1
1914
1915commit 51a0d57acc6ec0c9487d2dbc2dda806f05c49884
1916Author: Christian Linhart <chris@DemoRecorder.com>
1917Date:   Sun Nov 2 13:46:16 2014 +0100
1918
1919    generator: sumof with nested expression
1920    
1921    Support sumof with a nested expression.
1922    The nested expression is computed for every list-element
1923    and the result of the computation is added to the sum.
1924    
1925    This way, sumof can be applied to a list of structs,
1926    and, e.g., compute the sum of a specific field of that struct.
1927    
1928    example:
1929    <struct name="SumofTest_Element">
1930       <field type="CARD16" name="foo" />
1931       <field type="CARD16" name="bar" />
1932    </struct>
1933    
1934    <struct name="SumofTest_FieldAccess">
1935       <field type="CARD32" name="len" />
1936       <list type="SumofTest_Element" name="mylist1">
1937               <fieldref>len</fieldref>
1938       </list>
1939       <list type="CARD16" name="mylist2">
1940            <sumof ref="mylist1">
1941                    <fieldref>bar</fieldref>
1942            </sumof>
1943       </list>
1944    </struct>
1945    
1946    generated tmpvar:
1947        int xcb_pre_tmp_1; /* sumof length */
1948        int xcb_pre_tmp_2; /* sumof loop counter */
1949        int64_t xcb_pre_tmp_3; /* sumof sum */
1950        const xcb_input_sumof_test_element_t* xcb_pre_tmp_4; /* sumof list ptr */
1951    
1952    generated code:
1953        /* mylist2 */
1954        /* sumof start */
1955        xcb_pre_tmp_1 = _aux->len;
1956        xcb_pre_tmp_3 = 0;
1957        xcb_pre_tmp_4 = xcb_input_sumof_test_field_access_mylist_1(_aux);
1958        for ( xcb_pre_tmp_2 = 0; xcb_pre_tmp_2 < xcb_pre_tmp_1; xcb_pre_tmp_2++) {
1959            xcb_pre_tmp_3 += xcb_pre_tmp_4->bar;
1960            xcb_pre_tmp_4++;
1961        }
1962        /* sumof end. Result is in xcb_pre_tmp_3 */
1963        xcb_block_len += xcb_pre_tmp_3 * sizeof(uint16_t);
1964    
1965    changes for V2 of this patch:
1966    * explicitely set the member access operator in the prefix-tuple
1967      passed to function _c_helper_field_mapping.
1968      This enables us to simplify function "_c_helper_absolute_name"
1969      (which will be renamed "_c_helper_fieldaccess_expr" soon)
1970    
1971    V3: Changed style and formatting according to suggestions from Ran Benita
1972    
1973    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1974    Reviewed-by: Ran Benita <ran234@gmail.com>
1975    
1976    Message-ID: <54562798.8040500@DemoRecorder.com>
1977    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1978    Patch-Set: ListInputDevices
1979    Patch-Number: libxcb 5/9
1980    Patch-Version: V3
1981
1982commit 4a915c0dbadf326ea61349e29a0029d29f4bd339
1983Author: Christian Linhart <chris@DemoRecorder.com>
1984Date:   Sun Nov 2 13:45:40 2014 +0100
1985
1986    generator: sumof: support any type, generate explicit code
1987    
1988    A sumof-expression now generates explicit code ( for-loop etc )
1989    instead of calling xcb_sumof.
1990    
1991    This way, it supports any type which can be added.
1992    Previously, only uint_8 was supported.
1993    
1994    Here's an example and the generated code:
1995    
1996    xml:
1997    <struct name="SumofTest">
1998       <field type="CARD32" name="len" />
1999       <list type="CARD16" name="mylist1">
2000               <fieldref>len</fieldref>
2001       </list>
2002       <list type="CARD8" name="mylist2">
2003               <sumof ref="mylist1"/>
2004       </list>
2005    </struct>
2006    
2007    declaration of tempvars at the start of enclosing function:
2008        int xcb_pre_tmp_1; /* sumof length */
2009        int xcb_pre_tmp_2; /* sumof loop counter */
2010        int64_t xcb_pre_tmp_3; /* sumof sum */
2011        const uint16_t* xcb_pre_tmp_4; /* sumof list ptr */
2012    
2013    code:
2014        /* mylist2 */
2015        /* sumof start */
2016        xcb_pre_tmp_1 = _aux->len;
2017        xcb_pre_tmp_3 = 0;
2018        xcb_pre_tmp_4 = xcb_input_sumof_test_mylist_1(_aux);
2019        for ( xcb_pre_tmp_2 = 0; xcb_pre_tmp_2 < xcb_pre_tmp_1; xcb_pre_tmp_2++) {
2020            xcb_pre_tmp_3 += *xcb_pre_tmp_4;
2021            xcb_pre_tmp_4++;
2022        }
2023        /* sumof end. Result is in xcb_pre_tmp_3 */
2024        xcb_block_len += xcb_pre_tmp_3 * sizeof(uint8_t);
2025    
2026    This patch is also a preparation for sumof which can access
2027    fields of lists of struct, etc.
2028    
2029    V2: Changed style and formatting according to suggestions from Ran Benita
2030    
2031    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2032    Reviewed-by: Ran Benita <ran234@gmail.com>
2033    
2034    Message-ID: <54562774.8030306@DemoRecorder.com>
2035    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
2036    Patch-Set: ListInputDevices
2037    Patch-Number: libxcb 4/9
2038    Patch-Version: V2
2039
2040commit fda1fb4ed47a705744677a0074d83464af7aa4eb
2041Author: Christian Linhart <chris@DemoRecorder.com>
2042Date:   Sun Nov 2 13:45:29 2014 +0100
2043
2044    generator: expressions can generate pre-code
2045    
2046    This patch provides a mechanism for generating
2047    preparatory code for expressions.
2048    
2049    This is e.g. necessary when an expression needs computations
2050    which cannot be done in a C-Expression, like for-loops.
2051    
2052    This will be used for sumof expressions but may be useful
2053    elsewhere.
2054    
2055    Note: Patch 2 of this series has been removed during the review process.
2056    
2057    V2: adapt to changes in previous patches
2058    
2059    V3: some style and formatting changes according to suggestions from Ran Benita.
2060    
2061    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2062    Reviewed-by: Ran Benita <ran234@gmail.com>
2063    
2064    Message-ID: <54562769.3090405@DemoRecorder.com>
2065    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
2066    Patch-Set: ListInputDevices
2067    Patch-Number: libxcb 3/9
2068    Patch-Version: V3
2069
2070commit 265d38882cffce597367cc8bb2160b9e2482a80f
2071Author: Christian Linhart <chris@DemoRecorder.com>
2072Date:   Sun Nov 2 13:45:12 2014 +0100
2073
2074    generator: fix absname for fields with only accessor function
2075    
2076    Fix _c_helper_absolute_name for fields which cannot be accessed
2077    as a struct/union member but which can be accessed by an
2078    accessor function.
2079    
2080    The fix generates calls to the accessor function in these cases.
2081    
2082    Example:
2083    <struct name="AbsnameTest">
2084    <field type="CARD32" name="len" />
2085    <list type="CARD8" name="mylist1">
2086       <fieldref>len</fieldref>
2087    </list>
2088    <list type="CARD8" name="mylist2">
2089       <sumof ref="mylist1"/>
2090    </list>
2091    </struct>
2092    
2093    The sumof-expression ( <sumof ref="mylist1"/> ) refers to mylist1
2094    which is only acessible by an accessor function.
2095    
2096    Previously, sumof was only used inside bitcases,
2097    where such lists are accessible by members of the
2098    deserialized parent struct.
2099    (there is a difference between deserialization of switches
2100    and structs.)
2101    
2102    V2 of this patch:
2103    * replaced "!= None" with "is not None" because that's more pythonic.
2104    (according to suggestion from Ran Benita)
2105    
2106    V3 of this patch: simplification:
2107    * fixed the recursion in _c_type_setup
2108      so that _c_helper_absolute_name does not need check
2109      a gazillion things as a workaround anymore.
2110    
2111    * simplified _c_helper_absolute_name
2112      - remove unneeded check for empty string before
2113        append of last_sep to prefix_str
2114    
2115      - removed those if-conditions which are not
2116        needed anymore after fixing the recursion
2117        in _c_type_setup.
2118    
2119      - extract functionality for checking whether a field
2120        needs an accessor ( and which type of accessor )
2121        in functions.
2122        (also extracted from _c_accessors)
2123    
2124      - rearrange the condition branches and actions for
2125        more readability.
2126    
2127    V3 generates exactly the same *.c and *.h files as V2.
2128    
2129    V4 of this patch:
2130    * improve formatting as per suggestions of Ran
2131    
2132    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2133    Reviewed-by: Ran Benita <ran234@gmail.com>
2134    
2135    Message-ID: <54562758.5090107@DemoRecorder.com>
2136    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
2137    Patch-Set: ListInputDevices
2138    Patch-Number: libxcb 1/9
2139    Patch-Version: V4
2140
2141commit bbca7b82f803fa13fd30a2891ec06f2a213a28c2
2142Merge: 382d306 fdb291b
2143Author: Peter Harris <pharris@opentext.com>
2144Date:   Thu Oct 30 11:51:57 2014 -0400
2145
2146    Merge branch 'NestedStructTypenames-V5' of http://infra-srv1.demorecorder.com/git/free-sw/xcb/libxcb
2147
2148commit fdb291b414a7afc2c2326124b8ca11d6846cd7b9
2149Author: Christian Linhart <chris@demorecorder.com>
2150Date:   Wed Sep 3 10:10:49 2014 +0200
2151
2152    no typename for nested structs
2153    
2154    Nested structs which are generated for named case and bitcase
2155    do not get a typename anymore, i.e., they are anonymous structs.
2156    
2157    Reasons for this change:
2158    * Prior typenames have caused nameclashes
2159    * Prior typenames introduced names in the global namespace which
2160      did not start with the xcb prefix.
2161    
2162    This change is safe with respect to API compatibility because:
2163    I have searched for instances of named bitcases and there's only one place
2164    where they are used, and that's in xkb.xml: reply GetKbdByName.
2165    ( no need to search for <case> because it was introduced after the last release )
2166    
2167    The reply GetKbdByName is broken in its current form in the xkb.xml anyways,
2168    so it is most probably not used anywhere.
2169    
2170    So, my conclusion is that we can safely omit named types for nested structs.
2171    No need for an attribute.
2172    
2173    Message-ID: <1409731849-51897-1-git-send-email-chris@demorecorder.com>
2174    Patch-Thread-Subject: Re: [Xcb] names of nested structs of named bitcase/case are prone to nameclashes. Solution?
2175    Patch-Set: NestedStructTypenames
2176    Patch-Number: libxcb 1/1
2177    Patch-Version: V1
2178    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2179    Reviewed-By: Ran Benita <ran234@gmail.com>
2180
2181commit a7c75be5b1e2da32f7ce2c255972178e7d4b7598
2182Author: Christian Linhart <chris@DemoRecorder.com>
2183Date:   Tue Sep 9 23:26:40 2014 +0200
2184
2185    generator: fix align-pads for switches which start at unaligned pos
2186    
2187    Fix the alignment computation inside switches which start at
2188    an unaligned pos.
2189    This affects both explicit and implicit align pads.
2190    
2191    The alignment offset is derived from the lowest 3 bits of
2192    the pointer to the protocol-data at the start of the switch.
2193    This is sufficient for correcting all alignments up to 8-byte alignment.
2194    As far as I know there is no bigger alignment than 8-byte for the
2195    X-protocol.
2196    
2197    Example:
2198    struct InputState, where the switch starts after two 1-byte fields,
2199    which is a 2 byte offset for 4-byte and 8-byte alignment.
2200    
2201    The previous problem can be demonstrated when adding a
2202    <pad align="4"/> at the end of case "key".
2203    
2204    (Or when finding a testcase which reports the case "valuator" not
2205    at the last position of the QueryDeviceState-reply.
2206    I didn't find such a testcase, so I have used the pad align
2207    as described above.)
2208    
2209    V2: patch modified in order to fix bugs which I found when working on the
2210    next issue:
2211    * xcb_padding_offset has to be set 0 when xcb_block_len is set 0
2212    * xcb_padding_offset cannot be "const" therefore
2213    * for unpack and unserialize, the padding_offset must computed
2214      from _buffer instead of from the aux_var.
2215    
2216    V3: patch revised according to suggestion by Ran Benita:
2217    * only create and use xcb_padding_offset for switch
2218    
2219    Message-ID: <1410298000-24734-1-git-send-email-chris@demorecorder.com>
2220    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
2221    Patch-Set: QueryDeviceState
2222    Patch-Number: libxcb 4/4
2223    Patch-Version: V3
2224    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2225    Reviewed-By: Ran Benita <ran234@gmail.com>
2226
2227commit 277ea629def6728c9d826ff88e95b31c3e25915f
2228Author: Christian Linhart <chris@demorecorder.com>
2229Date:   Thu Aug 21 22:35:55 2014 +0200
2230
2231    generator: support lists of structs which contain a switch
2232    
2233    This essentially requires to have a correct sizeof-function
2234    for the struct.
2235    This in turn requires a sizeof-function for the switch, too.
2236    
2237    Making a sizeof-function for the switch is triggered by
2238    replacing "elif" by "if" in the first change of this patch.
2239    This way, c_need_sizeof is also set to True for switches if appropriate.
2240    
2241    The _c_serialize_helper_switch_field function has to support
2242    the context "sizeof":
2243    This is done in the second change of this patch
2244    
2245    The third change of this patch fixes an alignment error:
2246    It does not make sense to base the padding on the struct-type
2247    which is generated for switch because this struct does not
2248    represent the protocol. Rather it is the output of deserialization.
2249    ( The implicit padding for var-sized fields has other issues, IMHO,
2250    but I am not touching these now...)
2251    
2252    The effect on the generated code for the current xml-files
2253    is as follows:
2254    * several additional sizeof-functions are generated
2255    * the fix of the alignment error only changes one place
2256      in the XKB-extension for the GetKbdByName-reply.
2257      This is no problem because that reply in its current form
2258      is broken/unfinished anyways.
2259    
2260    Note:
2261    This patch also fixes a problem in the generator when
2262    a fixed-size list is the last field of a case or bitcase.
2263    
2264    Message-ID: <1408653356-21191-2-git-send-email-chris@demorecorder.com>
2265    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
2266    Patch-Set: QueryDeviceState
2267    Patch-Number: libxcb 2/3
2268    Patch-Version: V1
2269    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2270    Reviewed-By: Ran Benita <ran234@gmail.com>
2271
2272commit 382d306d6c44a9ece5551c210a932773b5cb94a5
2273Author: Emil Velikov <emil.l.velikov@gmail.com>
2274Date:   Fri Sep 5 01:46:40 2014 +0100
2275
2276    Move internal/private dependencies to Requires.private
2277    
2278    Program using the xcb sub-modules has indirect compile and runtime
2279    dependency of core xcb. To ensure this out we currently list xcb in
2280    the Requires field of the pkg-config files. While this provides all
2281    the required dependencies for successful compilation this causes
2282    over-linking and hides potential linking miss-use against the xcb modules.
2283    
2284    By moving to Requires.private we retain the compilation and runtime
2285    compatibility and avoids any runtime problems.
2286    
2287    Cc: Keith Packard <keithp@keithp.com>
2288    Cc: Alan Coopersmith <alan.coopersmith@oracle.com>
2289    References: http://people.freedesktop.org/~dbn/pkg-config-guide.html#faq
2290    References: https://wiki.mageia.org/en/Overlinking_issues_in_packaging
2291    References: http://err.no/personal/blog/2008/Mar/25
2292    Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
2293    Reviewed-by: Julien Cristau <jcristau@debian.org>
2294    Signed-off-by: Uli Schlachter <psychon@znc.in>
2295
2296commit d74d066949dbbbbbcb03bf7764e63f4347f99974
2297Author: Christian Linhart <chris@DemoRecorder.com>
2298Date:   Sat Sep 6 20:06:15 2014 +0200
2299
2300    generator: support fixed size lists in var-sized structs
2301    
2302    V2: patch revised according to suggestions from Ran Benita:
2303    * removed blanks before an after parentheses of function-calls or tuples
2304    * replaced if by elif in "if field.type.is_list". ( this fixes old code )
2305    
2306    Message-ID: <540B4D17.1080908@DemoRecorder.com>
2307    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
2308    Patch-Set: QueryDeviceState
2309    Patch-Number: libxcb 1/3
2310    Patch-Version: V2
2311    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
2312    Reviewed-By: Ran Benita <ran234@gmail.com>
2313
2314commit b0e6c2de09c7474868dd7185674fa113a5c2e0aa
2315Author: Uli Schlachter <psychon@znc.in>
2316Date:   Mon Aug 18 10:38:48 2014 +0200
2317
2318    xcb_get_setup(): Never return NULL
2319    
2320    The documentation doesn't mention it and it's unlikely that a lot of code out
2321    there handles this case correctly. So, instead of returning NULL, let
2322    xcb_get_setup() return a pointer to a static, invalid, all-zero setup
2323    information structure.
2324    
2325    Signed-off-by: Uli Schlachter <psychon@znc.in>
2326
2327commit c4e40f646b8da4fd112ea54a612c880be5e942a8
2328Author: Uli Schlachter <psychon@znc.in>
2329Date:   Mon Aug 18 10:38:41 2014 +0200
2330
2331    Make some functions also work on error connections
2332    
2333    There is no technical reason why xcb_get_setup() and xcb_get_file_descriptor()
2334    shouldn't work on non-static error connections. They cannot be used for many
2335    useful things, but at least they work.
2336    
2337    This works around bugs in lots of programs out there which assume that
2338    xcb_get_setup() does not return NULL and which just happily dereference the
2339    results. Since xcb_connect() never returns NULL, it's a bit weird that
2340    xcb_get_setup() can do so. xcb_get_file_descriptor() is just modified since this
2341    can be done here equally easily and because the fd isn't closed until the final
2342    xcb_disconnect() on the error connection.
2343    
2344    Non-static error connections are connections which entered an error state after
2345    xcb_connect() succeeded. If something goes wrong in establishing a connection,
2346    xcb_connect() will return a static error connection which doesn't have the
2347    fields used here.
2348    
2349    Signed-off-by: Uli Schlachter <psychon@znc.in>
2350
2351commit 355d4d6ab9f5c12c2ee4a91e8cf6eb4a2854d73c
2352Author: Christian Linhart <chris@demorecorder.com>
2353Date:   Tue Aug 19 15:57:34 2014 +0200
2354
2355    support switch case in the generator
2356    
2357    The implementation is rather simple:
2358    When a <case> is used instead of a <bitcase>
2359    then operator "==" is used instead of "&" in the if-condition.
2360    
2361    So it creates a series of "if" statements
2362    (instead of a switch-case statement in C )
2363    
2364    In practice this does not matter because a good
2365    optimizing compiler will create the same code
2366    as for a switch-case.
2367    
2368    With this simple implementation we get additional
2369    flexibility in the following forms:
2370    * a case value may appear in multiple case branches.
2371      for example:
2372            case C1 will be selected by values 1, 4, or 5
2373            case C2 will be selected by values 3, 4, or 7
2374    
2375    * mixing of bitcase and case is possible
2376            (this will usually make no sense but there may
2377            be protocol specs where this is needed)
2378    
2379    details of the impl:
2380    * replaced "is_bitcase" with "is_case_or_bitcase" in all places
2381      so that cases are treated like bitcases.
2382    
2383    * In function "_c_serialize_helper_switch": write operator "=="
2384      instead of operator "&" if it is a case.
2385
2386commit 966fba6ba4838949d0727dfa45eeb9392d1f85d9
2387Author: Daniel Martin <consume.noise@gmail.com>
2388Date:   Tue Jul 29 22:48:44 2014 +0200
2389
2390    Disable Xevie and Xprint by default
2391    
2392    Both extensions have been dropped from the X-Server in 2008:
2393        http://cgit.freedesktop.org/xorg/xserver/commit/?id=1c8bd31
2394        http://cgit.freedesktop.org/xorg/xserver/commit/?id=f4036f6
2395    
2396    Don't build them by default.
2397    
2398    Reviewed-by: Julien Cristau <jcristau@debian.org>
2399    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
2400    Signed-off-by: Uli Schlachter <psychon@znc.in>
2401
2402commit a5e90ae6a1543a681d95b831dc5c44e9c6e78610
2403Author: Gaetan Nadon <memsize@videotron.ca>
2404Date:   Wed Mar 26 15:40:57 2014 -0400
2405
2406    help text: do not report the insanly long list of Warning flags.
2407    
2408    Originally there was just one. Now that XCB has been integrated with X and
2409    uses the same compiler flags, it is a different story.
2410    
2411      Used CFLAGS:
2412        CPPFLAGS............:
2413        CFLAGS..............: -g -O2
2414        Warning CFLAGS......:  -Wall -Wpointer-arith AND SO ON FOR 8 lines...
2415    
2416    It completely defaces the otherwise excellent output.
2417    
2418    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2419    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2420    Signed-off-by: Uli Schlachter <psychon@znc.in>
2421
2422commit 529e3bfc2088dcd35e68a0a8394d0e9bff7c978b
2423Author: Gaetan Nadon <memsize@videotron.ca>
2424Date:   Wed Mar 26 15:40:56 2014 -0400
2425
2426    Add ChangeLog and INSTALL using xorg macros
2427    
2428    Same as all other X modules. The one in libxcb git is removed.
2429    
2430    Those files are created during 'make dist'
2431    
2432    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2433    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2434    Signed-off-by: Uli Schlachter <psychon@znc.in>
2435
2436commit 74f552c1b39eb6a1ab64451477a492a2b22d9d7a
2437Author: Gaetan Nadon <memsize@videotron.ca>
2438Date:   Wed Mar 26 15:24:46 2014 -0400
2439
2440    sendmsg: remove --enable-sendfds as it is superceeded by --enable-dri3
2441    
2442    DRI3 requires sendmsg support which is auto-detected. A builder can enable
2443    or disable dri3 feature. If sendmsg function is not available, dri3 cannot
2444    be enabled.
2445    
2446    This reverts af8067cbf4856 which was done at a time where --enable-dri3
2447    had not been added yet.
2448    
2449    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2450    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2451    Signed-off-by: Uli Schlachter <psychon@znc.in>
2452
2453commit 23f57ac8bf0812bbc851ec9a815e50a640b97db5
2454Author: Gaetan Nadon <memsize@videotron.ca>
2455Date:   Wed Mar 26 15:24:45 2014 -0400
2456
2457    config: issue an error if DRI3 is requested, but sendfds is not available
2458    
2459    When a user issues the --enable-dri3 option and sendfds is not available
2460    on the system, the configuration will abort with an error message.
2461    
2462    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2463    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2464    Signed-off-by: Uli Schlachter <psychon@znc.in>
2465
2466commit 7f07b57be587a2ebe0cadceba3fe67ed4a1e79db
2467Author: Gaetan Nadon <memsize@videotron.ca>
2468Date:   Wed Mar 26 15:24:44 2014 -0400
2469
2470    config: default option for enable-dri3 is not implemented
2471    
2472    The first symptom is the help text:
2473    
2474      --enable-dri3     Build XCB DRI3 Extension (default: "$sendfds")
2475    
2476    The implementation variable $sendfds leaked into the user interface.
2477    Testing the various user inputs:
2478      <nothing>                DRI3 is enabled     PASS
2479      --enable-dri3            DRI3 is enabled     PASS
2480      --enable-dri3=yes        DRI3 is enabled     PASS
2481      --enable-dri3=no         DRI3 is disabled    PASS
2482      --disable-dri3           DRI3 is disabled    PASS
2483      --enable-dri3=$sendfds   DRI3 is disabled    FAIL
2484    
2485    This patch implements the usual idiom for features that are enabled by
2486    default if the various conditions are met (sendfds is available).
2487    New help text:
2488    
2489      --enable-dri3     Build XCB DRI3 Extension (default: auto)
2490    
2491    With the additional user input:
2492    
2493      --enable-dri3=auto
2494    
2495    which is equivalent to providing no input at all.
2496    
2497    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2498    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2499    Signed-off-by: Uli Schlachter <psychon@znc.in>
2500
2501commit d1e8ec96fca4862f37ec9f0e9407bb989c4c161a
2502Author: Uli Schlachter <psychon@znc.in>
2503Date:   Fri Aug 1 15:56:52 2014 +0200
2504
2505    Release libxcb 1.11
2506
2507commit c5c6cfa4d2a27d0f9e139a1856d866433b6046bc
2508Author: Uli Schlachter <psychon@znc.in>
2509Date:   Fri Aug 1 16:03:24 2014 +0200
2510
2511    Bump xcb-proto requirement to 1.11
2512    
2513    This is needed for the new direct_imports field that we need from xcb-proto.
2514    
2515    Signed-off-by: Uli Schlachter <psychon@znc.in>
2516
2517commit 70ea5da64b34336bb0916f6c325545cb50746159
2518Author: Alexander Mezin <mezin.alexander@gmail.com>
2519Date:   Sun Jun 29 17:33:48 2014 +0700
2520
2521    xcb.h: add 'struct' before xcb_setup_t, xcb_query_extension_reply_t
2522    
2523    These structs are typedef'ed in xproto.h, so in xcb.h these types
2524    (without 'struct') are actually undefined.
2525    
2526    GCC reports this as error when building precompiled header.
2527    
2528    Signed-off-by: Alexander Mezin <mezin.alexander@gmail.com>
2529    Reviewed-by: Peter Harris <pharris@opentext.com>
2530
2531commit 7e6af51b4e984f661fe4f21596cab5cb8ee15ea0
2532Author: Ran Benita <ran234@gmail.com>
2533Date:   Tue Feb 25 14:11:35 2014 +0200
2534
2535    c_client.py: remove more trailing space from generated files
2536    
2537    Signed-off-by: Ran Benita <ran234@gmail.com>
2538    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2539    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2540
2541commit 8221d249b77131b338e3b35ce2229193f129e514
2542Author: Ran Benita <ran234@gmail.com>
2543Date:   Sun Feb 23 22:55:21 2014 +0200
2544
2545    c_client.py: remove trailing whitespace from generated files
2546    
2547    Signed-off-by: Ran Benita <ran234@gmail.com>
2548    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2549
2550commit e3c728ee3d9a2fd7478d5f57830c3483b774a16e
2551Author: Ran Benita <ran234@gmail.com>
2552Date:   Sun Feb 23 22:55:20 2014 +0200
2553
2554    c_client.py: remove useless generated comments
2555    
2556    They are bloated, don't add anything over the signature, in some cases
2557    duplicate the doxygen comments, and are not integrated with the <doc>
2558    tags in any way. Remove them and cut the generated LOC by half.
2559    
2560    Signed-off-by: Ran Benita <ran234@gmail.com>
2561    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2562
2563commit cae2e398563841c5b814596fd1f1c64354dcac71
2564Author: Ran Benita <ran234@gmail.com>
2565Date:   Sun Feb 23 22:55:19 2014 +0200
2566
2567    c_client.py: make the man page output deterministic
2568    
2569    Some parts of the man pages (SEE ALSO and ERRORS) are generated by
2570    iterating a Python dict. But the iteration order in a dict is random,
2571    so each build the output is ordered differently. Avoid that by iterating
2572    in sorted order.
2573    
2574    Signed-off-by: Ran Benita <ran234@gmail.com>
2575    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2576
2577commit bfbf83b1d8113ac398b57c2738706792946d1c03
2578Author: Ran Benita <ran234@gmail.com>
2579Date:   Sun Feb 23 22:55:18 2014 +0200
2580
2581    c_client.py: prefix all monkey-patched fields with c_
2582    
2583    The script adds many fields to the objects coming from xcbgen. To
2584    distinguish them, a c_ prefix is used, but for some it was missing.
2585    
2586    Signed-off-by: Ran Benita <ran234@gmail.com>
2587    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2588
2589commit 285d566a5c11c6edd3665beb42312e24bde77d16
2590Author: Ran Benita <ran234@gmail.com>
2591Date:   Sun Feb 23 22:55:17 2014 +0200
2592
2593    c_client.py: remove trailing whitespace
2594    
2595    These are extra annoying in python code.
2596    
2597    Signed-off-by: Ran Benita <ran234@gmail.com>
2598    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2599
2600commit 285838cfe41c212a2453d903497757d2392fd0ce
2601Author: Ran Benita <ran234@gmail.com>
2602Date:   Sun Feb 23 22:55:16 2014 +0200
2603
2604    c_client.py: remove useless 'today' variable
2605    
2606    Signed-off-by: Ran Benita <ran234@gmail.com>
2607    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2608
2609commit 49a61c8b459ab19c7f39e653bbb0d0339ea8f00f
2610Author: Alan Coopersmith <alan.coopersmith@oracle.com>
2611Date:   Fri Jul 11 20:41:15 2014 -0700
2612
2613    Fix typos & awkward wording in tutorial
2614    
2615    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2616
2617commit 125135452a554e89e49448e2c1ee6658324e1095
2618Author: Michael Haubenwallner <michael.haubenwallner@salomon.at>
2619Date:   Fri Jun 13 16:18:34 2014 +0200
2620
2621    bug#79986: include system headers early
2622    
2623    AIX <sys/poll.h> does redefine 'events' to 'reqevents' eventually.
2624    To not have this cause compilation errors, need to include the local
2625    header files after any system header file.
2626    
2627    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2628    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2629
2630commit bc5a1047548b578624cfbc44ca192cde7664ed78
2631Author: Alan Coopersmith <alan.coopersmith@oracle.com>
2632Date:   Fri Jun 13 21:26:21 2014 -0700
2633
2634    Document failure modes of xcb_connect*() functions
2635    
2636    Documentation was previously unclear that these always return a non-NULL
2637    pointer, and that callers need to check it for error values, instead of
2638    checking for a NULL return value.
2639    
2640    Triggered by having to dig through code to answer a user's question on
2641    the #xcb irc channel, since neither of us found it covered in the docs.
2642    
2643    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2644    Reviewed-by: Uli Schlachter <psychon@znc.in>
2645
2646commit 72e45969ff71204cee2dde3502841736cfd41c8a
2647Author: Daniel Martin <consume.noise@gmail.com>
2648Date:   Mon Jun 9 17:55:04 2014 +0200
2649
2650    Handle <pad align="n" /> between lists
2651    
2652    Without this patch we end up with invalid C code if we've a
2653    <pad align="n" /> between two variadic lists. Check for such a condition
2654    and take the alignment pad into account.
2655    
2656    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=79808
2657    
2658    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
2659    Signed-off-by: Peter Harris <pharris@opentext.com>
2660
2661commit d978a4f69b30b630f28d07f1003cf290284d24d8
2662Author: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2663Date:   Sat Apr 5 20:54:59 2014 -0700
2664
2665    xcb_open: Improve abstraction for launchd secure sockets
2666    
2667    This changes away from hard-coding the /tmp/launch-* path to now
2668    supporting a generic <path to unix socket>[.<screen>] format for
2669    $DISPLAY.
2670    
2671    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2672
2673commit 29e419c5840a1eeda3336a0802686ee723dcaab3
2674Author: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2675Date:   Sat Apr 5 21:09:42 2014 -0700
2676
2677    xcb_open: Minor code cleanup for better readability
2678    
2679    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2680    Reviewed-by: Uli Schlachter <psychon@znc.in>
2681
2682commit be0fe56c3bcad5124dcc6c47a2fad01acd16f71a
2683Author: Keith Packard <keithp@keithp.com>
2684Date:   Mon Dec 23 21:15:20 2013 -0800
2685
2686    Ensure xcb owns socket and no other threads are writing before send_request
2687    
2688    send_request may only write to out.queue if no other thread is busy
2689    writing to the network (as that thread may be writing from out.queue).
2690    
2691    send_request may only allocate request sequence numbers if XCB owns
2692    the socket.
2693    
2694    Therefore, send_request must make sure that both conditions are true
2695    when it holds iolock, which can only be done by looping until both
2696    conditions are true without having dropped the lock waiting for the
2697    second condition.
2698    
2699    We choose to get the socket back from Xlib first as get_socket_back
2700    has a complicated test and checking for other threads writing is a
2701    simple in-lined check.
2702    
2703    This also changes the sequence number checks (64k requests with no
2704    reply, 4M request wrapping) to ensure that both conditions are true
2705    before queueing the request.
2706    
2707    Signed-off-by: Keith Packard <keithp@keithp.com>
2708    Reviewed-by: Uli Schlachter <psychon@znc.in>
2709
2710commit e2813e1cde893f384fa620ff3c13493beebabe0c
2711Author: Keith Packard <keithp@keithp.com>
2712Date:   Wed Feb 12 14:15:46 2014 -0800
2713
2714    Update .pc file Requires lines to express full dependencies
2715    
2716    Some xcb libraries depend on others; make these dependencies explicit
2717    in the .pc files that are installed.
2718    
2719    This change was generated automatically by running 'check-pc-requires -fix'
2720    
2721    Signed-off-by: Keith Packard <keithp@keithp.com>
2722    Signed-off-by: Uli Schlachter <psychon@znc.in>
2723
2724commit 32de4c4213a49c61127c6957ea05fef3e5355291
2725Author: Keith Packard <keithp@keithp.com>
2726Date:   Wed Feb 12 14:15:45 2014 -0800
2727
2728    Validate .pc file Requires lines
2729    
2730    This walks through the .pc.in files and makes sure all of the Requires
2731    lines express sufficient dependency information.
2732    
2733    Signed-off-by: Keith Packard <keithp@keithp.com>
2734    Signed-off-by: Uli Schlachter <psychon@znc.in>
2735
2736commit 1f6cd9f1fcd3c07d323b678292c9cb00ae1f7504
2737Author: Keith Packard <keithp@keithp.com>
2738Date:   Wed Feb 12 14:15:44 2014 -0800
2739
2740    Only #include directly referenced module header files
2741    
2742    This avoids having the nested header files also included at the top
2743    level, which is more efficient.
2744    
2745    Signed-off-by: Keith Packard <keithp@keithp.com>
2746    Signed-off-by: Uli Schlachter <psychon@znc.in>
2747
2748commit cb686b576739deea00180c54697c8b62b8419ae0
2749Author: Uli Schlachter <psychon@znc.in>
2750Date:   Tue Feb 25 15:50:50 2014 +0100
2751
2752    Add doxygen documentation to functions in xcbext.h
2753    
2754    Signed-off-by: Uli Schlachter <psychon@znc.in>
2755    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2756
2757commit 2fb14e5883f2ea2f01d248674cfcc26ccb704753
2758Author: Uli Schlachter <psychon@znc.in>
2759Date:   Tue Dec 31 15:18:01 2013 +0100
2760
2761    Make xcb_disconnect(NULL) safe
2762    
2763    Code can be simplified if the deallocation functions can always be called in
2764    cleanup code. So if you have some code that does several things that can go
2765    wrong, one of which is xcb_connect(), after this change, the xcb_connection_t*
2766    variable can be initialized to NULL and xcb_disconnect() can always be called on
2767    the connection object.
2768    
2769    References: http://lists.freedesktop.org/archives/xcb/2013-September/008659.html
2770    
2771    Signed-off-by: Uli Schlachter <psychon@znc.in>
2772    Reviewed-by: Julien Cristau <jcristau@debian.org>
2773
2774commit 4dcbfd77b78ca6b016ce815af26235501f6cd75a
2775Author: Uli Schlachter <psychon@znc.in>
2776Date:   Tue Dec 31 15:05:36 2013 +0100
2777
2778    xcb_disconnect(): Fix leak with error connections
2779    
2780    There are two kind of error connections in XCB. First, if something goes wrong
2781    while the connection is being set up, _xcb_conn_ret_error() is used to return a
2782    static connection in an error state. If something goes wrong later,
2783    _xcb_conn_shutdown() is used to set c->has_error.
2784    
2785    This is important, because the static object that _xcb_conn_ret_error() returns
2786    must not be freed, while the dynamically allocated objects that go through
2787    _xcb_conn_shutdown() must obviously be properly deallocated.
2788    
2789    This used to work correctly, but in 769acff0da8, xcb_disconnect() was made to
2790    ignore all connections in an error state completely. Fix this by only ignoring
2791    the few static error connections that we have.
2792    
2793    This was tested with the following hack:
2794    
2795        xcb_connection_t *c = xcb_connect(NULL, NULL);
2796        close(xcb_get_file_descriptor(c));
2797        xcb_discard_reply(c, xcb_get_input_focus(c).sequence);
2798        xcb_flush(c);
2799        xcb_disconnect(c);
2800    
2801    Valgrind confirms that xcb has a memory leak before this patch that this patch
2802    indeed fixes.
2803    
2804    Signed-off-by: Uli Schlachter <psychon@znc.in>
2805    Reviewed-by: Julien Cristau <jcristau@debian.org>
2806
2807commit d84dd752ef571491b015443fefedca53c7f81282
2808Author: Uli Schlachter <psychon@znc.in>
2809Date:   Mon Sep 9 13:04:11 2013 +0200
2810
2811    Remove tabs and trailing whitespaces
2812    
2813    Signed-off-by: Uli Schlachter <psychon@znc.in>
2814
2815commit 4ffa6f83b92763eb901c7ddb7c20775e24d507ca
2816Author: Ran Benita <ran234@gmail.com>
2817Date:   Sat Jan 18 17:10:53 2014 +0200
2818
2819    Add comments about how _xcb_conn_ret_error() works
2820    
2821    If xcb_connect() fails, it doesn't return NULL. Instead, it always
2822    returns an xcb_connection_t*, and the user should check for errors with
2823    the xcb_connection_has_error() function. What this function does is
2824    check if conn->has_error contains a non-zero error code, and returns it.
2825    
2826    If an error did occur, xcb doesn't actually return a full
2827    xcb_connection_t though, it just returns (xcb_connection_t *)
2828    error_code. Since the 'has_error' field is the first, it is still
2829    possible to check conn->has_error.
2830    
2831    That last trick was not immediately obvious to me, so add some guiding
2832    comments. This also ensures no one obliviously rearranges the struct.
2833    
2834    Signed-off-by: Ran Benita <ran234@gmail.com>
2835    Signed-off-by: Uli Schlachter <psychon@znc.in>
2836
2837commit d7eb0bdf3b5b11ee9f40ee5e73df8fc0bdfa59f3
2838Author: Gaetan Nadon <memsize@videotron.ca>
2839Date:   Tue Jan 7 14:02:21 2014 -0500
2840
2841    generated man pages: use xorg footer and no hard coded extension
2842    
2843    The section number is no longer hard-coded
2844    The left footer is now "X Version 11".
2845    The center footer is the package name with the version, "libxcb 1.9"
2846    The three values above are provided through xorg-macros. They are passed-in
2847    to the python c_client code.
2848    
2849    Example of footer (last line, above dotted line)
2850    
2851    [...]
2852    AUTHOR
2853           Generated from xproto.xml. Contact xcb@lists.freedesktop.org for cor‐
2854           rections and improvements.
2855    
2856    X Version 11                      libxcb 1.9                 xcb_send_event(3)
2857    
2858    ------------------------------------------------------------------------------
2859    
2860    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2861    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2862
2863commit e4061b8f00d301a51d4c9eef4a07f1e87592fe85
2864Author: Gaetan Nadon <memsize@videotron.ca>
2865Date:   Tue Jan 7 14:02:16 2014 -0500
2866
2867    generated man pages: build without hard coded extension
2868    
2869    The automake MAN primary requires a hard coded extension to build
2870    man pages. Let's avoid that as the extension number may vary by platform.
2871    Take advantage of the fact that the man directory only contains man pages.
2872    Wildcards are not supported by Automake but it happens to work
2873    sufficiently well here.
2874    
2875    Normally xorg build man pages by converting a source .man file to a
2876    target file with the extension number. That would be too many files
2877    in this case.
2878    
2879    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2880    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2881
2882commit 3cdd524cadc4352ebd9e17b1f73134bec1838b40
2883Author: Gaetan Nadon <memsize@videotron.ca>
2884Date:   Tue Jan 7 14:02:05 2014 -0500
2885
2886    man: build static man pages using xorg patterns
2887    
2888    The section number is no longer hard-coded, supplied by xorg-macros.
2889    The left footer is now "X Version 11".
2890    The center footer is the package name with the version, "libxcb 1.9"
2891    The man directory is a sibbling to the doc directory. One can build
2892    or clean the man pages without disturbing the library code.
2893    
2894    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2895    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2896
2897commit c056adcd92daa06f4825d5c85a40e140a3e85b42
2898Author: Gaetan Nadon <memsize@videotron.ca>
2899Date:   Thu Jan 9 14:32:18 2014 -0500
2900
2901    autoconf: replace all tabs with spaces
2902    
2903    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2904    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2905
2906commit 412928f113c8a5e5b30c03a294a42b0b1cf5f5d7
2907Author: Gaetan Nadon <memsize@videotron.ca>
2908Date:   Tue Jan 7 14:00:44 2014 -0500
2909
2910    autoconf: use default xorg configuration for doxygen documentation
2911    
2912    No content or form changes for the xcb manual or tutorial.
2913    Only the configuration user visible bits change.
2914    
2915    Xcb will now have the same configuration options as the 30 other
2916    xorg modules.
2917    
2918    Xorg classifies documentation as "user", "developer" or "specifications".
2919    The xcb manual falls under the "developer" category. Developers docs
2920    are never installed under $prefix.
2921    
2922    A builder can selectively turn on/off any or all of the categories. He can
2923    also selectively turn on/off any of the many tools used to generate
2924    documentation such as doxygen, xmlto, etc... Each tool has an environment
2925    variable defined such as DOXYGEN.
2926    
2927    Other features are available, the user interface and the functionality
2928    is the same on all modules.
2929    
2930    --with-doxygen=FILE is replaced with DOXYGEN env variable
2931    --disable-build-docs is replaced with --disable-devel-docs
2932    
2933    The new interface displayed with ./configure --help:
2934    
2935      --enable-devel-docs     Enable building the developer documentation
2936                              (default: yes)
2937      --with-doxygen          Use doxygen to regenerate documentation (default:
2938                              auto)
2939      DOXYGEN     Path to doxygen command
2940      DOT         Path to the dot graphics utility
2941    
2942    The dot tool checking has been added to util-macros in version 1.18.
2943    
2944    Refer to the table of existing docs in xorg.
2945    XCB will be added for the doxygen generated API manual.
2946    
2947    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2948    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2949
2950commit 9ba6aa759e937e60b231b209b1293a40ad6c7a8a
2951Author: Gaetan Nadon <memsize@videotron.ca>
2952Date:   Tue Jan 7 14:00:43 2014 -0500
2953
2954    autoconf: fix warning by replacing deprecated AC_HELP_STRING
2955    
2956    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2957    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2958
2959commit 942eabaae3600e7277aa09a179ef10c9a06de62f
2960Author: Gaetan Nadon <memsize@videotron.ca>
2961Date:   Tue Jan 7 14:00:42 2014 -0500
2962
2963    autoconf: require libtool minimum level 2.2
2964    
2965    This is the updated minimum level as referenced in:
2966    http://www.x.org/wiki/Building_the_X_Window_System/#index2h3
2967    
2968    Libtool version 2 has been used for several years now. There should be
2969    no surprises.
2970    
2971    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2972    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2973
2974commit 517cb0e888eca4030b4bd00c316619db925032ea
2975Author: Gaetan Nadon <memsize@videotron.ca>
2976Date:   Tue Jan 7 14:00:41 2014 -0500
2977
2978    autoconf: comment and layout the initialization section
2979    
2980    No functional changes. Trying to make it clearer.
2981    
2982    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2983    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2984    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2985
2986commit 0a17b61a8971d3c4600a5c19fa7d753e7196abfb
2987Author: Gaetan Nadon <memsize@videotron.ca>
2988Date:   Tue Jan 7 14:00:40 2014 -0500
2989
2990    autoconf: AC_INIT: add bug url
2991    
2992    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2993    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2994
2995commit 414b1037c5a9770270022e62aa66fe7a4021a96b
2996Author: Gaetan Nadon <memsize@videotron.ca>
2997Date:   Tue Jan 7 14:00:39 2014 -0500
2998
2999    autoconf: use the warning variables from xorg
3000    
3001    The BASE_CFLAGS variable contains only warnings, just like the XCB
3002    version of CWARNFLAGS. This will result in no changes in the binaries
3003    produced. Xorg was missing -fd for SUNCC so it has been added to util-macros
3004    v 1.18.
3005    
3006    Do not get confused with the xorg deprecated CWARNFLAGS variable which
3007    contains an option that is not a warning, -fno-strict-aliasing. This
3008    option, should it be needed, can be added using the XORG_TESTSET_CFLAG
3009    macro.
3010    
3011    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3012    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
3013
3014commit c4f2c70bc37a592406b7693562c7513f2f99b34d
3015Author: Gaetan Nadon <memsize@videotron.ca>
3016Date:   Tue Jan 7 14:00:38 2014 -0500
3017
3018    autoconf: use XORG_DEFAULT_OPTIONS
3019    
3020    XCB has been part of X.Org for a while now. This patch will harmonize the XCB
3021    configuration, using xorg-macros series of macros. It is already used in the
3022    XCB utils packages and is needed to build xcb-proto.
3023    
3024    The XORG_DEFAULT_OPTIONS already includes the statement for the silent
3025    rules.
3026    
3027    The AC_PROG_CC statement is removed so as not to override AC_PROG_CC_C99
3028    in XORG_DEFAULT_OPTIONS. The effective change is that xcb now uses c99 as
3029    requested.
3030    
3031    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3032    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
3033
3034commit e7263931aff3e3450dc938ad465a7577f943549f
3035Author: Peter Harris <pharris@opentext.com>
3036Date:   Tue Jan 14 14:50:55 2014 -0500
3037
3038    Support <pad align="n" />
3039    
3040    Reviewed-By: Ran Benita <ran234@gmail.com>
3041    Signed-off-by: Peter Harris <pharris@opentext.com>
3042
3043commit 3b72a2c9d1d656c74c691a45689e1d637f669e3a
3044Author: Kenneth Graunke <kenneth@whitecape.org>
3045Date:   Fri Jan 3 15:08:33 2014 -0800
3046
3047    Force XCB event structures with 64-bit extended fields to be packed.
3048    
3049    With the advent of the Present extension, some events (such as
3050    PresentCompleteNotify) now use native 64-bit types on the wire.
3051    
3052    For XGE events, we insert an extra "uint32_t full_sequence" field
3053    immediately after the first 32 bytes of data.  Normally, this causes
3054    the subsequent fields to be shifted over by 4 bytes, and the structure
3055    to grow in size by 4 bytes.  Everything works fine.
3056    
3057    However, if event contains 64-bit extended fields, this may result in
3058    the compiler adding an extra 4 bytes of padding so that those fields
3059    remain aligned on 64-bit boundaries.  This causes the structure to grow
3060    by 8 bytes, not 4.  Unfortunately, XCB doesn't realize this, and
3061    always believes that the length only increased by 4.  read_packet()
3062    then fails to malloc enough memory to hold the event, and the event
3063    processing code uses the wrong offsets.
3064    
3065    To fix this, mark any event structures containing 64-bit extended
3066    fields with __attribute__((__packed__)).
3067    
3068    v2: Use any(...) instead of True in (...), as suggested by
3069        Daniel Martin.
3070    
3071    v3 (Alan Coopersmith): Fix build with Solaris Studio 12.3 by moving the
3072    attribute to after the structure definition.
3073    
3074    Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
3075    Reviewed-by: Keith Packard <keithp@keithp.com> [v1]
3076    Reviewed-by: Josh Triplett <josh@joshtriplett.org> [v1]
3077    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
3078    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3079
3080commit 010872f611a044ced4e96b18a7514796b2a443df
3081Author: Uli Schlachter <psychon@znc.in>
3082Date:   Sun Dec 22 15:59:24 2013 +0100
3083
3084    Release libxcb 1.10
3085    
3086    Signed-off-by: Uli Schlachter <psychon@znc.in>
3087
3088commit f653464554469b5767f1c99abced25a76bace047
3089Author: Julien Cristau <jcristau@debian.org>
3090Date:   Sat Dec 14 06:16:37 2013 +0100
3091
3092    Add NEWS for 1.10
3093    
3094    Signed-off-by: Julien Cristau <jcristau@debian.org>
3095
3096commit 9c2a6dc20c64ce93e0acd2fceec6d3cab8fc9134
3097Author: Uli Schlachter <psychon@znc.in>
3098Date:   Fri Nov 15 22:33:12 2013 +0100
3099
3100    Add NEWS entries for releases 1.9.1 to 1.9.3
3101    
3102    libxcb 1.9.1 was released from a branch and thus its NEWS entries never made it
3103    into the master branch. The other releases didn't update NEWS.
3104    
3105    Signed-off-by: Uli Schlachter <psychon@znc.in>
3106    Signed-off-by: Julien Cristau <jcristau@debian.org>
3107
3108commit ea4406a8e0193cad8effe569e7835f8b67894e54
3109Author: Julien Cristau <jcristau@debian.org>
3110Date:   Sat Dec 14 05:54:20 2013 +0100
3111
3112    Bump xcb-proto requirement to 1.10
3113    
3114    Makes sure we generate the new generic event struct.
3115    
3116    Signed-off-by: Julien Cristau <jcristau@debian.org>
3117
3118commit a1299eb2a210b5788a2b827b82a3d825caa1f201
3119Author: PHO <pho@cielonegro.org>
3120Date:   Tue Dec 3 12:43:04 2013 +0900
3121
3122    Test the value of msg_controllen for platforms whose CMSG_FIRSTHDR() does not test it for us
3123    
3124    As RFC 2292 points out, some platforms (e.g. Darwin 9.8.0) provide
3125    CMSG_FIRSTHDR(msg) which just returns msg.msg_control without first
3126    checking if msg.msg_controllen is non-zero. We need a workaround for
3127    such platforms not to let _xcb_in_read() segfault.
3128    
3129    https://bugs.freedesktop.org/show_bug.cgi?id=72253
3130    
3131    Signed-off-by: Julien Cristau <jcristau@debian.org>
3132
3133commit b30b11ac49d934541312b03c41d1ab83047a59f4
3134Author: Uli Schlachter <psychon@znc.in>
3135Date:   Mon Nov 18 20:28:08 2013 +0100
3136
3137    Increment the "current" version info for sync, xinput and xkb
3138    
3139    Sync: Due to commit e6a246e50e62cbcba3 "sync: Change value list param of
3140    CreateAlarm and ChangeAlarm into switch", various symbols disappeared,
3141    for example xcb_sync_{change,create}_alarm_sizeof.
3142    
3143    xinput: This extension was updated from version 1.4 to 2.3. This means
3144    that lots of new things are generated. However, this change is
3145    backwards-compatible and thus age gets set to 1.
3146    
3147    xkb: In commit 37d0f55392d6 "xkb: Work around alignment problems in
3148    GetNames and GetMap replies", some padding fields were introduced into
3149    structures for which an _unpack() function is generated. This changed
3150    the size of the struct and caused offsets into this struct to change.
3151    
3152    https://bugs.freedesktop.org/show_bug.cgi?id=71507
3153    
3154    Signed-off-by: Uli Schlachter <psychon@znc.in>
3155    Signed-off-by: Julien Cristau <jcristau@debian.org>
3156
3157commit ce5395eb4611341ba7c243ed524d023a616f73bb
3158Author: Uli Schlachter <psychon@znc.in>
3159Date:   Mon Nov 18 20:30:18 2013 +0100
3160
3161    Revert "Remove xcb_ge_event_t from xcb.h"
3162    
3163    This reverts commit f4d5b84800f960831e4fbb3ad9848bbb701020be.
3164    
3165    The version of this struct that the code generator produces breaks the API,
3166    because it gives the fields different (albeit better) names. Thus, we need to
3167    restore the old version of this struct.
3168    
3169    Additionally to the revert, this struct is documented as being deprecated. The
3170    replacement was added to xcb-proto.
3171    
3172    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71502
3173    Signed-off-by: Uli Schlachter <psychon@znc.in>
3174    Signed-off-by: Julien Cristau <jcristau@debian.org>
3175
3176commit 18f0afab3f0de68114fe185e89d8b25a8c072a2c
3177Author: Daniel Martin <consume.noise@gmail.com>
3178Date:   Fri Nov 22 23:27:28 2013 +0100
3179
3180    c_client.py: Fix _sizeof() functions
3181    
3182    Currently, it is not possible to correctly iterate over the replies of
3183    some requests. For example, the list of XIDeviceInfo returned by
3184    the XIQueryDevice request from xinput2 is read as garbage starting from
3185    the second entry.
3186    
3187    The culprits are the _sizeof() used by the iterators. In the above case:
3188    
3189        int
3190        xcb_input_xi_device_info_sizeof (const void  *_buffer  /**< */)
3191        {
3192            char *xcb_tmp = (char *)_buffer;
3193            [...]
3194            unsigned int xcb_block_len = 0;
3195            [...]
3196    
3197            xcb_block_len += sizeof(xcb_input_xi_device_info_t);
3198            xcb_tmp += xcb_block_len;
3199            /* name */
3200            xcb_block_len += (((_aux->name_len + 3) / 4) * 4) * sizeof(char);
3201            xcb_tmp += xcb_block_len;
3202            [...]
3203        }
3204    
3205    The problem here is that `xcb_block_len` is not zero'd right above the
3206    `/* name */` comment, causing `xcb_tmp` to be incremented by
3207    `sizeof(xcb_input_xi_device_info_t)` twice. The returned size is too
3208    large.
3209    
3210    https://bugs.freedesktop.org/show_bug.cgi?id=68387
3211    
3212    Tested-by: Ran Benita <ran234@gmail.com>
3213    Reviewed-by: Ran Benita <ran234@gmail.com>
3214    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
3215    Signed-off-by: Ran Benita <ran234@gmail.com>
3216    Signed-off-by: Julien Cristau <jcristau@debian.org>
3217
3218commit 5d1dbb468bb0f834eaa8adea6daf6729808ca429
3219Author: Uli Schlachter <psychon@znc.in>
3220Date:   Mon Nov 18 19:49:41 2013 +0100
3221
3222    Revert "fix deadlock with xcb_take_socket/return_socket v3"
3223    
3224    This reverts commit 9ae84ad187e2ba440c40f44b8eb21c82c2fdbf12.
3225    
3226    After this patch was merged, there were complaints about it not being a good
3227    idea. Revert this for now until we can agree on this.
3228    
3229    References: http://lists.freedesktop.org/archives/xcb/2013-June/008340.html
3230    Signed-off-by: Uli Schlachter <psychon@znc.in>
3231    
3232    Conflicts:
3233            src/xcbint.h
3234
3235commit c7c5b710f2cc0782412c9e159986c96b52aa0d02
3236Author: Mark Kettenis <kettenis@openbsd.org>
3237Date:   Mon Nov 11 23:11:56 2013 +0100
3238
3239    Fix alignment issues in FD passing code
3240    
3241    A char array on the stack is not guaranteed to have more than byte alignment.
3242    This means that casting it to a 'struct cmsghdr' and accessing its members
3243    may result in unaligned access.  This will generate SIGBUS on struct
3244    alignment architectures like OpenBSD/sparc64.  The canonical solution is to
3245    use a union to force proper alignment.
3246    
3247    Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
3248    Reviewed-by: Matthieu Herrb <matthieu@herrb.eu>
3249    Signed-off-by: Uli Schlachter <psychon@znc.in>
3250
3251commit e8663a935890ff366f49e356211049dfd0d9756a
3252Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3253Date:   Thu Nov 7 20:23:27 2013 -0800
3254
3255    Check if we need to define _XOPEN_SOURCE for struct msghdr.msg_control
3256    
3257    Required to expose the structure members in Solaris headers, since it
3258    was an XPG4/UNIX95 addition to the Solaris ABI.
3259    
3260    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3261
3262commit af8067cbf48561f1e2d43e153292e68e0376a8f9
3263Author: Keith Packard <keithp@keithp.com>
3264Date:   Thu Nov 7 17:36:01 2013 -0800
3265
3266    Add configure option to enable or disable fd passing with sendmsg
3267    
3268    --disable-sendfds or --enable-sendfds
3269    
3270    By default, configure auto-detects based on whether your system
3271    supports sendmsg at all.
3272    
3273    Signed-off-by: Keith Packard <keithp@keithp.com>
3274    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3275    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3276
3277commit 83f28ef8655acff746eab64eabe2e31f8cf0c892
3278Author: Keith Packard <keithp@keithp.com>
3279Date:   Thu Nov 7 17:28:45 2013 -0800
3280
3281    Switch to using the CMSG_* macros for FD passing
3282    
3283    Use these instead of computing the values directly so that it might
3284    work on BSD or other non-Linux systems
3285    
3286    Signed-off-by: Keith Packard <keithp@keithp.com>
3287    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3288    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3289
3290commit cca607409068ad0948e7283fb8d0465cabc51686
3291Author: Keith Packard <keithp@keithp.com>
3292Date:   Thu Jul 11 16:01:02 2013 -0700
3293
3294    Add Present extension
3295    
3296    Signed-off-by: Keith Packard <keithp@keithp.com>
3297    Reviewed-By: Uli Schlachter <psychon@znc.in>
3298
3299commit 7a9373078e69b2cb2753570f91e5c31062ba25f8
3300Author: Keith Packard <keithp@keithp.com>
3301Date:   Tue Apr 9 21:35:52 2013 -0700
3302
3303    Add DRI3 library
3304    
3305    Signed-off-by: Keith Packard <keithp@keithp.com>
3306    Reviewed-By: Uli Schlachter <psychon@znc.in>
3307
3308commit 93d733e85ded5e92292d36fc7025f0c8ff7b1167
3309Author: Keith Packard <keithp@keithp.com>
3310Date:   Thu Nov 7 05:20:06 2013 -0800
3311
3312    Require xcb proto version 1.9
3313    
3314    Signed-off-by: Keith Packard <keithp@keithp.com>
3315
3316commit 79019541e7c56ddfc3828b7bf96e6e5d3cf81c56
3317Author: Keith Packard <keithp@keithp.com>
3318Date:   Fri Apr 12 20:15:41 2013 -0700
3319
3320    Add event queue splitting
3321    
3322    This allows apps to peel off certain XGE events into separate queues
3323    for custom handling. Designed to support the Present extension
3324    
3325    Signed-off-by: Keith Packard <keithp@keithp.com>
3326    Reviewed-By: Uli Schlachter <psychon@znc.in>
3327
3328commit 7983bf0fbdc2725403f9db6154d0f5bc944040e5
3329Author: Keith Packard <keithp@keithp.com>
3330Date:   Fri Jan 18 01:29:40 2013 -0800
3331
3332    Add support for receiving fds in replies
3333    
3334    Requests signal which replies will have fds, and the replies report
3335    how many fds they expect in byte 1.
3336    
3337    Signed-off-by: Keith Packard <keithp@keithp.com>
3338    Reviewed-By: Uli Schlachter <psychon@znc.in>
3339
3340commit 7b53fb0f9bddae77b3ab8823743db57faee4e99b
3341Author: Keith Packard <keithp@keithp.com>
3342Date:   Mon Jan 14 11:23:00 2013 -0800
3343
3344    Add xcb_send_fd API
3345    
3346    This uses sendmsg to transmit file descriptors from the application to
3347    the X server
3348    
3349    Signed-off-by: Keith Packard <keithp@keithp.com>
3350    Reviewed-By: Uli Schlachter <psychon@znc.in>
3351
3352commit 98c227a2222fb5c7ca7e8101b1ed2bc096a33048
3353Author: Keith Packard <keithp@keithp.com>
3354Date:   Fri Jan 18 01:28:56 2013 -0800
3355
3356    -pendantic is too pendantic
3357    
3358    Many system headers have warnings when compiled with this flag.
3359    
3360    Signed-off-by: Keith Packard <keithp@keithp.com>
3361    Reviewed-By: Uli Schlachter <psychon@znc.in>
3362
3363commit d04a4a03a90f2721d507287938c90f1755d9da0e
3364Author: Keith Packard <keithp@keithp.com>
3365Date:   Fri Jul 12 10:32:03 2013 -0700
3366
3367    Make protocol C files depend on protocol XML files
3368    
3369    When new XML files get installed, make sure the C files are regenerated
3370    
3371    Signed-off-by: Keith Packard <keithp@keithp.com>
3372    Reviewed-By: Uli Schlachter <psychon@znc.in>
3373
3374commit f4d5b84800f960831e4fbb3ad9848bbb701020be
3375Author: Keith Packard <keithp@keithp.com>
3376Date:   Wed Nov 6 19:33:53 2013 -0800
3377
3378    Remove xcb_ge_event_t from xcb.h
3379    
3380    xcb proto now publishes this structure from an XML description
3381    
3382    Signed-off-by: Keith Packard <keithp@keithp.com>
3383
3384commit e4e0c6eec861f4c69da12060dc8dbe7a63fa5eb6
3385Author: Daphne Pfister <daphnediane@mac.com>
3386Date:   Sat Sep 14 17:36:22 2013 -0400
3387
3388    Use /usr/spool/sockets/X11/ on HP-UX for UNIX sockets (#69118).
3389
3390commit f1405d9fe4a6ddcae24585ba254389a4c4f4c8c9
3391Author: Daphne Pfister <daphnediane@mac.com>
3392Date:   Sun Sep 8 16:25:11 2013 -0400
3393
3394    Fix poll() if POLLIN == ROLLRDNORM|POLLRDBAND
3395    
3396    It seems like POLLIN is specified as equivalent to POLLRDNORM | POLLRDBAND. Some
3397    systems (e.g. QNX and HP-UX) take this literaly and have POLLIN defined as the
3398    above bit combination. Other systems (e.g. Linux) have POLLIN as just a single
3399    bit.
3400    
3401    This means that if no out-of-band data is available (which should never be the
3402    case), the result of poll() will not fulfil (fd.revents & POLLIN) == POLLIN on
3403    QNX, because the POLLRDBAND bit is not set.
3404    
3405    In other words, even though poll() signaled that the fd is readable, xcb would
3406    not read from the file descriptor.
3407    
3408    Fix this by checking if any bits from POLLIN are set in the result of poll(),
3409    instead of all of them.
3410    
3411    (This change was independently done by seanb@qnx.com as well)
3412    
3413    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=38001
3414    Acked-by: Julien Cristau <jcristau@debian.org>
3415    Signed-off-by: Uli Schlachter <psychon@znc.in>
3416
3417commit ac47e0ecdb46aa91b191a59364437a8f65947467
3418Author: Uli Schlachter <psychon@znc.in>
3419Date:   Sun Sep 8 22:16:39 2013 +0200
3420
3421    Fix documentation of xcb_poll_for_event()
3422    
3423    In commit 8eba8690adac2, the API documentation for xcb_poll_for_event() was
3424    fixed to remove an argument that was previously removed in commit 34168ab549.
3425    
3426    However, that commit only removed the first line of the documentation, leaving
3427    behind a spurious half-sentence. That commit happened seven years ago and now
3428    finally someone noticed...
3429    
3430    Thanks to Benjamin Herr for reporting this on IRC.
3431    
3432    v2: Thanks again to Benjamin Herr for noticing that my commit message blamed the
3433    wrong commit.
3434    
3435    Signed-off-by: Uli Schlachter <psychon@znc.in>
3436
3437commit c52f2891b43ae77008f63700982f800371458f4d
3438Author: Daniel Martin <consume.noise@gmail.com>
3439Date:   Sun Aug 11 13:25:18 2013 +0200
3440
3441    tests: Add files to .gitignore
3442    
3443    Add check_all.log, check_all.trs and test-suite.log to tests/.gitignore.
3444    
3445    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3446
3447commit a8d11c36edf5c49b718664dd7206f36be150f694
3448Author: Daniel Martin <consume.noise@gmail.com>
3449Date:   Thu Jul 25 11:09:26 2013 +0200
3450
3451    Sort gitignore, adjust pattern for config.h
3452    
3453    Don't ignore the files config.h and config.h.in, adjust the pattern to
3454    ignore config.h*. This matches an additional config.h.in~ too.
3455    
3456    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3457    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3458    Reviewed-by: Uli Schlachter <psychon@znc.in>
3459    Tested-By: Ran Benita <ran234@gmail.com>
3460
3461commit cbe54c97b3f0e4d40e0ee18796f8077cb4a6c16a
3462Author: Daniel Martin <consume.noise@gmail.com>
3463Date:   Thu Jul 25 10:56:30 2013 +0200
3464
3465    Use m4 directory
3466    
3467    - Follow the suggestion by libtoolize:
3468       "Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
3469        rerunning libtoolize, to keep the correct libtool macros in-tree.
3470        Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am."
3471      and add the macro and define.
3472    
3473    - Create the m4 directory and move acinclude.m4 as xcb.m4 there.
3474    
3475    - Ignore the m4 files libtoolize copies into the m4 directory
3476      (m4/l*.m4).
3477    
3478    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3479    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3480    Reviewed-by: Uli Schlachter <psychon@znc.in>
3481    Tested-By: Ran Benita <ran234@gmail.com>
3482
3483commit 6746ab1549d34a146c8383ed5acdabf0d48c1889
3484Author: Daniel Martin <consume.noise@gmail.com>
3485Date:   Wed Jul 24 14:13:41 2013 +0200
3486
3487    Use build-aux as autom4te cache directory
3488    
3489    Remove the generated directory ./autom4te.cache by reusing ./build-aux
3490    as cache directory.
3491    
3492    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3493    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3494    Reviewed-by: Uli Schlachter <psychon@znc.in>
3495    Tested-By: Ran Benita <ran234@gmail.com>
3496
3497commit b6d8c8fe61f25e4eb7b43d3d9b1de81a1f0638a3
3498Author: Daniel Martin <consume.noise@gmail.com>
3499Date:   Wed Jul 24 14:08:38 2013 +0200
3500
3501    Set AC_CONFIG_AUX_DIR to build-aux
3502    
3503    Do not clutter the project directory with generated/copied auxiliary
3504    files, save them in ./build-aux.
3505    
3506    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3507    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3508    Reviewed-by: Uli Schlachter <psychon@znc.in>
3509    Tested-By: Ran Benita <ran234@gmail.com>
3510
3511commit e3b34ad346efcaf0fd28b95e68d99388276802bc
3512Author: Daniel Martin <consume.noise@gmail.com>
3513Date:   Wed Jul 24 19:22:44 2013 +0200
3514
3515    Remove second AC_PREREQ, require version 2.60
3516    
3517    Remove a second AC_PREREQ and bump the required autoconf version to
3518    2.60.
3519    
3520    Version 2.59c was a testing release, published in April 2006. Version
3521    2.60 was the stable release afterwards, released in June 2006.
3522    
3523    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3524    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3525    Reviewed-by: Uli Schlachter <psychon@znc.in>
3526    Tested-By: Ran Benita <ran234@gmail.com>
3527
3528commit 50fb3a6312dd0b6b613fc886ffd6827952d1e286
3529Author: Daniel Martin <consume.noise@gmail.com>
3530Date:   Wed Jul 24 12:51:04 2013 +0200
3531
3532    Initialize automake earlier (bugfix for #66413)
3533    
3534    This fixes:
3535        https://bugs.freedesktop.org/show_bug.cgi?id=66413
3536        Bug 66413 - libxcb 1.9.1: Fails to build on Arch Linux: \
3537            /home/<user>/install-sh: No such file or directory
3538    
3539    Without that patch the search path for `install-sh` will become $HOME
3540    and the `install` target will fail, when DESTDIR doesn't exist in
3541    advance. (occured with automake 1.14 and autoconf 2.69)
3542    
3543    Initial patch by: Alain Kalker <a.c.kalker@gmail.com>
3544    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3545    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3546    Reviewed-by: Uli Schlachter <psychon@znc.in>
3547    Tested-By: Ran Benita <ran234@gmail.com>
3548
3549commit dd01db570c34dd3e2e0f5d07b8d40c837fa51057
3550Author: Daniel Martin <consume.noise@gmail.com>
3551Date:   Mon Dec 31 11:57:49 2012 +0100
3552
3553    Make xsltproc optional
3554    
3555    Fix Bug 23863 - xcb still checks for xsltproc:
3556        https://bugs.freedesktop.org/show_bug.cgi?id=23863
3557    
3558    xsltproc is used to generate the optional html page for `check` results,
3559    only. So, it's not a hard build dependency.
3560    
3561    Additionally, use yes/no instead of true/false in the HTML_CHECK_RESULT
3562    variable for consistent output after a configure run.
3563    
3564    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3565    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3566
3567commit 0289348f2c4ed3b1b286c51df19d82c6787c2b36
3568Author: Daniel Martin <consume.noise@gmail.com>
3569Date:   Fri Dec 28 23:25:16 2012 +0100
3570
3571    c_client.py: Do not create pointers in unions
3572    
3573    Do not create pointers in unions for fields of variadic length.
3574    
3575    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3576    Reviewed-by: Ran Benita <ran234@gmail.com>
3577
3578commit b9efd2a09a45616f6238e4da49b8f2127b6ec6d9
3579Author: Daniel Martin <consume.noise@gmail.com>
3580Date:   Wed Jan 9 12:52:15 2013 +0100
3581
3582    c_client.py: Always initialize xcb_align_to
3583    
3584    to get rid of:
3585        warning: 'xcb_align_to' may be used uninitialized in this function
3586    
3587    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3588    Reviewed-by: Peter Harris <pharris@opentext.com>
3589
3590commit 5648ddd2b97068f549268284129a438a6845e14c
3591Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3592Date:   Sat Aug 3 20:25:23 2013 -0700
3593
3594    Define _xcb_map_new with explicit void arg list instead of empty one
3595    
3596    Fixes Solaris Studio compiler warning:
3597    "xcb_list.c", line 50: warning: old style function definition
3598    
3599    and gcc warning:
3600    xcb_list.c: In function '_xcb_map_new':
3601    xcb_list.c:50:11: warning: old-style function definition [-Wold-style-definition]
3602    
3603    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3604    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3605
3606commit cb51f271b26c6abc76d415553f202bc5139273ca
3607Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3608Date:   Sat Aug 3 20:22:25 2013 -0700
3609
3610    Enable warnings for pre-C89 style definitions for gcc & Solaris Studio
3611    
3612    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3613    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3614
3615commit bc6a4f557ff4e497acdafdcebb006e5a7b4c5b11
3616Author: Michael Stapelberg <michael@stapelberg.de>
3617Date:   Mon Aug 5 22:14:18 2013 +0200
3618
3619    Build xcb-xkb by default
3620    
3621    There have not been any big issues with xcb-xkb recently.
3622    
3623    Also, Wayland is using xcb-xkb actively, making distributions compile
3624    libxcb with xkb support anyway, so let’s reflect reality :).
3625    
3626    Reviewed-by: Jamey Sharp <jamey@minilop.net>
3627    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3628
3629commit 45619dc71e9411a526d7c69595cf615b1b1206cf
3630Author: Daniel Martin <consume.noise@gmail.com>
3631Date:   Sat Jun 8 11:20:39 2013 +0200
3632
3633    c_client.py: Inject full_sequence into GE events
3634    
3635    The generic event structure xcb_ge_event_t has the full_sequence field
3636    at the 32byte boundary. That's why we've to inject this field into GE
3637    events while generating the structure for them. Otherwise we would read
3638    garbage (the internal full_sequence) when accessing normal event fields
3639    there.
3640    
3641    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3642    Reviewed-by: Keith Packard <keithp@keithp.com>
3643    Signed-off-by: Peter Harris <pharris@opentext.com>
3644
3645commit a1e67b141a57d39cbcaff1a703d6fc0da1fbb56d
3646Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3647Date:   Mon Jul 8 17:54:35 2013 -0700
3648
3649    Fix "indention" typos in xcb-examples.3 man page
3650    
3651    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3652
3653commit 8b6bb1a71977116d382f45eef803aedd3e313d37
3654Author: Marc Deslauriers <marc.deslauriers@canonical.com>
3655Date:   Wed Jun 5 16:38:00 2013 -0400
3656
3657    Update Makefile.am for newer automake
3658    
3659    Debian Bug #710344
3660    
3661    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
3662
3663commit 9ae84ad187e2ba440c40f44b8eb21c82c2fdbf12
3664Author: Christian König <christian.koenig@amd.com>
3665Date:   Wed May 15 11:21:36 2013 +0200
3666
3667    fix deadlock with xcb_take_socket/return_socket v3
3668    
3669    To prevent different threads from stealing the socket from each other the
3670    caller of "xcb_take_socket" must hold a lock that is also acquired in
3671    "return_socket". Unfortunately xcb tries to prevent calling return_socket
3672    from multiple threads and this can lead to a deadlock situation.
3673    
3674    A simple example:
3675    - X11 has taken the socket
3676    - Thread A has locked the display.
3677    - Thread B does xcb_no_operation() and thus ends up in libX11's return_socket(),
3678      waiting for the display lock.
3679    - Thread A calls e.g. xcb_no_operation(), too, ends up in return_socket() and
3680      because socket_moving == 1, ends up waiting for thread B
3681    => Deadlock
3682    
3683    This patch allows calling return_socket from different threads at the same time
3684    an so resolves the deadlock situation.
3685    
3686    Partially fixes: https://bugs.freedesktop.org/show_bug.cgi?id=20708
3687    
3688    v2: fixes additional pthread_cond_wait dependencies,
3689        rework comments and patch description
3690    
3691    v3: separate pthread_cond_wait dependencies and unrelated whitespace
3692        change into their own patch, use unsigned for socket_seq
3693    
3694    Signed-off-by: Christian König <christian.koenig@amd.com>
3695    Signed-off-by: Uli Schlachter <psychon@znc.in>
3696
3697commit 1b33867fa996034deb50819ae54640be501f8d20
3698Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3699Date:   Wed May 1 17:59:31 2013 -0700
3700
3701    integer overflow in read_packet() [CVE-2013-2064]
3702    
3703    Ensure that when calculating the size of the incoming response from the
3704    Xserver, we don't overflow the integer used in the calculations when we
3705    multiply the int32_t length by 4 and add it to the default response size.
3706    
3707    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3708
3709commit e602b653c191e18cbb63db6526aac77c368ed70b
3710Author: Daniel Martin <consume.noise@gmail.com>
3711Date:   Mon May 13 23:33:04 2013 +0200
3712
3713    c_client.py: Handle multiple expr. in a bitcase
3714    
3715    Adopt a change from xcbgen. With that modification the expression in a
3716    bitcase became a list of expressions to support multiple <enumref> in a
3717    <bitcase>.
3718    
3719    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3720    Signed-off-by: Peter Harris <pharris@opentext.com>
3721
3722commit 6b6044cb8aacdf1b637da7b5dda392f9ff41ed39
3723Author: Christian König <christian.koenig@amd.com>
3724Date:   Wed May 15 11:21:35 2013 +0200
3725
3726    whitespace fix in xcb_take_socket
3727    
3728    Signed-off-by: Christian König <christian.koenig@amd.com>
3729    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3730
3731commit 0dd8f8d26a758bc385e79d9239bf6ef2e3d7bf13
3732Author: Colin Walters <walters@verbum.org>
3733Date:   Wed Jan 4 17:37:06 2012 -0500
3734
3735    autogen.sh: Implement GNOME Build API
3736    
3737    http://people.gnome.org/~walters/docs/build-api.txt
3738    
3739    Signed-off-by: Adam Jackson <ajax@redhat.com>
3740
3741commit 76a2166de9c80b35f987fdc3f3a228bafa0de94e
3742Author: Chí-Thanh Christopher Nguyễn <chithanh@gentoo.org>
3743Date:   Thu Oct 11 01:14:12 2012 +0200
3744
3745    c_client.py: Fix python-3 invalid except statement
3746    
3747    Replace except statement with a PEP-3110 compliant one. This fixes a regression
3748    introduced by c3deeaf714630531d693a6a902b8dabf791858b1
3749    https://bugs.freedesktop.org/show_bug.cgi?id=55690
3750    
3751    Reviewed-by: Peter Harris <pharris@opentext.com>
3752    Signed-off-by: Uli Schlachter <psychon@znc.in>
3753
3754commit 9db4517c87f56bb0ac82b647a08db30850ee2b04
3755Author: Chí-Thanh Christopher Nguyễn <chithanh@gentoo.org>
3756Date:   Thu Oct 11 01:14:11 2012 +0200
3757
3758    c-client.py: Fix python-3 AttributeError: 'dict' object has no attribute 'iteritems'
3759    
3760    This fixes a regression introduced by ea71d7d7e3f5d8189b80747678e9ca9a417b1d37
3761    https://bugs.freedesktop.org/show_bug.cgi?id=55690
3762    
3763    Reviewed-by: Peter Harris <pharris@opentext.com>
3764    Signed-off-by: Uli Schlachter <psychon@znc.in>
3765
3766commit 4ffe54f69049e6792a35a287fd9ff83abbd4fd8d
3767Author: Uli Schlachter <psychon@znc.in>
3768Date:   Fri Oct 5 14:53:37 2012 +0200
3769
3770    Release libxcb 1.9
3771    
3772    Signed-off-by: Uli Schlachter <psychon@znc.in>
3773
3774commit 4f52f884f42b72087f3323f2bab204223664a488
3775Author: Uli Schlachter <psychon@znc.in>
3776Date:   Fri Oct 5 11:23:26 2012 +0200
3777
3778    Include static man pages in "make dist"
3779    
3780    This was found by distcheck. It tried to install src/man/xcb-examples.3 and
3781    src/man/xcb-requests.3, but those files weren't in the distribution.
3782    
3783    Fix this by explicitly telling automake to distribute those files.
3784    
3785    Signed-off-by: Uli Schlachter <psychon@znc.in>
3786
3787commit 23911a707b8845bff52cd7853fc5d59fb0823cef
3788Author: Uli Schlachter <psychon@znc.in>
3789Date:   Mon Sep 24 22:07:51 2012 +0200
3790
3791    Fix a multi-thread deadlock
3792    
3793    This fixes a deadlock which was seen in-the-wild with wine.
3794    
3795    It could happen that two threads tried to read from the socket at the same time
3796    and one of the thread got stuck inside of poll()/select().
3797    
3798    The fix works by making sure that the writing thread doesn't steal the reading
3799    thread's reply.
3800    
3801    Debugged-by: Erich Hoover <ehoover@mines.edu>
3802    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54671
3803    Signed-off-by: Uli Schlachter <psychon@znc.in>
3804
3805commit c16cc5467eb0af7c5cdee16e6efaee54eb53bba6
3806Author: Uli Schlachter <psychon@znc.in>
3807Date:   Mon Sep 24 22:07:30 2012 +0200
3808
3809    Add a .gitignore for src/man/
3810    
3811    Signed-off-by: Uli Schlachter <psychon@znc.in>
3812
3813commit 08cc068ead7b8e678cdb119b38ada5261d5cc3ea
3814Author: Peter Harris <pharris@opentext.com>
3815Date:   Thu Aug 16 11:59:14 2012 -0400
3816
3817    Allow xcb_send_request with >MAX_IOV iovecs
3818    
3819    This allows an application to do a scatter/gather operation on a large
3820    image buffer to avoid the extra memcpy.
3821    
3822    Use autoconf to use UIO_MAXIOV where IOV_MAX is not available (and the
3823    POSIX minimum of 16 where neither are available).
3824    
3825    Reviewed-by: Uli Schlachter <psychon@znc.in>
3826    Signed-off-by: Peter Harris <pharris@opentext.com>
3827
3828commit ff53285ae3f604e9f2cc5f4837255220459b5e44
3829Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3830Date:   Sat Aug 25 13:53:37 2012 -0700
3831
3832    Return connection failure if display string specifies non-existent screen
3833    
3834    Matches the behaviour of Xlib - if you set DISPLAY to :0.1 but only have
3835    one screen, closes connection and returns error.
3836    
3837    This introduces a new connection error code:
3838    XCB_CONN_CLOSED_INVALID_SCREEN
3839    
3840    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
3841    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3842    Reviewed-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
3843    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3844
3845commit 90889794ad882a6847bcffe52c4cc5dfd168f1f4
3846Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3847Date:   Fri Aug 24 23:35:41 2012 -0700
3848
3849    Add AC_USE_SYSTEM_EXTENSIONS to allow use of more system functionality
3850    
3851    Copied from libX11 configure.ac
3852    
3853    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3854
3855commit b52790e8ed4bb077eabdeca803935d2910558acc
3856Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3857Date:   Fri Aug 24 23:32:32 2012 -0700
3858
3859    Always include "config.h" at the start of all C source files.
3860    
3861    Allows configure to set defines such as _POSIX_SOURCE in config.h
3862    that affect functions exposed by system headers and get consistent
3863    results across all the source files.
3864    
3865    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3866
3867commit ed93a6a2a8e23f12380709de6e04b2d833df7e71
3868Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3869Date:   Sat Aug 25 12:33:35 2012 -0700
3870
3871    Fix "sppported" typo in doxygen comment for XCB_CONN_CLOSED_EXT_NOTSUPPORTED
3872    
3873    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3874
3875commit c3deeaf714630531d693a6a902b8dabf791858b1
3876Author: Colin Walters <walters@verbum.org>
3877Date:   Mon Aug 13 15:32:31 2012 -0400
3878
3879    c_client: Fix parallel-make issue creating 'man' directory
3880    
3881    With make -j, it was possible to hit a race condition in the code to
3882    make the 'man' directory.
3883    
3884    Signed-off-by: Julien Danjou <julien@danjou.info>
3885
3886commit 5f8f2ba1c4f9ac74c8f301dcca8566e296e37995
3887Author: Jeremy Huddleston <jeremyhu@apple.com>
3888Date:   Sat Apr 21 22:42:51 2012 -0700
3889
3890    xcb_connect: launchd: Don't fall back on tcp if $DISPLAY is a path to a launchd socket
3891    
3892    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
3893
3894commit 71a295082e07ff20d4c4cc97feed03b94cceb251
3895Author: Julien Danjou <julien@danjou.info>
3896Date:   Tue Mar 27 12:10:59 2012 +0200
3897
3898    Move static man to man
3899    
3900    Signed-off-by: Julien Danjou <julien@danjou.info>
3901
3902commit df217bf7c930d4433c991d86f857ecf63cc2d25a
3903Author: Julien Danjou <julien@danjou.info>
3904Date:   Tue Mar 27 12:10:15 2012 +0200
3905
3906    Do not list manpages, use a wildcard
3907    
3908    Signed-off-by: Julien Danjou <julien@danjou.info>
3909
3910commit a4417b1611f0bf02b81b54be90ef3353010da10a
3911Author: Julien Danjou <julien@danjou.info>
3912Date:   Tue Mar 27 12:06:54 2012 +0200
3913
3914    Split manpage list in two (static/built)
3915    
3916    Signed-off-by: Julien Danjou <julien@danjou.info>
3917
3918commit cc7fab2d5e912629d4a2a6adfb7666dc2ba45db2
3919Author: Julien Danjou <julien@danjou.info>
3920Date:   Mon Mar 26 18:29:35 2012 +0200
3921
3922    Allow undocumented code to be built
3923    
3924    Signed-off-by: Julien Danjou <julien@danjou.info>
3925
3926commit ea71d7d7e3f5d8189b80747678e9ca9a417b1d37
3927Author: Michael Stapelberg <michael@stapelberg.de>
3928Date:   Sun Nov 27 10:38:26 2011 +0000
3929
3930    c_client.py: generate manpages
3931    
3932    Signed-off-by: Julien Danjou <julien@danjou.info>
3933
3934commit 57a62e99b1241d5a0e6cf0b72f52090862a9c07d
3935Author: Julien Danjou <julien@danjou.info>
3936Date:   Fri Mar 9 15:38:56 2012 +0100
3937
3938    Release libxcb 1.8.1
3939    
3940    Signed-off-by: Julien Danjou <julien@danjou.info>
3941
3942commit 236f914ea7205f5f74e87fcc1b06d87bd0789a7a
3943Author: Uli Schlachter <psychon@znc.in>
3944Date:   Thu Mar 1 20:26:39 2012 +0100
3945
3946    Fix a busy loop on BSD and Mac OS
3947    
3948    On FreeBSD MSG_WAITALL on a non-blocking socket fails immediately if less bytes
3949    than were asked for are available. This is different than the behavior on linux
3950    where as many bytes as are available are returned in this case. Other OS
3951    apparently follow the FreeBSD behavior.
3952    
3953    _xcb_in_read() is used to fill xcb's read buffer, thus this function will call
3954    recv() with a big length argument (xcb's read buffer is by default 16 KiB
3955    large). That many bytes are highly unlikely to be available in the kernel
3956    buffer.
3957    
3958    This means that _xcb_in_read() always failed on FreeBSD. Since the socket was
3959    still signaled as readable by poll(), this bug even resulted in a busy loop.
3960    
3961    The same issue is present in read_block(), but here it is slightly different.
3962    read_block() is called when we read the first few bytes of an event or a reply,
3963    so that we already know its length. This means that we should be able to use
3964    MSG_WAITALL here, because we know how many bytes there have to be.
3965    
3966    However, that function could busy loop, too, when only the first few bytes of
3967    the packet were sent while the rest is stuck somewhere on the way to us. Thus,
3968    MSG_WAITALL should be removed here, too.
3969    
3970    Thanks to Christoph Egger from Debian for noticing the problem, doing all the
3971    necessary debugging and figuring out what the problem was! This patch is 99%
3972    from debian. Thanks for all the work.
3973    
3974    This bug was introduced in commit 2dcf8b025be88a25d4333abdc28d425b88238d96.
3975    
3976    This commit also reverts commit 9061ee45b8dbe5431c23e3f628089d703ccad0b1.
3977    
3978    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=45776
3979    
3980    Signed-off-by: Uli Schlachter <psychon@znc.in>
3981    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3982
3983commit 9061ee45b8dbe5431c23e3f628089d703ccad0b1
3984Author: Jeremy Huddleston <jeremyhu@apple.com>
3985Date:   Thu Mar 8 00:50:13 2012 -0800
3986
3987    darwin: Use read(2) rather than recv(2)
3988    
3989    2dcf8b025be88a25d4333abdc28d425b88238d96 was causing some regressions on
3990    darwin, so go back to using read(2) there until I have time to investigate
3991    further.
3992    
3993    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
3994
3995commit 7d235c62f0d5bd0df1236cc52141c10c5d272a18
3996Author: Julien Cristau <jcristau@debian.org>
3997Date:   Sun Feb 19 13:43:54 2012 +0100
3998
3999    Fallback to TCP if no protocol is specified and the UNIX connection fails
4000    
4001    Signed-off-by: Julien Cristau <jcristau@debian.org>
4002    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4003
4004commit f7bd139616d228b20eeb2c96b85e4093606c93fc
4005Author: Arnaud Fontaine <arnau@debian.org>
4006Date:   Wed Jan 25 17:18:20 2012 +0900
4007
4008    Add xkb_internals and xkb_issues to EXTRA_DIST.
4009
4010commit 68d30adde982f1be33a934707fa105c0db6f7f8f
4011Author: Jon TURNEY <jon.turney@dronecode.org.uk>
4012Date:   Tue Jan 17 20:59:56 2012 +0000
4013
4014    Update use of error_connection under WIN32 to _xcb_conn_ret_error()
4015    
4016    Unfortunately, commit 31b57676 adding WSACleanup/WSAShutdown on Win32 adds a new use
4017    of error_connection, which was removed in commit 769acff0, applied 5 minutes earlier.
4018    
4019    src/xcb_util.c: In function 'xcb_connect_to_display_with_auth_info':
4020    src/xcb_util.c:433:39: error: 'error_connection' undeclared (first use in this function)
4021    
4022    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4023    Reviewed-by: Arvind Umrao <arvind.umrao@oracle.com>
4024    Signed-off-by: Uli Schlachter <psychon@znc.in>
4025
4026commit 87b7bf875e0105924ae306e90ca79512d6c0cf47
4027Author: Jon TURNEY <jon.turney@dronecode.org.uk>
4028Date:   Wed Jan 18 14:11:07 2012 +0000
4029
4030    Fix build of xcb_auth.c with XDMCP on WIN32
4031    
4032    Fix a redefinition problem which shows up when building for _WIN32 and
4033    libXdmcp is installed, so HASXDMAUTH is enabled
4034    
4035    It seems this is a special place in xcb as it uses other X11 library headers here
4036    
4037    If HASXDMAUTH is defined, include the wrapped windows.h before any header which
4038    includes it unwrapped, to avoid conflicts with types defined in X headers
4039    
4040    We need to include config.h and check HASXDMAUTH to avoid an unconditional dependency
4041    on x11proto headers
4042    
4043    In file included from install/include/X11/Xdmcp.h:19:0,
4044                     from git/xcb/libxcb/src/xcb_auth.c:52:
4045    install/include/X11/Xmd.h:120:14: error: conflicting types for 'INT32'
4046    /usr/i686-pc-mingw32/sys-root/mingw/include/basetsd.h:54:13: note: previous declaration of 'INT32' was here
4047    install/include/X11/Xmd.h:143:15: error: conflicting types for 'BOOL'
4048    /usr/i686-pc-mingw32/sys-root/mingw/include/windef.h:234:17: note: previous declaration of 'BOOL' was here
4049    
4050    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4051    Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
4052
4053commit 6db1a2686f0f073438d36f5fa0f97b787842b0f2
4054Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
4055Date:   Tue Jan 17 23:55:23 2012 -0800
4056
4057    Revert "Fix include order with Xdmcp on WIN32"
4058    
4059    This reverts commit 0e9246def562be97cc816f824f702233a826ec56.
4060    
4061    This change caused build failures because <X11/Xdmcp.h> was never
4062    included under any circumstance.  This is because the check for
4063    HASXDMAUTH was moved before the inclusion of config.h (via xcbint.h)
4064    which defined it.
4065    
4066    Found-by: Tinderbox
4067    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4068    Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4069
4070commit da1d15082baab844a3b9b2d5cc48620af0b806ec
4071Author: Julien Danjou <julien@danjou.info>
4072Date:   Thu Jan 12 10:25:07 2012 +0100
4073
4074    Bump xcb-proto requirement
4075    
4076    We are now unable to build xcb-proto before 1.7.
4077    
4078    Signed-off-by: Julien Danjou <julien@danjou.info>
4079
4080commit b95b33e8c04c90c2240df19acea0c841d6e3450b
4081Author: Julien Danjou <julien@danjou.info>
4082Date:   Wed Jan 11 18:27:38 2012 +0100
4083
4084    Release libxcb 1.8
4085    
4086    Signed-off-by: Julien Danjou <julien@danjou.info>
4087
4088commit 31b57676e8d7ab6048dbfb145187833fac5e478c
4089Author: Ryan Pavlik <rpavlik@iastate.edu>
4090Date:   Wed Jan 11 18:06:50 2012 +0100
4091
4092    Use WSAStartup()/WSACleanup() on WIN32
4093    
4094    The alternative is to use these in every WIN32 application which uses xcb. Doing
4095    it this way should be safe, as, according to MSDN, "There must be a call to
4096    WSACleanup for each successful call to WSAStartup. Only the final WSACleanup
4097    function call performs the actual cleanup. The preceding calls simply decrement
4098    an internal reference count"
4099    
4100    (We should probably also include ws2_32 in Libs.private for libxcb, as anything
4101    which links with libxcb will also need that, but there seems to be some pkg-config
4102    issues to resolve first...)
4103    
4104    v2: Check for errors so WSAStartup()/WSACleanup() uses are balanced
4105    v3: Use same indentation style as surrounding code
4106    
4107    Reviewed-by: Peter Harris <pharris@opentext.com>
4108    Signed-off-by: Julien Danjou <julien@danjou.info>
4109
4110commit 0e9246def562be97cc816f824f702233a826ec56
4111Author: Ryan Pavlik <rpavlik@iastate.edu>
4112Date:   Thu Jan 5 20:57:53 2012 +0000
4113
4114    Fix include order with Xdmcp on WIN32
4115    
4116    Fix a redefinition problem due to include order which shows up when
4117    building for _WIN32 and libXdmcp is installed, so HASXDMAUTH is enabled
4118    
4119    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4120    Reviewed-by: Peter Harris <pharris@opentext.com>
4121    Signed-off-by: Julien Danjou <julien@danjou.info>
4122
4123commit 4aa7a2c849a9536febb2dc7773e06c12a69c5213
4124Author: Jon TURNEY <jon.turney@dronecode.org.uk>
4125Date:   Thu Jan 5 20:57:52 2012 +0000
4126
4127    Fix WIN32 compilation after commit 163c47bdc0d32785d831e4c93fea9ab7e023446b
4128    
4129    WIN32 does not have arpa/inet.h, so do not try to include it unless _WIN32 is
4130    not defined
4131    
4132    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4133    Reviewed-by: Peter Harris <pharris@opentext.com>
4134    Signed-off-by: Julien Danjou <julien@danjou.info>
4135
4136commit 769acff0da8b9859ebdf052dce80045465c7598c
4137Author: Arvind Umrao <arvind.umrao@oracle.com>
4138Date:   Fri Nov 4 15:42:05 2011 +0530
4139
4140    Added more error states and removed global error_connection
4141    
4142    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=41443
4143    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=42304
4144    
4145    I have added more xcb connection error states at xcb.h header.
4146    Also I have removed global error_connection variable, and added
4147    an interface that returns connection error state.
4148    
4149    TBD:
4150    I will segregate errors states in a separate header file and try to
4151    provide more precise error states, in future. Also I will give patch
4152    for libX11, in that patch xcb_connection_t::has_error will be passed
4153    to default io handler of libX11. This value can then be used for
4154    displaying error messages.
4155    
4156    Reviewed-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4157    Reviewed-by: Uli Schlachter <psychon@znc.in>
4158    Signed-off-by: Arvind Umrao <arvind.umrao@oracle.com>
4159
4160commit 9b4d6f30a339e2d18ebaea98149da81aba453216
4161Author: Keith Packard <keithp@keithp.com>
4162Date:   Thu Dec 1 10:28:51 2011 +0000
4163
4164    Make xcb_take_socket keep flushing until idle
4165    
4166    _xcb_out_flush_to will drop the iolock in pthread_cond_wait allowing
4167    other threads to queue new requests. When this happened,
4168    there would be requests queued for the socket after _xcb_out_flush_to
4169    returned, and xcb_take_socket would throw an assert.
4170    
4171    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29875
4172    Signed-off-by: Keith Packard <keithp@keithp.com>
4173    Signed-off-by: Julien Danjou <julien@danjou.info>
4174
4175commit 5ceeaaa4294201b3f613c07f9ec610c0e5f673c7
4176Author: Uli Schlachter <psychon@znc.in>
4177Date:   Thu Aug 25 14:18:16 2011 +0200
4178
4179    Fix a dead-lock due to xcb_poll_for_reply
4180    
4181    Imagine two threads:
4182    
4183    Thread#1: for(;;) { xcb_get_input_focus_reply(c, xcb_get_input_focus(c), 0); }
4184    
4185    Thread#2: for(;;) { xcb_poll_for_event(c); }
4186    
4187    Since xcb_poll_for_event() calls _xcb_in_read() directly without synchronizing
4188    with any other readers, this causes two threads to end up calling recv() at the
4189    same time. We now have a race because any of these two threads could get read
4190    the GetInputFocus reply.
4191    
4192    If thread#2 reads this reply, it will be put in the appropriate queue and
4193    thread#1 will still be stuck in recv(), although its reply was already received.
4194    If no other reply or event causes this thread to wake up, the process deadlocks.
4195    
4196    To fix this, we have to make sure that there is only ever one thread reading
4197    from the connection. The obvious solution is to check in poll_for_next_event()
4198    if another thread is already reading (in which case c->in.reading != 0) and not
4199    to read from the wire in this case.
4200    
4201    This solution is actually correct if we assume that the other thread is blocked
4202    in poll() which means there isn't any data which can be read. Since we already
4203    checked that there is no event in the queue this means that
4204    poll_for_next_event() didn't find any event to return.
4205    
4206    There might be a small race here where the other thread already determined that
4207    there is data to read, but it still has to wait for c->iolock. However, this
4208    means that the next poll_for_next_event() will be able to read the event, so
4209    this shouldn't cause any problems.
4210    
4211    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=40372
4212    
4213    Signed-off-by: Uli Schlachter <psychon@znc.in>
4214    Signed-off-by: Peter Harris <pharris@opentext.com>
4215
4216commit b12038e9ae5343c4176f11d68c963c752bc35c03
4217Author: Jamey Sharp <jamey@minilop.net>
4218Date:   Wed Aug 24 08:52:02 2011 -0700
4219
4220    Keep ALIGNOF definition out of the public namespace.
4221    
4222    Uli's patch is an excellent solution; I just want to keep the new
4223    ALIGNOF macro hidden from XCB's users, as they don't need it to call
4224    XCB.
4225    
4226    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4227
4228commit 163c47bdc0d32785d831e4c93fea9ab7e023446b
4229Author: Markus Duft <mduft@gentoo.org>
4230Date:   Wed Aug 24 10:49:06 2011 -0400
4231
4232    Support pre-IPv6 systems (without getaddrinfo)
4233    
4234    Some of these systems (eg. Interix on XP) are still in use.
4235    
4236    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
4237    Signed-off-by: Peter Harris <pharris@opentext.com>
4238
4239commit aa02096b8e7f94ad3c998a8d5af54963ee860b13
4240Author: Uli Schlachter <psychon@znc.in>
4241Date:   Wed Aug 24 12:47:16 2011 +0200
4242
4243    Compute alignment correctly
4244    
4245    The code previously assumed that everything has to be aligned to a 4 byte
4246    boundary. This assumption is wrong as e.g. the STR struct from xproto shows.
4247    
4248    Instead, each type has to be aligned to its natural alignment. So a char doesn't
4249    need any alignment, a INT16 gets aligned to a 2-byte-boundary and a INT32 gets
4250    the old 4 byte alignment.
4251    
4252    I'm not 100% sure that this commit is correct, but some quick tests with awesome
4253    and cairo-xcb went well.
4254    
4255    This commit causes lots of dead assignments to xcb_align_to since only the last
4256    field's alignment is actually used, but this simplified this patch a lot.
4257    
4258    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=34037
4259    
4260    Signed-off-by: Uli Schlachter <psychon@znc.in>
4261    Signed-off-by: Peter Harris <pharris@opentext.com>
4262
4263commit 4f25ee16443b29e1a25bd26a724e1e0a577e21ff
4264Author: Michael Stapelberg <michael@stapelberg.de>
4265Date:   Thu Aug 18 21:38:28 2011 +0200
4266
4267    Drop AI_ADDRCONFIG when resolving TCP addresses
4268    
4269    When a system is completely offline (no interface has an IP address but 'lo'),
4270    xcb could not connect to localhost via TCP, e.g. connections with
4271    DISPLAY=127.0.0.1:0 fail.
4272    
4273    AI_ADDRCONFIG will only return IPv4 addresses if the system has an IPv4
4274    address configured (likewise for IPv6). This also takes place when
4275    resolving localhost (or 127.0.0.0/8 or ::1). Also, as per RFC 3493,
4276    loopback addresses are not considered as valid addresses when
4277    determining whether to return IPv4 or IPv6 addresses.
4278    
4279    As per mailing-list discussion on the xcb list started with message
4280    20110813215405.5818a0c1@x200, the AI_ADDRCONFIG flag is there for historical
4281    reasons:
4282    
4283        In the old days, the "default on-link" assumption in IPv6 made the flag vey
4284        much indispensable for dual-stack hosts on IPv4-only networks. Without it,
4285        there would be long timeouts trying non-existent IPv6 connectivity. Nowadays,
4286        this assumption has been flagged as historic bad practice by IETF, and hosts
4287        should have been updated to not make it anymore.
4288    
4289        Then AI_ADDRCONFIG became mostly cosmetic: it avoids phony "Protocol family
4290        not supported" or "Host unreachable" errors while trying to connect to a dual-
4291        stack mode from a host with no support for source address selection.
4292    
4293        Nowadays, on up-to-date systems, this flag is completely useless. Then again,
4294        I understood only the very latest MacOS release is "up-to-date" with this
4295        definition.
4296
4297commit 662ad589c5d6f03757ae57a926d3800bfb528b30
4298Author: James Jones <jajones@nvidia.com>
4299Date:   Wed May 11 23:22:22 2011 -0700
4300
4301    Insert, not append explicit xcbgen dir python path
4302    
4303    If a the path to the xcb python generate libs is
4304    explicitly specified to c_client.py, insert it in
4305    the python path list just after the local dir entry,
4306    rather than appending it to the existing paths.
4307    This keeps a global/distro install of xcb from
4308    overriding a local build of the xcb proto files.
4309    
4310    Signed-off-by: James Jones <jajones@nvidia.com>
4311    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4312
4313commit 294c9f455c3534d836b010dacd2e7aa62a7dde9d
4314Author: David Coles <dcoles@gaikai.com>
4315Date:   Fri Apr 8 17:47:05 2011 -0700
4316
4317    Add support for building with Python 3
4318    
4319    Python 3 introduces some language changes that cause issues when running
4320    c_client.py. This also breaks compatibility with Python 2.5 since it does not
4321    support the "as" statement in try/except blocks and does not have reduce() in
4322    the functools package.
4323    
4324    The main changes are:
4325    * try/except blocks require `except ... as ...:` to resolve syntactical ambiguity
4326    * map() and filter() return iterators rather than lists in Python 3
4327    * reduce() is now in functools package (and not built-in in Python 3)
4328    * Dictionaries don't have a has_key() method in Python 3
4329    * None and int types can't be directly compared in Python 3
4330    * print() is a statement in Python 3
4331    
4332    See http://diveintopython3.org/porting-code-to-python-3-with-2to3.html and
4333    PEP-3110 for details.
4334    
4335    Verified on Python 2.6.5 and 3.1.3.
4336    
4337    Signed-off-by: David Coles <dcoles@gaikai.com>
4338    Signed-off-by: Julien Danjou <julien@danjou.info>
4339
4340commit e300ee4920bf4618f58618f3063b362f811154c1
4341Author: Jamey Sharp <jamey@minilop.net>
4342Date:   Tue Apr 12 13:09:23 2011 -0700
4343
4344    Revert "Introduce xcb_wait_for_event_until, for consuming responses in wire-order."
4345    
4346    This function was intended to allow libX11 to fix a multi-threaded hang,
4347    but the corresponding libX11 patch caused single-threaded apps to spin
4348    sometimes. Since I've retracted that patch, this patch has no users and
4349    shouldn't go into a release unless/until that changes.
4350    
4351    This reverts commit 2415c11dec5e5adb0c17f98aa52fbb371a4f8f23.
4352    
4353    Conflicts:
4354    
4355            src/xcb.h
4356            src/xcb_in.c
4357    
4358    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4359
4360commit 527df3c84bd71113cedc7f55089c02d1c099ecad
4361Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
4362Date:   Tue Mar 22 14:33:23 2011 +0200
4363
4364    Introduce a variant of xcb_poll_for_event for examining event queue.
4365    
4366    In some circumstances using xcb_poll_for_event is suboptimal because
4367    it checks the connection for new events. This may lead to a lot of
4368    failed nonblocking read system calls.
4369    
4370    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4371    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4372
4373commit b64cd0df884e7901ff13def0272df74962035920
4374Author: Alan Coopersmith <alan.coopersmith@oracle.com>
4375Date:   Mon Apr 4 21:36:47 2011 -0700
4376
4377    If protocol is "unix", use a Unix domain socket, not TCP
4378    
4379    Fixes fallback to local connections from Xlib's XOpenDisplay(), which
4380    will try with protocol "unix" if a hostname is specified and tcp fails
4381    (as it usually will now that most OS'es ship with -nolisten tcp enabled).
4382    
4383    Also fixes explicitly specifying DISPLAY as "unix/foo:0", which Xlib
4384    previously accepted for Unix domain sockets.
4385    
4386    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4387
4388commit b027922ebf1931885e00629c20e26f14f184998d
4389Author: Alan Coopersmith <alan.coopersmith@oracle.com>
4390Date:   Mon Apr 4 21:28:00 2011 -0700
4391
4392    Make launchd code in xcb_util.c match surrounding code indent levels
4393    
4394    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4395
4396commit 82b1f3919a82a730f6b2f952d4090fe15702694e
4397Author: Carlos Garnacho <carlosg@gnome.org>
4398Date:   Tue Oct 5 18:03:22 2010 +0200
4399
4400    Handle XGE events with the "send event" flag
4401    
4402    This patch is necessary so xcb reads the payload after the message
4403    for GenericEvents with the 0x80 flag turned on.
4404    
4405    Signed-off-by: Carlos Garnacho <carlosg@gnome.org>
4406    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4407
4408commit 42c4adeff4a6aedfba30e22f71800c1b73942923
4409Author: Alan Coopersmith <alan.coopersmith@oracle.com>
4410Date:   Mon Apr 4 20:20:16 2011 -0700
4411
4412    Add #include <sys/socket.h> to xcb_conn.c
4413    
4414    Solves compiler warning on Solaris:
4415    "xcb_conn.c", line 304: warning: implicit function declaration: shutdown
4416    
4417    Also provides system definition of SHUT_RDWR on Solaris 11.
4418    
4419    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4420    Reviewed-by: Jamey Sharp <jamey@minilop.net>
4421
4422commit 4b502dd696cf7f59a961bcf71c9255ae28f0765a
4423Author: Alan Coopersmith <alan.coopersmith@oracle.com>
4424Date:   Mon Apr 4 20:12:56 2011 -0700
4425
4426    Remove unused DECnet code
4427    
4428    "unifdef -UDNETCONN src/xcb_util.c" plus re-indenting code that was
4429    formerly in the else clause after a DECnet check.
4430    
4431    DECnet support has been removed from most of the X.Org code base for
4432    several years, and it appears DNETCONN was never defined in XCB.
4433    
4434    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4435    Reviewed-by: Jamey Sharp <jamey@minilop.net>
4436
4437commit 7131d5d0706f2b63caad13c815e893627872114c
4438Author: Alan Coopersmith <alan.coopersmith@oracle.com>
4439Date:   Mon Apr 4 16:32:45 2011 -0700
4440
4441    Use special path to sockets when running under Solaris Trusted Extensions
4442    
4443    Solaris Trusted Extensions puts the endpoints for the X server's Unix
4444    domain sockets in a special directory shared from the global zone to
4445    each of the labeled zones, since each labeled zone has a separate /tmp.
4446    
4447    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4448    Reviewed-by: Peter Harris <pharris@opentext.com>
4449
4450commit 70976d87f18d15c2ccc28eb7728e4822d3849e0d
4451Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
4452Date:   Wed Mar 23 17:47:50 2011 +0200
4453
4454    Prevent theoretical double free and leak on get_peer_sock_name.
4455    
4456    Variable new_sockname will leak and sockname will be double freed if
4457    both of the cases shown below are true.
4458    
4459    1. realloc succeeds and doesn't return the original pointer
4460    2. calling socket_func fails
4461    
4462    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4463    Signed-off-by: Erkki Seppälä <erkki.seppala@vincit.fi>
4464    Reviewed-by: Arnaud Fontaine <arnau@debian.org>
4465    Signed-off-by: Peter Harris <pharris@opentext.com>
4466
4467commit 3678159e4ed64502f9ce218a63c8d069649f2215
4468Author: Jamey Sharp <jamey@minilop.net>
4469Date:   Sat Mar 19 20:04:55 2011 -0700
4470
4471    Delete the old c-client.xsl.
4472    
4473    It hasn't been used since libxcb 1.1.90.1, released in 2008.
4474    
4475    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4476
4477commit 2415c11dec5e5adb0c17f98aa52fbb371a4f8f23
4478Author: Jamey Sharp <jamey@minilop.net>
4479Date:   Fri Mar 18 20:56:07 2011 -0700
4480
4481    Introduce xcb_wait_for_event_until, for consuming responses in wire-order.
4482    
4483    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4484    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4485
4486commit 29a974f212aae9eeff4fde99f110cee08f0312f3
4487Author: Jamey Sharp <jamey@minilop.net>
4488Date:   Fri Mar 18 17:36:32 2011 -0700
4489
4490    Dequeue readers that can't receive any new responses.
4491    
4492    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4493    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4494
4495commit 131e867fca5cda94e634af69214ad54e066ac871
4496Author: Jamey Sharp <jamey@minilop.net>
4497Date:   Fri Mar 18 15:37:34 2011 -0700
4498
4499    Factor reader_list management out of wait_for_reply.
4500    
4501    Later patches will insert reader_list entries from other entry points.
4502    
4503    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4504    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4505
4506commit 1469e879655b20351530059538a7b89612028ae2
4507Author: Jamey Sharp <jamey@minilop.net>
4508Date:   Fri Mar 18 18:18:41 2011 -0700
4509
4510    Enable AM_SILENT_RULES on automake 1.11 or newer.
4511    
4512    This incantation is supposed to be a no-op on earlier automake versions.
4513    
4514    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4515    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4516
4517commit 6310475e23eac6917db54f1425e20d8434bee679
4518Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
4519Date:   Wed Oct 13 17:48:13 2010 +0300
4520
4521    Prevent reply waiters from being blocked.
4522    
4523    It's possible to call xcb_wait_for_reply more than once for a single
4524    request. In this case we are nice and let reply waiters continue so
4525    that they can notice that the reply is not available
4526    anymore. Otherwise an event waiter could just signal the reply waiter
4527    that got its reply to continue but leave a waiter for an earlier reply
4528    blocked.
4529    
4530    Below is an example sequence for reproducing this problem.
4531    
4532    thread #1 (XNextEvent)
4533      - waits for events
4534    thread #2 (XSync)
4535      - executes request #2
4536      - waits for reply #2
4537    thread #1
4538      - reads reply #2
4539      - signals waiter of reply #2 to continue
4540      - waits for events
4541    thread #2
4542      - handles reply #2
4543    thread #3 (XCloseDisplay)
4544      - executes request #3
4545      - waits for reply #2
4546    thread #1
4547      - reads reply #3
4548      - nobody is waiting for reply #3 so don't signal
4549      - wait for events
4550    
4551    Of course it may be questionable to wait for a reply twice, but XCB
4552    should be smart enough to let clients continue if they choose to do
4553    so.
4554    
4555    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4556    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4557
4558commit 29ab5aeb9b1b1daf7f0659b134a4cfe9f42ca71a
4559Author: Jamey Sharp <jamey@minilop.net>
4560Date:   Sun Mar 13 09:41:10 2011 -0700
4561
4562    Include XKB in ./configure's summary output.
4563    
4564    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4565
4566commit 2edfd5c375bf980b484b7cfbfc1f4fb751621859
4567Merge: ed37b08 8c3325f
4568Author: Jamey Sharp <jamey@minilop.net>
4569Date:   Sun Mar 13 09:18:24 2011 -0700
4570
4571    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
4572    
4573    Apparently I forgot to push these months ago.
4574
4575commit 8c3325f8bbdb1e8975bdb01525a52d6b0f80cfa3
4576Author: Jeremy Huddleston <jeremyhu@apple.com>
4577Date:   Fri Mar 4 12:41:55 2011 -0800
4578
4579    darwin: Don't use poll() when expected to run on darwin10 and prior
4580    
4581    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4582
4583commit c2e023646298da05e212056fbc6b329e4dd9a100
4584Author: Peter Harris <pharris@opentext.com>
4585Date:   Thu Dec 23 13:32:17 2010 -0500
4586
4587    Don't try to sizeof(void)
4588    
4589    sizeof(void) is a gcc extension, and not portable.
4590    
4591    Xorg Bugzilla 31959
4592    http://bugs.freedesktop.org/show_bug.cgi?id=31959
4593    http://lists.freedesktop.org/archives/xcb/2010-May/006039.html
4594    
4595    Signed-off-by: Peter Harris <pharris@opentext.com>
4596    Tested-by: Cyril Brulebois <kibi@debian.org>
4597
4598commit 9efced72a3cb8072fa60fbed4f04d61cde412494
4599Author: Vincent Torri <vincent dot torri at gmail dot com>
4600Date:   Sun Nov 28 14:02:40 2010 +0100
4601
4602    fix Windows build and installation
4603
4604commit 3c5813697169a33ecfd6ac0ab5641dec654f6612
4605Merge: 69b78ce b672d15
4606Author: Peter Harris <pharris@opentext.com>
4607Date:   Thu Dec 23 13:04:40 2010 -0500
4608
4609    Merge branch 'master' of git://anongit.freedesktop.org/~peterh/libxcb
4610
4611commit 69b78ced1a7bcdca538c0720fde9cf3e6f70d040
4612Author: Jeetu Golani <jeetu.golani@gmail.com>
4613Date:   Sun Dec 12 16:48:41 2010 -0500
4614
4615    Don't validate FD_SETSIZE on Win32
4616    
4617    Windows' file handles have never been small or consecutive, so Windows'
4618    select has always been implemented the same way as everyone else's poll.
4619    
4620    On Windows, FD_SETSIZE is the size of the poll array, not the maximum
4621    SOCKET number.
4622    
4623    Signed-off-by: Peter Harris <git@peter.is-a-geek.org>
4624
4625commit 8ecd754b168a0352783bf1ba0f0887f7ff479ee8
4626Author: Uli Schlachter <psychon@znc.in>
4627Date:   Wed Nov 17 20:26:11 2010 +0100
4628
4629    xcb_take_socket: Document sequence wrap requirements
4630    
4631    If lots of requests are send without one causing a reply, xcb can get confused
4632    about the current sequence number of a reply. Document the requirements of an
4633    external socket owner to avoid this problem.
4634    
4635    The return_socket callback for xcb_take_socket() originally was supposed to
4636    return the last sequence number used, but the version committed to libxcb never
4637    actually had this signature. This fixes the function's documentation not to
4638    mention this non-existent return value.
4639    
4640    Signed-off-by: Uli Schlachter <psychon@znc.in>
4641    Signed-off-by: Julien Danjou <julien@danjou.info>
4642
4643commit 5755582444ad0ba79e661ab3173cc38e9e588d83
4644Author: Nick Bowler <nbowler@draconx.ca>
4645Date:   Wed Nov 10 20:49:41 2010 -0500
4646
4647    xcb_auth: Fix memory leak in _xcb_get_auth_info.
4648    
4649    If the initial get_peer_sock_name(getpeername ...) succeeds, the
4650    pointer to allocated memory is overwritten by the later call to
4651    get_peer_sock_name(getsockname ...).  Fix that up by freeing
4652    the allocated memory before overwriting the pointer.
4653    
4654    Signed-off-by: Nick Bowler <nbowler@draconx.ca>
4655    Signed-off-by: Julien Danjou <julien@danjou.info>
4656
4657commit ed37b087519ecb9e74412e4df8f8a217ab6d12a9
4658Author: Jamey Sharp <jamey@minilop.net>
4659Date:   Sat Oct 9 17:13:45 2010 -0700
4660
4661    xcb_in: Use 64-bit sequence numbers internally everywhere.
4662    
4663    Widen sequence numbers on entry to those public APIs that still take
4664    32-bit sequence numbers.
4665    
4666    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4667
4668commit 6c8b539c2a2e53bf3deb0e749a941ab52b7e8834
4669Author: Jamey Sharp <jamey@minilop.net>
4670Date:   Sat Oct 9 13:19:05 2010 -0700
4671
4672    xcb_discard_reply: Simplify by re-using poll_for_reply helper.
4673    
4674    If you discard a sequence number that has multiple responses already
4675    read, this will do more allocations than necessary. But nobody cares
4676    about ListFontsWithInfo.
4677    
4678    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4679
4680commit 3a74b5e7a1aab0619b7e34d90d2b8b2b1e386129
4681Author: Jamey Sharp <jamey@minilop.net>
4682Date:   Sat Oct 9 12:32:05 2010 -0700
4683
4684    xcb_request_check: Hold the I/O lock while deciding to sync.
4685    
4686    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4687
4688commit ee1bc1d28a1bda0526db90139edc1304d2ef3d7c
4689Author: Jamey Sharp <jamey@minilop.net>
4690Date:   Sat Oct 9 04:08:18 2010 -0700
4691
4692    xcb_send_request: Send all requests using a common internal send_request.
4693    
4694    This simplifies the critical section of xcb_send_request and fixes a
4695    couple of subtle bugs:
4696    
4697    - It's possible for xcb_send_request to need to issue two sync requests
4698      before it can issue the real request. Previously, we counted sequence
4699      numbers as if both were issued, but only one went out on the wire.
4700    
4701    - The test for whether to sync at 32-bit sequence number wrap has been
4702      incorrect since we switched to 64-bit sequence numbers internally.
4703    
4704    This change means that if the output queue was already full and the
4705    current request is bigger than the output queue, XCB will do one more
4706    write syscall than it did before. But syncs are rare and small requests
4707    are the norm, so this shouldn't be a measurable difference.
4708    
4709    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4710
4711commit b672d1514c88e119f3aaeded8c8a488cad36db52
4712Author: Peter Harris <pharris@opentext.com>
4713Date:   Wed Sep 22 23:15:38 2010 -0400
4714
4715    Fix _unserialize of reply headers
4716    
4717    This cleans up a number of warnings, and passes the sequence number
4718    through correctly.
4719    
4720    Signed-off-by: Peter Harris <pharris@opentext.com>
4721
4722commit 29cca33b9001961fa2c33bb9d9fe4a9983913fce
4723Author: Peter Harris <pharris@opentext.com>
4724Date:   Wed Sep 22 22:32:34 2010 -0400
4725
4726    Clean up a couple of warnings in xprint
4727    
4728    Signed-off-by: Peter Harris <pharris@opentext.com>
4729
4730commit 8c1d2021ca611c1452a8d2ff2a705569e4ebd056
4731Author: Peter Harris <pharris@opentext.com>
4732Date:   Wed Sep 22 21:16:51 2010 -0400
4733
4734    Make *_unserialize safe to use on buffers in-place
4735    
4736    By calling memmove instead of memcpy, and walking the buffer backward
4737    from the end, *_unserialize is safe to use in-place.
4738    
4739    Signed-off-by: Peter Harris <pharris@opentext.com>
4740
4741commit 28a71c6567d08272dc9c1c2b32f0529f11f62b9e
4742Author: Peter Harris <pharris@opentext.com>
4743Date:   Fri Sep 10 15:51:56 2010 -0400
4744
4745    Fix memory leak in _sizeof implemented with _unserialize
4746    
4747    Signed-off-by: Peter Harris <pharris@opentext.com>
4748
4749commit a22909c0f513fe28347c56be65a345831c3ce744
4750Author: Peter Harris <pharris@opentext.com>
4751Date:   Wed Sep 8 15:57:00 2010 -0400
4752
4753    Don't emit out-of-module sizeof definitions
4754    
4755    Signed-off-by: Peter Harris <pharris@opentext.com>
4756
4757commit 1c4717de367fe3bf1cf56bd8ef2bd30586bed023
4758Author: Josh Triplett <josh@joshtriplett.org>
4759Date:   Sun Sep 19 20:38:06 2010 +0200
4760
4761    Allow disconnecting connections that are in error state.
4762    
4763    In support of this, consolidate the two static error_connection
4764    definitions into one so we don't try to free the static out-of-memory
4765    error_connection.
4766    
4767    Commit by Josh Triplett and Jamey Sharp.
4768    
4769    Signed-off-by: Josh Triplett <josh@joshtriplett.org>
4770    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4771
4772commit 03bcccb132249142ba908a075e0bd5075fc20d97
4773Author: Peter Harris <pharris@opentext.com>
4774Date:   Fri Sep 10 15:53:13 2010 -0400
4775
4776    Add xkb.* to gitignore
4777    
4778    Signed-off-by: Peter Harris <pharris@opentext.com>
4779
4780commit 28d39258008fcc8ced84dc6c1dd2644e2c908c87
4781Merge: f0565e8 5e8a7ad
4782Author: Peter Harris <pharris@opentext.com>
4783Date:   Wed Sep 8 14:41:52 2010 -0400
4784
4785    Merge branch 'gsoc2010' of git://anongit.freedesktop.org/~chr/libxcb
4786
4787commit f0565e8f06aadf760a9065a97b8cf5ab9cbd18de
4788Author: Jamey Sharp <jamey@minilop.net>
4789Date:   Sat Sep 4 10:17:21 2010 -0700
4790
4791    _xcb_conn_wait: Shut down the connection on unexpected poll() events.
4792    
4793    If a client calls close(2) on the connection's file descriptor and then
4794    flushes writes, libxcb causes a hang in the client.
4795    
4796    Any flush eventually calls _xcb_out_send() with has the following loop:
4797       while(ret && *count)
4798           ret = _xcb_conn_wait(c, &c->out.cond, vector, count);
4799    
4800    _xcb_conn_wait(), if built with USE_POLL, gets the POLLNVAL error. It only
4801    checks for POLLIN and POLLOUT though, ignoring the error. Return value is 1,
4802    count is unmodified, leaving us with an endless loop and a client hang.
4803    
4804    XTS testcase Xlib3/XConnectionNumber triggers this bug. It creates a display
4805    connection, closes its file descriptor, tries to send a no-op, and then expects
4806    an error.
4807    http://cgit.freedesktop.org/xorg/test/xts/tree/xts5/Xlib3/XConnectionNumber.m
4808    
4809    If poll returned POLLHUP or POLLERR, we might see the same result.
4810    
4811    If poll returns any event we didn't ask for, this patch causes
4812    _xcb_conn_shutdown() to be invoked and an error returned. This matches the
4813    behaviour if select(2) is used instead of poll(2): select(2) returns -1 and
4814    EBADF for an already closed file descriptor.
4815    
4816    I believe this fix both is safe and will handle any similar error. POSIX says
4817    that the only bits poll is permitted to set in revents are those bits that were
4818    set in events, plus POLLHUP, POLLERR, and POLLNVAL. So if we see any flags we
4819    didn't ask for then something has gone wrong.
4820    
4821    Patch inspired by earlier proposals from Peter Hutterer and Aaron
4822    Plattner--thanks!
4823    
4824    Reported-by: Peter Hutterer <peter.hutterer@who-t.net>
4825    Reported-by: Aaron Plattner <aplattner@nvidia.com>
4826    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4827    Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
4828    Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
4829    Tested-by: Aaron Plattner <aplattner@nvidia.com>
4830    Cc: Peter Hutterer <peter.hutterer@who-t.net>
4831    Cc: Dan Nicholson <dbn.lists@gmail.com>
4832    Signed-off-by: Peter Harris <pharris@opentext.com>
4833
4834commit 20da10490f8dac75ec9fe1df28cb9e862e171be5
4835Merge: 7f5cfcc 2dcf8b0
4836Author: Peter Harris <pharris@opentext.com>
4837Date:   Tue Aug 31 18:33:36 2010 -0400
4838
4839    Merge branch 'master' of git://github.com/topcat/xcb-win32
4840    
4841    Conflicts:
4842            src/xcb_conn.c
4843            src/xcb_util.c
4844    
4845    Signed-off-by: Peter Harris <pharris@opentext.com>
4846
4847commit 7f5cfcc2fd0168d505504cc088bfdcba5c71f0ea
4848Author: Aaron Plattner <aplattner@nvidia.com>
4849Date:   Tue Aug 17 08:04:41 2010 -0700
4850
4851    xcb_disconnect: call shutdown() to force a disconnect
4852    
4853    Fixes the X Test Suite's XCloseDisplay-6 test, which has this (admittedly
4854    ridiculous) behavior:
4855    
4856     1. Create a window w.
4857     2. Open two display connections, dpy1, and dpy2.
4858     3. Grab the server using dpy1.
4859     4. Fork.
4860     5 (child). XSetProperty on w using dpy2.
4861     5 (parent). Verify that no event was recieved on dpy1.
4862     6 (parent). XCloseDisplay(dpy1).
4863     6 (child). Verify that an event was received on dpy2.
4864    
4865    It was failing because at step 6 (child), the server had not actually ungrabbed
4866    yet because the file descriptor for dpy1 was still open in the child process.
4867    
4868    Shutting down the socket during XCloseDisplay matches the behavior of non-XCB
4869    Xlib, which calls shutdown() from _X11TransSocketDisconnect.
4870    
4871    Thanks to Julien Cristau for noticing this.
4872    
4873    Signed-off-by: Aaron Plattner <aplattner at nvidia.com>
4874    Reviewed-by: Julien Cristau <jcristau@debian.org>
4875    Signed-off-by: Peter Harris <pharris@opentext.com>
4876
4877commit 2040f10a4efa95092bc9409c5b20347989b5f0a2
4878Author: Jamey Sharp <jamey@minilop.net>
4879Date:   Tue Aug 24 09:17:38 2010 -0700
4880
4881    xcb_request_check: Sync even if an event was read for this sequence.
4882    
4883    This fixes the test case I have so far for Havoc's report that
4884    xcb_request_check hangs.
4885    
4886    Rationale: Since we have a void cookie, request_expected can't have been
4887    set equal to this sequence number when the request was sent; it can only
4888    have become equal due to the arrival of an event or error. If it became
4889    equal due to an event then we still need to sync. If it became equal due
4890    to an error, then request_completed will have been updated, which means
4891    we correctly won't sync.
4892    
4893    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=29599
4894    
4895    However, Havoc reports that he can still reproduce the problem, so we
4896    may be revisiting this later.
4897    
4898    Reported-by: Havoc Pennington <hp@pobox.com>
4899    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4900
4901commit 5e8a7ade2dc8aeeeb8013785ca3f24c6057ae443
4902Author: Christoph Reimann <oss@arcor.de>
4903Date:   Mon Aug 16 20:24:40 2010 +0200
4904
4905    small fix to get rid of some compiler warnings
4906    also added very basic documentation for xkb
4907
4908commit b89f634ff9b321f21874cd45e398d661a6ff726e
4909Author: Christoph Reimann <oss@arcor.de>
4910Date:   Mon Aug 16 18:22:42 2010 +0200
4911
4912    small fix in the xkb pkg file
4913
4914commit 22e1013131984a217e9bddeac3a6a4183e35f0c1
4915Author: Christoph Reimann <oss@arcor.de>
4916Date:   Mon Aug 16 18:19:16 2010 +0200
4917
4918    added accessors for special cases
4919    major bugfixes include: rewrite of prefix related functions, merge of serialize/unserialize/... generators, extended field name resolution
4920
4921commit 35f901a0f2bc3f5bb30dc6ff9d791679c9e84c05
4922Author: Julien Danjou <julien@danjou.info>
4923Date:   Fri Aug 13 13:46:37 2010 +0200
4924
4925    Release libxcb 1.7
4926    
4927    Signed-off-by: Julien Danjou <julien@danjou.info>
4928
4929commit fe0e32b5fa3923fae97210e974c0f96a085116cb
4930Author: Christoph Reimann <oss@arcor.de>
4931Date:   Sun Aug 8 21:25:13 2010 +0200
4932
4933    special case 'intermixed variable and fixed size fields': fixed reply side, needs testing
4934
4935commit 77b594f9583ea0247ff27130316d8e045da7f921
4936Author: Christoph Reimann <oss@arcor.de>
4937Date:   Thu Aug 5 15:55:28 2010 +0200
4938
4939    renamed most _unserialize() functions to _sizeof() and fixed _unserialize() for the special case of intermixed variable and fixed size fields
4940
4941commit dd1a4dbe20d6b5fd33aeb65e662bb2ca18665518
4942Author: Eamon Walsh <efw@eamonwalsh.com>
4943Date:   Thu Aug 5 00:48:08 2010 -0400
4944
4945    Tutorial uses wrong function.
4946    
4947    https://bugs.freedesktop.org/show_bug.cgi?id=29392
4948    
4949    Signed-off-by: Eamon Walsh <efw@eamonwalsh.com>
4950
4951commit b187f029d6bb693f0294bad5261ec486b140f185
4952Author: Christoph Reimann <oss@arcor.de>
4953Date:   Mon Aug 2 23:30:42 2010 +0200
4954
4955    attempt to fix special case: variable fields followed by fixed size fields
4956
4957commit a700eeb502b5fe902e4c93cc707100ec868ce946
4958Author: Christoph Reimann <oss@arcor.de>
4959Date:   Sun Aug 1 23:40:20 2010 +0200
4960
4961    bug fixes for all kinds of 'special cases'
4962
4963commit 1c590d5a86ae854e53f388e40c952e92f11d59e6
4964Author: Christoph Reimann <oss@arcor.de>
4965Date:   Thu Jul 22 16:41:15 2010 +0200
4966
4967    partial rewrite of serialize helper functions completed;
4968    _serialize() & _unserialize() have been tested for switch derived from valueparam
4969
4970commit 566ae9baee20fb6147b94b89a26796087461bae8
4971Author: Christoph Reimann <oss@arcor.de>
4972Date:   Tue Jul 20 22:46:37 2010 +0200
4973
4974    preliminary handling of further special cases in unserialize
4975    first attempts to unify serialize and unserialize
4976
4977commit 4e665e1580ece7bc9505f3a2f657959669ffcd05
4978Author: Christoph Reimann <oss@arcor.de>
4979Date:   Thu Jul 15 18:43:43 2010 +0200
4980
4981    added generating code for _serialize() in case of variable sized structs (largely untested)
4982
4983commit d0031456097f009bdb82fb979924e71ca38c539b
4984Author: Christoph Reimann <oss@arcor.de>
4985Date:   Tue Jul 13 20:08:51 2010 +0200
4986
4987    xkb: added pkg config file
4988
4989commit 86704124b1fd62c30441ace1f3f8e2c316801c53
4990Author: Christoph Reimann <oss@arcor.de>
4991Date:   Tue Jul 13 20:06:08 2010 +0200
4992
4993    new and still preliminary functions for switch; feautures include
4994    - API compatibility with valueparam
4995    - request _aux() auxiliary functions
4996    - _serialize() and _unserialize() auxiliary functions
4997    - new data type that allows mixing of fixed and variable size members
4998
4999commit 80322d11636dd638902660d80481080d2fad40fe
5000Author: Christoph Reimann <oss@arcor.de>
5001Date:   Tue Jul 13 19:59:23 2010 +0200
5002
5003    xkb: updated configure.ac/Makefile.am
5004
5005commit 8c2707773b3621fb8bbda9021d23944f5be34aab
5006Author: Christoph Reimann <oss@arcor.de>
5007Date:   Tue Jul 13 19:56:44 2010 +0200
5008
5009    added xcb_sumof() with restriction to uint8_t
5010
5011commit 496efb7624d14b4ca0391f44926edc448cbd605e
5012Author: Jamey Sharp <jamey@minilop.net>
5013Date:   Tue Jul 13 07:01:06 2010 -0700
5014
5015    _xcb_conn_wait: Fix whitespace.
5016    
5017    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5018
5019commit 74057c7eb6836353960ce3849703ce20e45089bc
5020Author: Jeremy Huddleston <jeremyhu@apple.com>
5021Date:   Mon Jul 12 16:53:53 2010 -0700
5022
5023    AC_CHECK_PROG(LAUNCHD, [launchd], [yes], [no], [$PATH$PATH_SEPARATOR/sbin])
5024    
5025    launchd: Explicitly search /sbin
5026    
5027    Previously, launchd wasn't found if /sbin wasn't in the user's PATH.
5028    https://bugs.freedesktop.org/show_bug.cgi?id=29028
5029    
5030    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
5031
5032commit 75ff427d41fc10d00b780c965c289fc02c81aaac
5033Author: Vincent Torri <vtorri@univ-evry.fr>
5034Date:   Sun Jun 20 20:50:06 2010 -0700
5035
5036    configure.ac: Report which extensions are being built.
5037    
5038    I was surprised to see that xinput was not installed. Looking at
5039    configure.ac, it seems that it is disabled by default. Maybe configure
5040    should output the status of the different extensions.
5041
5042commit de3cdad87a341c238c544425f4dd574b19c58ae3
5043Author: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
5044Date:   Fri Jun 11 16:30:47 2010 +0300
5045
5046    xcb_connect_to_display_with_auth_info: Fix memory leak
5047    
5048    protocol and host are allocated in _xcb_parse_display but ownership of
5049    them is passed to the caller. They have to be freed in
5050    xcb_connect_to_display_with_auth_info.
5051    
5052    Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
5053    Signed-off-by: Peter Harris <pharris@opentext.com>
5054
5055commit 18718d483e0982c779a61c71176fb0e64f850015
5056Author: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
5057Date:   Fri Jun 11 16:30:46 2010 +0300
5058
5059    _xcb_parse_display: Fix error path
5060    
5061    xcb_parse_display claims that there is no side effects when failing.
5062    That requires _xcb_parse_display to free the memory in failure case.
5063    
5064    Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
5065    Signed-off-by: Peter Harris <pharris@opentext.com>
5066
5067commit 3f79628becbd3b0eff1aef804902eb739fac4403
5068Author: Jeremy Huddleston <jeremyhu@apple.com>
5069Date:   Wed May 12 19:53:45 2010 -0700
5070
5071    xcb_open: Improve protocol/host parsing
5072    
5073    Support scenarios where host is not set and protocol is.  eg:
5074    
5075    DISPLAY=tcp/:0
5076    
5077    as well as the "inet" and "inet6" alias for "tcp" for compatability
5078    with Xlib
5079    
5080    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
5081    Reviewed-by: Jamey Sharp <jamey@minilop.net>
5082
5083commit e4b746ac13e89b99abd80b3d1fa2a16796da3b6d
5084Author: Marcin Kościelnicki <koriakin@0x04.net>
5085Date:   Thu May 13 21:05:57 2010 +0000
5086
5087    Add ~ operator support in code generator
5088    
5089    Reviewed-by: Julien Cristau <jcristau@debian.org>
5090    Signed-off-by: Julien Danjou <julien@danjou.info>
5091
5092commit 5e86cb05666c448de2f61c23ae94e94ef4b38d64
5093Author: Arnaud Fontaine <arnau@debian.org>
5094Date:   Fri Apr 30 18:49:18 2010 +0100
5095
5096    Fix GCC error on undeclared variable when not using abstract socket
5097    
5098    This is a regression found by tinderbox in previous commit:
5099    
5100    xcb_util.c: In function '_xcb_open':
5101    xcb_util.c:213: error: 'fd' undeclared (first use in this function)
5102
5103commit a546d00091de0ab16374dec55e8e2fa87d6bbebf
5104Author: Arnaud Fontaine <arnau@debian.org>
5105Date:   Fri Apr 30 14:47:16 2010 +0200
5106
5107    Get rid of PATH_MAX and MAXPATHLEN
5108    
5109    There could be no upper limit on the length of a path according
5110    to POSIX, therefore these macros may not be defined at all on
5111    some systems (such as GNU Hurd).
5112    
5113    Signed-off-by: Arnaud Fontaine <arnau@debian.org>
5114    Reviewed-by: Peter Harris <pharris@opentext.com>
5115
5116commit d06857217328c2283a8956788d72646fc67216fb
5117Author: Jeremy Huddleston <jeremyhu@apple.com>
5118Date:   Fri Apr 23 21:57:26 2010 -0700
5119
5120    Use limits.h instead of syslimits.h
5121    
5122    Regression found by tinderbox in 89b3485dadef47a30264a5bf150b96522183376b
5123    
5124    xcb_util.c:31:27: error: sys/syslimits.h: No such file or directory
5125    xcb_util.c: In function '_xcb_open':
5126    xcb_util.c:148: error: 'PATH_MAX' undeclared (first use in this function)
5127    
5128    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
5129
5130commit 89b3485dadef47a30264a5bf150b96522183376b
5131Author: Jeremy Huddleston <jeremyhu@apple.com>
5132Date:   Fri Apr 23 17:29:25 2010 -0700
5133
5134    Reworked launchd support to work better with _xcb_parse_display
5135    
5136    Fixes: http://xquartz.macosforge.org/trac/ticket/390
5137    
5138    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
5139
5140commit 2dcf8b025be88a25d4333abdc28d425b88238d96
5141Author: Jeetu Golani <jeetu.golani@gmail.com>
5142Date:   Fri Apr 23 00:47:16 2010 +0530
5143
5144    Replaced read() in read_block and _xcb_in_read() with recv for all
5145    platforms. MSG_WAITALL is undefined in MinGW so it's been explicitly
5146    defined in xcb_in.c
5147
5148commit 56962e42a509dc4d0d9541e46b93689dac61c4fd
5149Author: Jeetu Golani <jeetu.golani@gmail.com>
5150Date:   Thu Apr 22 23:23:27 2010 +0530
5151
5152    Set errno=0 in read_block. On Win32 there is no errno and this makes the
5153    do..while loop execute only once. Also set the return value to -1 in
5154    _xcb_open if control reaches the end - if all goes well it shouldn't
5155    reach there.
5156
5157commit b0525e242368fffbc77ebb45293f34e80847e65a
5158Author: Jamey Sharp <jamey@minilop.net>
5159Date:   Sat Apr 17 17:59:11 2010 -0700
5160
5161    Always wake up readers after writing.
5162    
5163    Since writers must make sure they read as well, threads may have gone to
5164    sleep waiting for the opportunity to read. The writer must wake up one
5165    of those readers or the application can hang.
5166    
5167    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5168    Reviewed-by: Josh Triplett <josh@freedesktop.org>
5169
5170commit eff3851ba80c42b5b3ba240f7e9049d7b0fac6f0
5171Author: Jamey Sharp <jamey@minilop.net>
5172Date:   Sun Mar 28 10:31:55 2010 -0700
5173
5174    Fix strict-aliasing warning when getting generic event length.
5175    
5176    xcb_ge_event_t has its length field in the same place that
5177    xcb_generic_reply_t does, so there's no need to cast the generic reply.
5178    
5179    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5180    Cc: Peter Hutterer <peter.hutterer@who-t.net>
5181    Reviewed-by: Julien Danjou <julien@danjou.info>
5182
5183commit 6dd8228a137d280ce24cec604a419129d8ed0e8e
5184Author: Jamey Sharp <jamey@minilop.net>
5185Date:   Fri Feb 12 12:25:05 2010 -0800
5186
5187    Delete a useless level of indirection from _xcb_out_send's parameters.
5188    
5189    _xcb_out_send needs _xcb_conn_wait to store back its progress so it can
5190    be reinvoked to pick up where it left off---but then _xcb_out_send
5191    guarantees that it leaves either an empty output vector or a shut-down
5192    connection, so *its* callers never care how much progress was made.
5193    
5194    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5195    Reviewed-by: Josh Triplett <josh@freedesktop.org>
5196
5197commit a63fbc9d6c484e5ad7a5e9d56f81b8e1a2f38a82
5198Author: Julien Danjou <julien@danjou.info>
5199Date:   Fri Apr 9 16:57:51 2010 +0200
5200
5201    Release libxcb 1.6
5202
5203commit a1d9aa6e07a297f4108b4ad787336f74c86a3312
5204Author: Rémi Denis-Courmont <remi.denis-courmont@nokia.com>
5205Date:   Fri Mar 26 23:12:47 2010 +0200
5206
5207    Fail if fd is too big and poll() is not available
5208    
5209    Depending on the process file limit, a file descriptor can be larger
5210    than the capacity of fd_set. There is no portable way to create a
5211    large enough fd_set at run-time. So we just fail if the file descriptor
5212    number is too high and poll() is not available.
5213    
5214    Reviewed-by: Jamey Sharp <jamey@minilop.net>
5215    Signed-off-by: Julien Danjou <julien@danjou.info>
5216
5217commit 0e0c80e749eccf121e55c1e855c48d03b54f33ef
5218Author: Jeetu Golani <jeetu.golani@gmail.com>
5219Date:   Wed Mar 31 22:00:04 2010 +0530
5220
5221    xcb_in.c #ifndef _WIN32 inside of #if USE_POLL redundant and removed
5222
5223commit e8009194c9f5a6995c4a9b03a7a49d5bc09e96fc
5224Author: Jeetu Golani <jeetu.golani@gmail.com>
5225Date:   Wed Mar 31 09:50:51 2010 +0530
5226
5227    restablished inclusion of fcntl.h and netinet/tcp.h in xcb_util.c -- without these the code no longer compiled on *ix
5228
5229commit d302f1e9b158d0a51936c28e5dc66251d90d1d56
5230Author: Jeetu Golani <jeetu.golani@gmail.com>
5231Date:   Mon Mar 29 22:37:33 2010 +0530
5232
5233    changes in xcb_windefs.h - the flag  _XCB_WINDEFS_H replaces WINDEFS_H
5234
5235commit 36c9a985aaee655c118c9f7b8425d3ac9ce0f840
5236Author: Jeetu Golani <jeetu.golani@gmail.com>
5237Date:   Mon Mar 29 22:31:49 2010 +0530
5238
5239    windefs.h is now called xcb_windefs.h - changed all includes to reflect that.Replaced one instance ofWIN32 with _WIN32 in each xcb_in.c and xcb_conn.c
5240
5241commit bce72f63d2dfb61661f81e305ad3a7db0334403c
5242Author: Jeetu Golani <jeetu.golani@gmail.com>
5243Date:   Fri Mar 26 09:40:09 2010 +0530
5244
5245    Win32 code for xcb-1.5
5246
5247commit d18d03d6f37ec220805855d840950716e22354e4
5248Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
5249Date:   Fri Mar 12 23:51:32 2010 +0100
5250
5251    Fix authentication on hpux and Hurd
5252    
5253    libxcb's 010e5661 (Fix XDM-AUTHORIZATION-1 (bug #14202)) mistakenly
5254    inverted a few lines of code, making local socket authentication fail on
5255    hpux and Hurd: when getpeername fails, sockname needs to be initialized
5256    by getsockname before its address family can be checked.
5257    
5258    Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
5259    Signed-off-by: Julien Danjou <julien@danjou.info>
5260
5261commit 53a9834e4cdd11aba8c1cc49347f09a958107de8
5262Author: Rémi Denis-Courmont <remi@remlab.net>
5263Date:   Sat Feb 13 12:23:51 2010 +0200
5264
5265    Open the X11 socket with close-on-exec flag
5266    
5267    This saves the X11 connection from leaking into children processes.
5268    On Linux, this is fully thread-safe using SOCK_CLOEXEC. On other
5269    systems, there is a small race condition.
5270    
5271    Signed-off-by: Julien Danjou <julien@danjou.info>
5272
5273commit 367882fa32489ebafcd9afc04fbf732b02ceb33a
5274Author: Peter Harris <pharris@opentext.com>
5275Date:   Tue Feb 9 16:27:22 2010 -0500
5276
5277    Support xcb_discard_reply
5278    
5279    This function is useful for dynamic language garbage collectors. Frequently
5280    a GC cycle may run before you want to block wainting for a reply.
5281    
5282    This function is also marginally useful for libxcb apps that issue
5283    speculative requests (eg. xlsclients).
5284    
5285    Reviewed-by: Jamey Sharp <jamey@minilop.net>
5286    Tested-by: Eamon Walsh <efw@eamonwalsh.com>
5287    Signed-off-by: Peter Harris <pharris@opentext.com>
5288
5289commit be7e528eae62ddee14fa50f2c0e9036bafbc9f81
5290Author: Rémi Denis-Courmont <remi@remlab.net>
5291Date:   Thu Jan 7 18:08:35 2010 +0200
5292
5293    xcb_connect_to_fd: fix descriptor leak on memory error path
5294    
5295    Signed-off-by: Julien Danjou <julien@danjou.info>
5296
5297commit c7a57043da1717c18703a38772555fea6aa811bf
5298Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5299Date:   Fri Dec 11 17:24:43 2009 -0800
5300
5301    configure.ac: Fix a typo on the last commit.
5302    
5303    Signed-off-by: Jeremy Huddleston <jeremyhu@freedesktop.org>
5304
5305commit 58c96da9283d3c2f65c818bd952f611ec4e4dad0
5306Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5307Date:   Fri Dec 11 17:15:16 2009 -0800
5308
5309    darwin: xnu doesn't support poll on ttys on the master side.
5310    
5311    <rdar://problem/7360546>
5312    
5313    Signed-off-by: Jeremy Huddleston <jeremyhu@freedesktop.org>
5314
5315commit 1e746fd89ae93965183c759b969ff9f4d5dbc9d8
5316Author: Jim Ingram <ingramj@gmail.com>
5317Date:   Mon Dec 7 14:41:18 2009 -0600
5318
5319    Tutorial code fix
5320    
5321    Accessed elements of names[] after freeing them in the first example.
5322    
5323    Signed-off-by: Julien Danjou <julien@danjou.info>
5324
5325commit a470579ba29c8f39e77668558a08bb173297711f
5326Author: Rémi Cardona <remi@gentoo.org>
5327Date:   Fri Dec 4 22:06:20 2009 +0100
5328
5329    DRI2 extension requires xcb-proto 1.6 or newer
5330    
5331    Signed-off-by: Rémi Cardona <remi@gentoo.org>
5332    Signed-off-by: Julien Danjou <julien@danjou.info>
5333
5334commit 6e875a82a8a1a324067970a5f9d7442585ee0eef
5335Author: Julien Danjou <julien@danjou.info>
5336Date:   Thu Dec 3 11:34:18 2009 +0100
5337
5338    build: simplify extension building
5339    
5340    Signed-off-by: Julien Danjou <julien@danjou.info>
5341
5342commit 48217ac986d77eba40f3af4df597a734f4fd3690
5343Author: Julien Danjou <julien@danjou.info>
5344Date:   Thu Dec 3 10:08:59 2009 +0100
5345
5346    Release libxcb 1.5
5347    
5348    Signed-off-by: Julien Danjou <julien@danjou.info>
5349
5350commit 1cf2a87def76f4646fe05e282b45605d572f2513
5351Author: Adam Jackson <ajax@redhat.com>
5352Date:   Wed Dec 2 14:31:56 2009 -0500
5353
5354    setsockopt(SO_KEEPALIVE) on TCP display connections.
5355    
5356    This matches xtrans behaviour in SocketINETConnect, and makes it so apps
5357    don't hang forever if their display dies.
5358    
5359    Signed-off-by: Adam Jackson <ajax@redhat.com>
5360
5361commit e82c34c2f7ac3fbb23ab14cbee8df2dd9178f5a4
5362Author: Eric Anholt <eric@anholt.net>
5363Date:   Thu Oct 15 13:13:47 2009 -0700
5364
5365    Add DRI2 support. (v2)
5366    
5367    v2: Build fix from jcristau.
5368
5369commit a27c77ccae059fb64e0506648b81677858d3b05b
5370Author: Eric Anholt <eric@anholt.net>
5371Date:   Thu Oct 15 13:14:49 2009 -0700
5372
5373    Fix typo in the tutorial.
5374
5375commit 17af34b4a438a80d3dcad93bb5254366791a488f
5376Author: Paulo R. Zanoni <pzanoni@mandriva.com>
5377Date:   Fri Sep 18 15:05:46 2009 -0300
5378
5379    Don't build docs if 'dot' is not found Signed-off-by: Paulo R. Zanoni <pzanoni@mandriva.com>
5380
5381commit 29207e2943ad56fb8b4d2b7f07b1530cb1d7d9be
5382Author: Julien Cristau <jcristau@debian.org>
5383Date:   Mon Aug 31 17:51:36 2009 +0200
5384
5385    Fix check dependency
5386    
5387    Bugzilla #21992
5388    
5389    make -j check fails because the check-local rule gets executed before
5390    the tests actually ran, so CheckLog*.xml doesn't exist.
5391    
5392    Signed-off-by: Julien Danjou <julien@danjou.info>
5393
5394commit ff2e2e35bcac1d835c9eecf9ac8b3072005989a5
5395Author: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
5396Date:   Sat Jun 6 15:52:20 2009 +0100
5397
5398    Cygwin build fix: Add -no-undefined to libtool flags
5399    
5400    -no-undefined is needed to tell libtool a shared library can be built
5401    on platforms which require all references to be statisfied at link time.
5402    
5403    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
5404    Signed-off-by: Julien Danjou <julien@danjou.info>
5405
5406commit 96ff5fb635bc333e80ebce6cc03008a1b9992ff6
5407Author: Arnaud Fontaine <arnau@debian.org>
5408Date:   Wed Jul 15 16:15:02 2009 +0100
5409
5410    Release libxcb 1.4
5411
5412commit f4c2794bf5990a0b2f6168f2b22b60b15e08ac44
5413Author: Arnaud Fontaine <arnau@debian.org>
5414Date:   Wed Jul 15 16:03:56 2009 +0100
5415
5416    Add majorCode, minorCode and resourceID fields to X generic error
5417
5418commit e06955ed66cb499ada52b8add6709edd38d70054
5419Author: Jamey Sharp <jamey@minilop.net>
5420Date:   Mon Jul 6 13:14:35 2009 -0700
5421
5422    Fix precedence bug: wrong length for big-requests preceded by sync.
5423    
5424    Also replace excessively clever use of bitwise OR with equivalent
5425    addition.
5426    
5427    Reported-by: Geoffrey Li <geoffrey@seitopos.com>
5428    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5429    Signed-off-by: Josh Triplett <josh@joshtriplett.org>
5430
5431commit 9e191c722a0be4b5f3d993165055a71f85c21882
5432Author: Julien Cristau <jcristau@debian.org>
5433Date:   Fri May 29 14:41:59 2009 +0200
5434
5435    Fix libxcb-randr version info
5436    
5437    The SONAME shouldn't have been bumped in 1.3, only new symbols were
5438    added.
5439    
5440    Signed-off-by: Julien Danjou <julien@danjou.info>
5441
5442commit f44dc519c5b324335d05f6e7fb31b78062a8c5f5
5443Author: Julien Danjou <julien@danjou.info>
5444Date:   Fri May 29 08:26:05 2009 +0200
5445
5446    Release libxcb 1.3
5447
5448commit ee89850e68205a7f8961ace0839b5be86040dade
5449Author: elupus <elupus@ecce.se>
5450Date:   Tue May 26 16:14:48 2009 +0200
5451
5452    Disable Nagle on TCP socket
5453    
5454    Signed-off-by: Julien Danjou <julien@danjou.info>
5455
5456commit 62fe187e2d617eb0feb1ca03d8b4a64db9dd952b
5457Author: Bob Ham <rah@bash.sh>
5458Date:   Mon May 25 12:20:23 2009 +0200
5459
5460    Store xcbproto version libxcb was compiled with
5461    
5462    Signed-off-by: Julien Danjou <julien@danjou.info>
5463
5464commit efbe96ee0ab0c5511035eee99f8fe7b38d5f65fa
5465Author: Julien Danjou <julien@danjou.info>
5466Date:   Sat May 9 17:39:34 2009 +0200
5467
5468    depends on recent xcb-proto and bump version of randr
5469    
5470    Signed-off-by: Julien Danjou <julien@danjou.info>
5471
5472commit 010e566126625c56bdf9989085bacf731520ff87
5473Author: Bart Massey <bart@cs.pdx.edu>
5474Date:   Tue Apr 21 08:39:52 2009 +0200
5475
5476    Fix XDM-AUTHORIZATION-1 (bug #14202)
5477    
5478    With this patch, we know use correctly the socket address or peer
5479    address for authentication purpose.
5480    
5481    Signed-off-by: Julien Danjou <julien@danjou.info>
5482
5483commit ca978a9dae621126075712f9e2c29591208570bc
5484Author: Julien Danjou <julien@danjou.info>
5485Date:   Tue Apr 7 14:22:57 2009 +0200
5486
5487    util: remove useless strlen calls from decnet opening
5488    
5489    Signed-off-by: Julien Danjou <julien@danjou.info>
5490
5491commit cc191431412d8764c645a51b0f106c0dfe652213
5492Author: Julien Danjou <julien@danjou.info>
5493Date:   Tue Apr 7 14:18:40 2009 +0200
5494
5495    util: merge common code for xcb_connect
5496    
5497    Many code was duplicated between xcb_connect_to_display_with_auth_info
5498    and xcb_connect(). We merge both, since the difference is just about the
5499    xcb_auth_info_t pointer being supplied, or not.
5500    
5501    Signed-off-by: Julien Danjou <julien@danjou.info>
5502
5503commit 8797e053b2b2ee989f47490c7687b9a2fbdb0021
5504Author: Julien Danjou <julien@danjou.info>
5505Date:   Tue Apr 7 13:37:40 2009 +0200
5506
5507    util: open_abstract gets filelen as parameters
5508    
5509    That saves us from a couple of strlen() calls.
5510    
5511    Signed-off-by: Julien Danjou <julien@danjou.info>
5512
5513commit f0b29819749b769e5a8d313bf1bab80d6513208b
5514Author: Julien Danjou <julien@danjou.info>
5515Date:   Tue Apr 7 11:55:30 2009 +0200
5516
5517    auth: use snprintf() return value
5518    
5519    That save us from a strlen().
5520    
5521    Signed-off-by: Julien Danjou <julien@danjou.info>
5522
5523commit 9f24c91f91dd68a52e46191b686283b0df38d2f5
5524Author: Julien Danjou <julien@danjou.info>
5525Date:   Tue Apr 7 11:49:13 2009 +0200
5526
5527    auth: precompute authnameslen
5528    
5529    Signed-off-by: Julien Danjou <julien@danjou.info>
5530
5531commit 2a4f1cf3801080276694f5026d35220b65201038
5532Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5533Date:   Thu Apr 9 05:12:02 2009 -0700
5534
5535    darwin: Don't use poll() on versions of darwin before darwin10
5536
5537commit 6e2e87d0bbdff87f127986a0666445160d52e6a5
5538Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
5539Date:   Mon Apr 6 03:31:23 2009 +0200
5540
5541    Local socket connections do not work on hurd-i386
5542    
5543    Local socket connections currently do not work on hurd-i386 because
5544    xcb_auth calls getpeername() on the client socket, but hurd-i386 does
5545    not implement anything in that case (I actually wonder what reasonable
5546    value could be returned).  In such case the xcb code does not actually
5547    need the peer name anyway.
5548    
5549    Signed-off-by: Julien Danjou <julien@danjou.info>
5550
5551commit f916062edf0e04cd4e0a78f6975892f59fae3b60
5552Author: Michael Ost <most@museresearch.com>
5553Date:   Mon Mar 30 11:09:32 2009 +0200
5554
5555    use poll() instead of select() when available
5556    
5557    Signed-off-by: Julien Danjou <julien@danjou.info>
5558
5559commit beccb0be15f5699c942a0af33307d9e4bf797e2a
5560Author: Bart Massey <bart@cs.pdx.edu>
5561Date:   Tue Mar 24 16:24:04 2009 -0700
5562
5563    kludgily hand-merged xid fixes
5564    
5565    Signed-off-by: Bart Massey <bart@cs.pdx.edu>
5566    Signed-off-by: Julien Danjou <julien@danjou.info>
5567
5568commit 1e9c0f1012b6d349f92ea7246194b8667d27f849
5569Author: Julien Danjou <julien@danjou.info>
5570Date:   Mon Mar 16 10:26:02 2009 +0100
5571
5572    Fix do_append() arguments
5573    
5574    Signed-off-by: Julien Danjou <julien@danjou.info>
5575
5576commit fcb433db80315a44154248a9229c9cfcbab63f04
5577Author: Julien Danjou <julien@danjou.info>
5578Date:   Sun Mar 15 10:18:50 2009 +0100
5579
5580    Copy full IPv4 mapping (Bug #20665)
5581    
5582    Signed-off-by: Julien Danjou <julien@danjou.info>
5583
5584commit eaa71eac02c6a862ab23e8afcce12d9f38590338
5585Author: Peter Harris <pharris@opentext.com>
5586Date:   Fri Mar 13 15:24:55 2009 -0400
5587
5588    Avoid name collisions between xidtype and enum.
5589    
5590    These changes are necessary to build with latest xcb/proto.
5591    
5592    Signed-off-by: Peter Harris <pharris@opentext.com>
5593
5594commit e986d1ee5a126dc38113125075a1e986235ba7c7
5595Author: Peter Harris <pharris@opentext.com>
5596Date:   Fri Mar 13 15:25:30 2009 -0400
5597
5598    Revert "Don't use enums in generated C code"
5599    
5600    This commit broke xcb/util.
5601    
5602    This reverts commit 9984b72888108a038d6b3f7dee374d17e26ef9e2.
5603    
5604    Signed-off-by: Peter Harris <pharris@opentext.com>
5605
5606commit 9984b72888108a038d6b3f7dee374d17e26ef9e2
5607Author: Peter Harris <pharris@opentext.com>
5608Date:   Wed Feb 25 18:48:50 2009 -0500
5609
5610    Don't use enums in generated C code - use integer constants instead.
5611    
5612    Signed-off-by: Peter Harris <pharris@opentext.com>
5613
5614commit b08e1535cf0716fc917eaa1f5d5f6d1960bf1e3c
5615Author: Julien Danjou <julien@danjou.info>
5616Date:   Tue Feb 17 13:37:29 2009 +0100
5617
5618    Release libxcb 1.2
5619    
5620    Signed-off-by: Julien Danjou <julien@danjou.info>
5621
5622commit 2e65188f91800d7e6a8d74fa077b46f8f67b6893
5623Author: Julien Danjou <julien@danjou.info>
5624Date:   Mon Feb 16 11:44:20 2009 +0100
5625
5626    Stop providing autogenerated files in tarball
5627    
5628    Signed-off-by: Julien Danjou <julien@danjou.info>
5629
5630commit 37f5ce3ef46c8af38ec0103b9d960615a0a1e058
5631Author: Julien Danjou <julien@danjou.info>
5632Date:   Thu Dec 11 11:17:13 2008 +0100
5633
5634    Release libxcb 1.1.93
5635    
5636    Signed-off-by: Julien Danjou <julien@danjou.info>
5637
5638commit f896ae8c53e99f39b347f9f0ac2b4a8cc12cad6f
5639Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5640Date:   Sun Nov 23 17:25:21 2008 -0800
5641
5642    Shutup compiler warning about unused variable...
5643
5644commit d79621b25ba6784135b1a3aa51e9561fcf72fe7f
5645Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5646Date:   Sun Nov 23 17:23:17 2008 -0800
5647
5648    Apple: Apple launchd cleanup
5649    Added --with-launchd option instead of just using __APPLE__
5650    Fixed opening launchd fd when displayname=NULL
5651
5652commit 9b79ae49f709707e99b8487b01f9d3f102754bd3
5653Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5654Date:   Sat Nov 8 14:41:23 2008 -0800
5655
5656    Apple: Enable support for launchd DISPLAY socket
5657
5658commit 43b612a5e99ccdfb787a11b2b7c3179fc242edf3
5659Author: Peter Harris <peter.harris@hummingbird.com>
5660Date:   Wed Nov 12 14:45:04 2008 -0500
5661
5662    Treat XIDs the same as other cardinal values.
5663    
5664    This fixes a bug where c_client.py wasn't generating *_end functions,
5665    but expected them to exist in order to find the subsequent list's start.
5666    
5667    Signed-off-by: Peter Harris <peter.harris@hummingbird.com>
5668
5669commit 902cade8ec9e7eb5d29f6f6d61cac0470c300aec
5670Author: Julien Danjou <julien@danjou.info>
5671Date:   Tue Nov 4 10:19:14 2008 +0100
5672
5673    Release libxcb 1.1.92
5674    
5675    Signed-off-by: Julien Danjou <julien@danjou.info>
5676
5677commit fa452cc9b2bb69fa0603dfd97e00e540b6b52840
5678Author: Josh Triplett <josh@freedesktop.org>
5679Date:   Sun Mar 16 23:16:31 2008 -0700
5680
5681    Support handing off socket write permission to external code.
5682    
5683    Libraries like Xlib, some XCB language bindings, and potentially others
5684    have a common problem: they want to share the X connection with XCB. This
5685    requires coordination of request sequence numbers.  Previously, XCB had an
5686    Xlib-specific lock, and allowed Xlib to block XCB from making requests.
5687    Now we've replaced that lock with a handoff mechanism, xcb_take_socket,
5688    allowing external code to ask XCB for permission to take over the write
5689    side of the socket and send raw data with xcb_writev.  The caller of
5690    xcb_take_socket must supply a callback which XCB can call when it wants
5691    the write side of the socket back to make a request.  This callback
5692    synchronizes with the external socket owner, flushes any output queues if
5693    appropriate, and then returns the sequence number of the last request sent
5694    over the socket.
5695    
5696    Commit by Josh Triplett and Jamey Sharp.
5697    Handoff mechanism inspired by Keith Packard.
5698
5699commit baff35a04b0e8d21821850a405a550d86a8aeb6f
5700Author: Jamey Sharp <jamey@minilop.net>
5701Date:   Wed May 21 14:44:16 2008 -0700
5702
5703    Track 64-bit sequence numbers internally.
5704    
5705    External APIs that used 32-bit sequence numbers continue to do so.
5706    
5707    Commit by Josh Triplett and Jamey Sharp.
5708
5709commit 96e55444b9b9500420f9132a1ace720100a26398
5710Author: Jamey Sharp <jamey@minilop.net>
5711Date:   Fri Mar 14 20:18:52 2008 -0700
5712
5713    Use sequence number ranges in pending replies
5714    
5715    This allows optimizing adjacent pending replies with the same flags, and
5716    will help support default flags for a range of future requests.
5717    
5718    Commit by Josh Triplett and Jamey Sharp.
5719
5720commit 059ca642c76639fee958dc6054070de85e257e98
5721Author: Jamey Sharp <jamey@minilop.net>
5722Date:   Fri Mar 14 12:08:58 2008 -0700
5723
5724    Inline _xcb_lock_io, _xcb_unlock_io, and _xcb_wait_io.
5725    
5726    These functions are once again a single pthread call, so just make that
5727    call directly.
5728
5729commit d989656cde2ee7a4a66b2065209ef389495f3452
5730Author: Jamey Sharp <jamey@minilop.net>
5731Date:   Fri Mar 14 12:08:32 2008 -0700
5732
5733    Remove libxcb-xlib and xcbxlib.h.
5734
5735commit dcbef23d730b95ef7dc9ef524a4c3fc0017b63d2
5736Author: Julien Danjou <julien@danjou.info>
5737Date:   Tue Oct 14 23:39:07 2008 +0200
5738
5739    build: fix configure.ac AC_DEFINE
5740    
5741    Rather use AC_DEFINE_UNQUOTED and only once.
5742    
5743    Signed-off-by: Julien Danjou <julien@danjou.info>
5744
5745commit cebd482a20fcc2b2dae0683c38e917a3740638a6
5746Author: Julien Danjou <julien@danjou.info>
5747Date:   Fri Sep 26 15:26:42 2008 +0200
5748
5749    allow compile-time setting for XCB queue buffer size
5750    
5751    Signed-off-by: Julien Danjou <julien@danjou.info>
5752
5753commit 625ed596cae6dd8175aeb6cb6f26784928042f22
5754Author: Josh Triplett <josh@freedesktop.org>
5755Date:   Wed Oct 8 16:04:25 2008 -0700
5756
5757    Remove duplicate XCB_EXTENSION calls for Composite extension
5758
5759commit db332dcda989b5b021dc220c102666f695c772cf
5760Author: Henning Sten <henning.sten@yahoo.com>
5761Date:   Sat Sep 20 13:08:58 2008 +0200
5762
5763    fix tiny memory leak in read_packet (leak only happens when malloc returns NULL so it's very rare)
5764    
5765    Signed-off-by: Julien Danjou <julien@danjou.info>
5766
5767commit 9afadd2aef8af89a4f4ab70baeae0b848904c367
5768Author: Carsten Meier <cm@trexity.de>
5769Date:   Tue Sep 9 12:11:37 2008 +0200
5770
5771    Added generation of extern "C" for compatibility with C++
5772    
5773    The auto-generated header files now include an extern "C"
5774    declaration for compatibility with C++.
5775    
5776    Signed-off-by: Julien Danjou <julien@danjou.info>
5777
5778commit 25e59ccc0dc8baf344145d6d739229e8120330db
5779Author: Julien Cristau <jcristau@debian.org>
5780Date:   Tue Sep 9 04:42:36 2008 +0100
5781
5782    Add support for the abstract socket namespace under Linux
5783    
5784    Based on same in Xtrans.
5785    
5786    Signed-off-by: Julien Danjou <julien@danjou.info>
5787
5788commit f3f8738436d09f7e590b76e22a7a2cc4b16abd1d
5789Author: Julien Cristau <jcristau@debian.org>
5790Date:   Tue Sep 9 04:42:35 2008 +0100
5791
5792    Fix some fd leaks in _xcb_open_*()
5793    
5794    Signed-off-by: Julien Danjou <julien@danjou.info>
5795
5796commit d6d44e1bf09cca8aefbf4ce9875d7f794bf19fb1
5797Author: Bart Massey <bart@cs.pdx.edu>
5798Date:   Wed Sep 3 13:52:58 2008 -0700
5799
5800    fixed overly aggressive warning about fixed field following variable
5801
5802commit 7e0674e76186ee4491a089350511fc0d22fb3af3
5803Author: Bart Massey <bart@cs.pdx.edu>
5804Date:   Sun Aug 31 00:42:23 2008 -0700
5805
5806    added small fix to support trailing fixed fields; also warning for non-pad fixed fields
5807
5808commit 2d04a1e6cedcdc832e2db3c65ababc3aff904ec4
5809Author: Vincent Torri <vtorri at univ-evry dot fr>
5810Date:   Sun Aug 31 10:33:31 2008 +0200
5811
5812    factorize m4 macros and add one to set X extensions
5813    
5814    Signed-off-by: Julien Danjou <julien@danjou.info>
5815
5816commit bcf662c1b433b296060c66ae1656fcb5c6e697ef
5817Author: Julien Danjou <julien@danjou.info>
5818Date:   Thu Aug 28 14:35:54 2008 +0200
5819
5820    Initialize all fields of addrinfo
5821    
5822    Signed-off-by: Julien Danjou <julien@danjou.info>
5823
5824commit baf31b1bf20b49ec00d0f64bb7cc9c788a28c088
5825Author: Julien Danjou <julien@danjou.info>
5826Date:   Thu Aug 28 13:51:38 2008 +0200
5827
5828    Use ifdef instead of if for defined value
5829    
5830    Signed-off-by: Julien Danjou <julien@danjou.info>
5831
5832commit 38d5de3a5573b2e89e97d04a809a3dd38a0fe8a7
5833Author: Julien Danjou <julien@danjou.info>
5834Date:   Wed Aug 27 13:56:28 2008 +0200
5835
5836    Set namelen unsigned
5837    
5838    Signed-off-by: Julien Danjou <julien@danjou.info>
5839
5840commit 9c9c09b376fe1ddcedd03c52cfc0b06867d998c9
5841Author: Julien Danjou <julien@danjou.info>
5842Date:   Wed Aug 27 13:56:26 2008 +0200
5843
5844    Rename index to idx to avoid shadowing
5845    
5846    Signed-off-by: Julien Danjou <julien@danjou.info>
5847
5848commit c5b2e53abf0b113d4cc4105127cf848ee450aa98
5849Author: Julien Danjou <julien@danjou.info>
5850Date:   Wed Aug 27 13:56:25 2008 +0200
5851
5852    Use a signed size in read_block()
5853    
5854    Signed-off-by: Julien Danjou <julien@danjou.info>
5855
5856commit 1bbdba52116f127bed3ce812a00240b4009bbf22
5857Author: Julien Danjou <julien@danjou.info>
5858Date:   Wed Aug 27 13:56:24 2008 +0200
5859
5860    Use unsigned to compare and rename sync
5861    
5862    - i must be unsigned to be compare in the loop
5863    - sync shadow global sync() function
5864    
5865    Signed-off-by: Julien Danjou <julien@danjou.info>
5866
5867commit 6438584285de72858f97be891e16a125d13471d8
5868Author: Julien Danjou <julien@danjou.info>
5869Date:   Wed Aug 27 13:56:23 2008 +0200
5870
5871    Fix htonl() arg & convert sizeof() to signed
5872    
5873    Signed-off-by: Julien Danjou <julien@danjou.info>
5874
5875commit 1ead02e88eb9f37757eeb1cc8c762fc48d6d08ee
5876Author: Julien Danjou <julien@danjou.info>
5877Date:   Wed Aug 27 13:56:22 2008 +0200
5878
5879    initialize global_id to 0
5880    
5881    Signed-off-by: Julien Danjou <julien@danjou.info>
5882
5883commit f209d0ef7ad57a395c01ca09ecf48117a648e39c
5884Author: Peter Hutterer <peter.hutterer@who-t.net>
5885Date:   Thu Jul 17 13:57:41 2008 +0930
5886
5887    Bump to 1.1.91.
5888
5889commit cdc347938702dddbacb5af5c24988e9152b5447c
5890Author: Julien Cristau <jcristau@debian.org>
5891Date:   Wed Jul 16 23:25:04 2008 +0930
5892
5893    Make EXTHEADERS, EXTSOURCES, EXTENSION_XML unconditional of configure flags.
5894    
5895    yay, make distcheck works now even when some extensions are disabled.
5896    
5897    Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
5898
5899commit a9d15a08451c76a9250642c9f662f296196f60a0
5900Author: Petr Salinger <Petr.Salinger@seznam.cz>
5901Date:   Mon Jul 7 17:57:37 2008 +0200
5902
5903    fix FreeBSD support
5904    
5905    The GNU/kFreeBSD (and BSDs in general) have a different
5906    layout of struct sockaddr, sockaddr_in, sockaddr_un ...
5907    
5908    The first member do not have to be "sa_family",
5909    they also have "sa_len" field.
5910    
5911    Signed-off-by: Julien Danjou <julien@danjou.info>
5912
5913commit ee78071902e93ce22a3170f0937c158fd16894d8
5914Author: Peter Hutterer <peter@cs.unisa.edu.au>
5915Date:   Wed May 28 17:41:35 2008 +0930
5916
5917    Bump to 1.1.90.
5918
5919commit 424ad131b7d97d6196181c31677655e6d52b41df
5920Author: Josh Triplett <josh@freedesktop.org>
5921Date:   Wed May 28 12:26:13 2008 -0700
5922
5923    Fix variable declaration formatting
5924
5925commit 6532c715c3805128b9976ab208f1426f691056a2
5926Author: Peter Hutterer <peter@cs.unisa.edu.au>
5927Date:   Tue May 15 16:28:19 2007 +0930
5928
5929    Add xcb_ge_event_t and handling for long events.
5930    
5931    GenericEvent can be more than 32 bytes long. Ensure that the required data is
5932    pulled off the wire and tack it onto the event.
5933    
5934    Due to the structure of the xcb_generic_event_t, the data is appended AFTER
5935    the full_sequence field.
5936
5937commit b08a5909daf589d5e06c17c55d044f39c1d3479a
5938Author: Oswald Buddenhagen <ossi@kde.org>
5939Date:   Thu May 1 16:17:55 2008 -0400
5940
5941    Fix libxcb/src compile with srcdir != builddir.
5942
5943commit 4a405feba8cde8490d847a57b7e833176e18b90f
5944Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5945Date:   Wed Apr 23 20:26:28 2008 -0400
5946
5947    Replace a stray c-client.xsl in the libxcb SOURCES.  Fixes make distcheck.
5948
5949commit 40566c36d543edc2118cbb358e0303d9e8862892
5950Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5951Date:   Wed Apr 23 20:25:57 2008 -0400
5952
5953    Use the python install path from xcb-xproto.pc to locate the xcbgen package.
5954
5955commit b3832bcc46d85110fdb2b553df6249a831cfe0fa
5956Author: Jeremy Kolb <jkolb@glorfindel.hsd1.ma.comcast.net>
5957Date:   Sun Apr 20 16:26:51 2008 -0400
5958
5959    Add mention of PYTHONPATH if xcbgen cannot be found.
5960
5961commit 5ee915e12a102e86e141981bbce60ed81037dfdc
5962Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5963Date:   Fri Apr 18 16:30:08 2008 -0400
5964
5965    Add Python parser C language-dependent part.
5966
5967commit 947a2e26e4217531e612a5110e6f95296c94614f
5968Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5969Date:   Thu Jan 24 16:02:34 2008 -0500
5970
5971    Add SELinux extension support, disabled by default.
5972
5973commit 0b5f684eb0f8a8ad1887514568532037bb46020c
5974Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5975Date:   Thu Jan 24 15:57:35 2008 -0500
5976
5977    Add XInput extension support, disabled by default.
5978
5979commit c72581c844efbaaa7e632377b57678f6668327fe
5980Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5981Date:   Tue Jan 15 17:06:14 2008 -0500
5982
5983    Inputs to AC_CONFIG_FILES are automatically distributed.
5984    Hence, it is not necessary to explicitly add them to EXTRA_DIST.
5985
5986commit 7a74ba3d0212f9bfe021d6da9070f71cbc53f85b
5987Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5988Date:   Fri Dec 7 16:22:04 2007 -0500
5989
5990    Generated the configure.ac and Makefile.am's in libxcb with the idea of
5991    making each extension library individually selectable for build.
5992    
5993    Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
5994
5995commit bcd1dcec9b242d7e2185c1ae83d3884844a2647f
5996Merge: baae582 9bf8329
5997Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5998Date:   Fri Dec 7 16:18:00 2007 -0500
5999
6000    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6001
6002commit 9bf8329b50a8edacf00efb074f73c3bb759f7c8d
6003Author: Bart Massey <bart@cs.pdx.edu>
6004Date:   Sat Nov 24 14:53:54 2007 -0800
6005
6006    make IPv6 optional
6007
6008commit 0593989103c48cc10165066d985c9f2d3355926b
6009Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6010Date:   Fri Nov 16 19:38:40 2007 -0500
6011
6012    Remove file that wasn't meant to be committed.
6013
6014commit baae5826a6f51490e842be931c8b9f76086c4d98
6015Merge: c3136d1 46413cd
6016Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6017Date:   Fri Nov 16 19:36:08 2007 -0500
6018
6019    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6020
6021commit 46413cd85ee4f3d51a3a3e1d8ee13bc5fa6c2d5d
6022Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6023Date:   Fri Nov 16 19:34:42 2007 -0500
6024
6025    Add comment noting the requirement to free replies when finished.
6026
6027commit c3136d18321df31caa7f582d475132c2e02155de
6028Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6029Date:   Fri Nov 16 19:33:20 2007 -0500
6030
6031    Add generated comment noting requirement to free replies after use.
6032
6033commit 3c6c8f127c2bce4f45bface7dd45cc719af9de0d
6034Author: Jamey Sharp <jamey@minilop.net>
6035Date:   Sun Nov 4 17:29:13 2007 -0800
6036
6037    Release libxcb 1.1
6038
6039commit af50de26c10c93ccc4cd3bc61e92aff47651b961
6040Author: Jamey Sharp <jamey@minilop.net>
6041Date:   Sun Nov 4 17:26:21 2007 -0800
6042
6043    Revert "Generate error constants as XCB_BAD_*, similar to Xlib."
6044    
6045    Since several extensions named their errors like "BadFoo", this patch
6046    results in names like XCB_EXT_BAD_BAD_FOO, which is really awful. Those
6047    extensions are already kind of awful, as they produce structure names
6048    like xcb_ext_bad_foo_error_t, which is redundant.
6049    
6050    A patch that removes "Bad" from the XML extension descriptions, while
6051    maintaining API and ABI compatibility in XCB, is needed before this
6052    patch can be released.
6053    
6054    This reverts commit 158c9b6ba18b39f424bd524fceb66f3fec0d1616.
6055
6056commit a29fbc2645fabb96d02c382ffef499b48fb1514a
6057Author: Jamey Sharp <jamey@minilop.net>
6058Date:   Sun Oct 28 13:28:18 2007 -0700
6059
6060    Don't hold the xlib-xcb lock while sleeping: that allows deadlock.
6061    
6062    With this patch, `ico -threads 2` runs without deadlock.
6063    
6064    Many thanks to Christoph Pfister <christophpfister@gmail.com> for
6065    pointing out the problem, providing detailed analyses, explaining it to
6066    me repeatedly until I understood what was going on, and proposing and
6067    reviewing possible solutions.
6068    
6069    Signed-off-by: Jamey Sharp <jamey@minilop.net>
6070    Acked-by: Christoph Pfister <christophpfister@gmail.com>
6071
6072commit f6b75d6090dc40918196d2b902e9616d0199af42
6073Author: Jamey Sharp <jamey@minilop.net>
6074Date:   Sun Oct 28 11:56:08 2007 -0700
6075
6076    Factor pthread_cond_wait(iolock) to _xcb_wait_io.
6077    
6078    This parallels the _xcb_lock_io and _xcb_unlock_io factoring.
6079
6080commit 4d828c5eba9fc7161c5f18650f2dbe218e1db06f
6081Author: Jamey Sharp <jamey@minilop.net>
6082Date:   Tue Oct 23 11:03:33 2007 -0700
6083
6084    Don't abort() on locking assertions if LIBXCB_ALLOW_SLOPPY_LOCK is set.
6085    
6086    But do still print a full backtrace, on platforms where that's
6087    supported.
6088    
6089    This commit follows the spirit of Novell's libxcb-sloppy-lock.diff.
6090    
6091    I strongly opposed proposals like this one for a long time. Originally I
6092    had a very good reason: libX11, when compiled to use XCB, would crash
6093    soon after a locking correctness violation, so it was better to have an
6094    informative assert failure than a mystifying crash soon after.
6095    
6096    It took some time for me to realize that I'd changed the libX11
6097    implementation (for unrelated reasons) so that it could survive most
6098    invalid locking situations, as long as it wasn't actually being used
6099    from multiple threads concurrently.
6100    
6101    The other thing that has changed is that most of the code with incorrect
6102    locking has now been fixed. The value of the assert is accordingly
6103    lower.
6104    
6105    However, remaining broken callers do need to be fixed. That's why libXCB
6106    will still noisily print a stacktrace (if possible) on each assertion
6107    failure, even when assert isn't actually invoked to abort() the program;
6108    and that's why aborting is still default. This environment variable is
6109    provided only for use as a temporary workaround for broken applications.
6110    
6111    Signed-off-by: Jamey Sharp <jamey@minilop.net>
6112    Acked-by: Josh Triplett <josh@freedesktop.org>
6113
6114commit 09045eaac34973662aaa820a94ca8ed66d9dcb4e
6115Author: Egbert Eich <eich@freedesktop.org>
6116Date:   Thu Jul 19 17:00:18 2007 +0200
6117
6118    Allow unix:<screen> style display names again.
6119    
6120    https://bugzilla.novell.com/show_bug.cgi?id=289007
6121    This notion is used in a lot of scripts.
6122
6123commit 65ffbc6cfdb97b14689d3baef183cd50fbd31a7f
6124Author: Josh Triplett <josh@freedesktop.org>
6125Date:   Wed Jun 13 23:46:37 2007 -0700
6126
6127    Send locking assertion backtraces to stderr.  Improve the heading on the backtrace.
6128
6129commit 605c778e695a4535c35c5324325f310b5faf80e2
6130Author: Christoph Pfister <christophpfister@gmail.com>
6131Date:   Wed Jun 6 17:17:49 2007 +0200
6132
6133    Print backtraces in case an assert fails inside xlib/xcb.
6134    
6135    As you know there are some nasty libs / apps doing locking
6136    incorrectly. In order to improve the information given to the user
6137    when he encounters such a situation (people don't run apps in gdb
6138    normally) I created the patch attached.
6139    It's very non-intrusive (and affects only xlib/xcb, Josh told me on
6140    irc that it could be useful for other areas too, personally I don't
6141    think that it's really needed at other places ...).
6142    
6143    Some same outputs and the discussion of them:
6144    
6145        lxuser@pdln:/tmp$ ./main
6146        Got a backtrace:
6147        #0 /tmp/usr/lib/libxcb-xlib.so.0 [0xb7f9d728]
6148        #1 /tmp/usr/lib/libxcb-xlib.so.0(xcb_xlib_unlock+0x31) [0xb7f9d861]
6149        #2 ./test.so(function_a+0x11) [0xb7f9f3fd]
6150        #3 ./test.so(function_b+0x11) [0xb7f9f410]
6151        #4 ./main [0x80484a7]
6152        #5 /lib/libc.so.6(__libc_start_main+0xdc) [0xb7e60ebc]
6153        #6 ./main [0x80483f1]
6154        main: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
6155        Aborted
6156    
6157    That's kinda the normal situation.
6158    
6159        lxuser@pdln:/tmp$ ./main
6160        Got a backtrace:
6161        #0 /tmp/usr/lib/libxcb-xlib.so.0 [0xb7f90728]
6162        #1 /tmp/usr/lib/libxcb-xlib.so.0(xcb_xlib_unlock+0x31) [0xb7f90861]
6163        #2 /tmp/test.so [0xb7f923cd]
6164        #3 /tmp/test.so(function_b+0x11) [0xb7f923e0]
6165        #4 ./main [0x80484ab]
6166        #5 /lib/libc.so.6(__libc_start_main+0xdc) [0xb7e53ebc]
6167        #6 ./main [0x80483f1]
6168        main: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
6169        Aborted
6170    
6171    There are two possible reasons that the name doesn't appear in #2:
6172    a) a hidden symbol or a symbol with statical linkage in a library
6173    b) a symbol in an app not compiled with -rdynamic.
6174    But in both cases you still know _where_ the caller is.
6175    
6176    Note that in this example test.so was compiled with
6177    -fomit-frame-pointer; this isn't an issue as _one_ (= the caller)
6178    stack trace is still valid (as long as you don't have the insane idea
6179    to compile xcb with -fo-f-p).
6180    
6181    Another issue that may appear is "tail call elimination" (some entries
6182    are mysteriously missing; this is quite ugly, but you still get enough
6183    information so that you can do something useful with the issue e.g. by
6184    disassembling the relevant parts with gdb).
6185    
6186    Signed-off-by: Jamey Sharp <jamey@minilop.net>
6187
6188commit e20a31d72b8838cdf31b568431b5ad78492c1481
6189Author: Jamey Sharp <jamey@minilop.net>
6190Date:   Sat Jun 2 18:29:37 2007 -0700
6191
6192    xcb_poll_for_event: Return already-read events before read(2)ing again.
6193
6194commit 2ec1383a68bf9f4baf7125a7d6544167f38d8d62
6195Merge: 3abd416 158c9b6
6196Author: Matthias Hopf <mhopf@suse.de>
6197Date:   Wed Apr 18 12:46:08 2007 +0200
6198
6199    Merge branch 'master' of git://anongit.freedesktop.org/git/xcb/libxcb
6200
6201commit 158c9b6ba18b39f424bd524fceb66f3fec0d1616
6202Author: Ian Osgood <iano@quirkster.com>
6203Date:   Fri Apr 13 15:14:12 2007 -0700
6204
6205    Generate error constants as XCB_BAD_*, similar to Xlib.
6206    
6207    The previous constants remain for compatibility, but should be deprecated.
6208
6209commit f7279d8c8ae33d6e3029e85b14ed8495d3b00e30
6210Merge: 0925e47 eaa380e
6211Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6212Date:   Thu Apr 12 15:48:44 2007 +0200
6213
6214    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6215
6216commit 0925e470b2abf8837c62e81428660a01bbb990fa
6217Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6218Date:   Thu Apr 12 15:46:05 2007 +0200
6219
6220    remove the 7th bit of the response_type for the event loops
6221
6222commit 3abd41625c7d6db6d01f3167d6bac2b7481965cf
6223Author: Matthias Hopf <mhopf@suse.de>
6224Date:   Wed Apr 11 17:31:31 2007 +0200
6225
6226    Update autogen.sh to one that does objdir != srcdir
6227
6228commit eaa380efefd347abcd11a6c24c008686beaf8257
6229Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6230Date:   Tue Apr 10 11:56:06 2007 -0400
6231
6232    Modify new attribute from previous patch so that it is necessary only on
6233    extensions with split names.  Tested with diff and found no difference with
6234    previous stylesheet header-file output.
6235
6236commit 91be36f845352dea7838853d0f98ae144fe571a4
6237Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
6238Date:   Thu Mar 29 12:28:07 2007 -0700
6239
6240    Replaces special-casing in c-client stylesheet with support
6241    for new attribute.  Tested with diff and found no difference with
6242    previous stylesheet header-file output.
6243
6244commit 4a60950b7433eb41e08cb0c74dc8ced8f92fd78a
6245Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6246Date:   Tue Feb 27 13:35:02 2007 +0100
6247
6248    add the first step toward the documentation of the request/reply functions. The arguments of the requests are not found yet. Josh, can you look at it ?
6249
6250commit bca41cdcdb11e2a610d7b7a3d077cbb77bc81e45
6251Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6252Date:   Wed Feb 7 20:08:27 2007 +0100
6253
6254    and make the html code valid...
6255
6256commit acefe83bea9b2b17a956d68f5be504866d34196f
6257Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6258Date:   Wed Feb 7 20:04:10 2007 +0100
6259
6260    font part
6261
6262commit 8a8c1fa184939ef23f96421990c171b49d16ee33
6263Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6264Date:   Wed Feb 7 18:57:46 2007 +0100
6265
6266    no more xid or id fields
6267
6268commit 6cedaece0e294d39fe090b87e2f60569b40a4554
6269Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6270Date:   Tue Feb 6 09:31:54 2007 +0100
6271
6272    add the complete cursor example. Make the html code valid
6273
6274commit ab22a4d61665bbe637b8c1f349fcfaf04e386e88
6275Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6276Date:   Mon Jan 22 11:40:15 2007 +0100
6277
6278    add doxygen doc for the *_end functions
6279
6280commit 342e9cefcc7d942714488f6b424fee7a5bdc741f
6281Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6282Date:   Sat Jan 13 11:20:55 2007 +0100
6283
6284    fix all the occurence where a drawable where considered as a union
6285
6286commit 70a72f65e438888a5530e9911c36aad68833790f
6287Author: Josh Triplett <josh@freedesktop.org>
6288Date:   Sun Dec 10 21:10:06 2006 -0800
6289
6290    Add autogen.sh to EXTRA_DIST.
6291
6292commit 867ae5eb5c19091365daad621b923f56ff699415
6293Author: Josh Triplett <josh@freedesktop.org>
6294Date:   Sun Dec 10 21:09:30 2006 -0800
6295
6296    Add tools/* to EXTRA_DIST.
6297
6298commit 67af2d24e4c46580479570cf09586a54b84b1b63
6299Author: TORRI Vincent <torri@doursse.(none)>
6300Date:   Tue Nov 28 20:26:54 2006 +0100
6301
6302    add doc tag for the _next functions. It creates doxygen doc in the header files for these functions
6303
6304commit 4c8777f87a28ff5bf45cbdddce509163dbcf8137
6305Merge: e624cca f486075
6306Author: TORRI Vincent <torri@doursse.(none)>
6307Date:   Tue Nov 28 20:15:27 2006 +0100
6308
6309    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6310
6311commit e624cca7df4c2c5bc5e44af1e851e5a0d17c88bd
6312Author: TORRI Vincent <torri@doursse.(none)>
6313Date:   Tue Nov 28 20:14:16 2006 +0100
6314
6315    add some output informations to configure script
6316
6317commit f486075fa093846c3f32b3e4b9624c039ea26ba4
6318Author: Tilman Sauerbeck <tilman@code-monkey.de>
6319Date:   Fri Nov 24 15:48:08 2006 +0100
6320
6321    If enable_build_docs is "no", we don't even try to look for doxygen.
6322    
6323    Also set BUILD_DOCS in that branch. Now the disable-build-docs works
6324    as expected.
6325
6326commit 2e8e6debac39864450c0a69633086ad92459c25e
6327Author: Tilman Sauerbeck <tilman@code-monkey.de>
6328Date:   Fri Nov 24 15:43:39 2006 +0100
6329
6330    Fixed evaluation of the disable-build-docs argument.
6331    
6332    Now at least enable_build_docs is set correctly.
6333
6334commit 19dfaf93ac1b5e9d3ce09b1f8e2338e53a5d7324
6335Author: Ian Osgood <iano@quirkster.com>
6336Date:   Sun Nov 26 09:26:32 2006 -0800
6337
6338    Fix unit tests for FreeBSD
6339    
6340    putenv() string must contain '='
6341    environment failure test is invalid if argument is NULL
6342
6343commit 98e2a5617ef1c9955b3b5553224c34f55c7c5d29
6344Author: Josh Triplett <josh@freedesktop.org>
6345Date:   Sat Nov 25 22:27:33 2006 -0800
6346
6347    Use substitition variables in xcb-xinerama.pc.in, not instances of their values
6348    
6349    xcb-xinerama.pc.in looked more like a generated .pc file; replace specific
6350    instances of values provided by an invocation of configure with the general
6351    substitution variables configure replaces.
6352
6353commit e74cdcd02e6814222a76c0a237efca16c423bb26
6354Author: Ian Osgood <iano@quirkster.com>
6355Date:   Sat Nov 25 11:00:14 2006 -0800
6356
6357    Bug #9119: test xcb_popcount
6358
6359commit 05d23a724d4dde42b11d6e9dec9ccaf5a516e287
6360Author: Jamey Sharp <jamey@minilop.net>
6361Date:   Fri Nov 24 16:22:13 2006 -0800
6362
6363    libXau didn't have a correct pkg-config file until 0.99.2: fail if an older version is found.
6364
6365commit 3360d0c79e98cf6b7f30b2d84f117aea0a28595d
6366Author: Jamey Sharp <jamey@minilop.net>
6367Date:   Fri Nov 24 13:24:05 2006 -0800
6368
6369    NetBSD doesn't have AI_ADDRCONFIG: use it only if it's available.
6370
6371commit 818811a7ac660e46d0dca1cbf9e53ad0475af330
6372Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6373Date:   Fri Nov 24 12:33:15 2006 +0100
6374
6375    replace all the _new functions with xcb_generate_id. repalce X11/XCB with xcb. Fix a description of the default background of a window
6376
6377commit 27f98afc2f7039f8546d6cab5a72f609c72299d6
6378Author: Josh Triplett <josh@freedesktop.org>
6379Date:   Thu Nov 23 07:30:51 2006 -0800
6380
6381    Release libxcb 1.0
6382
6383commit 11738b2af0d8bfcf5b2f7c0d3e6ade1a14866b94
6384Author: Diego 'Flameeyes' Pettenò <flameeyes@gentoo.org>
6385Date:   Thu Nov 23 15:15:30 2006 +0100
6386
6387    Avoid race condition when using multiple make jobs
6388    
6389    Avoid race condition when symlinking XML files.
6390    When declaring a rule with many files as target, the rule is called
6391    when any of them is requested, resulting in multiple for loops happening
6392    during a make process using more than one job.
6393    Also, use '$(LN_S) -f' rather than removing and recreating a file,
6394    that one should be as supported as 'rm -f' and requires one less command.
6395
6396commit 30c768b322f613d697a61997e081a19aecd0b07f
6397Author: Josh Triplett <josh@freedesktop.org>
6398Date:   Thu Nov 23 02:11:02 2006 -0800
6399
6400    Rewrite automake's data installation rules, because they suck.
6401    
6402    Specifically, they didn't handle installing data from both srcdir and builddir.
6403    We have the tutorial in the srcdir, and build the manual in the builddir.
6404    
6405    Also, stop rebuilding the manual for each make target in the doc directory, and
6406    every time any of those targets get called.  This change now makes the manual
6407    never rebuild once built; we plan to fix that later, by rewriting the makefiles
6408    to avoid recursive make, and then making the manual depend on the source files.
6409    
6410    Commit by Jamey Sharp and Josh Triplett.
6411
6412commit af3a15838fc7dc5b4e52197854deb54da6baeaf1
6413Author: Josh Triplett <josh@freedesktop.org>
6414Date:   Thu Nov 23 00:08:30 2006 -0800
6415
6416    Rework doxygen build and install to work with srcdir != builddir
6417    
6418    The documentation generation with doxygen now works when built out of tree,
6419    with srcdir != builddir.  xcb.doxygen now gets generated from xcb.doxygen.in,
6420    so that it can use top_builddir and top_srcdir to find source and to output
6421    documentation.  Also fill in PROJECT_NUMBER from @VERSION@, now that we have
6422    it readily available via autoconf.
6423
6424commit 608058ec80edb041ca012d530b42d97474f80320
6425Author: Josh Triplett <josh@freedesktop.org>
6426Date:   Wed Nov 22 22:47:15 2006 -0800
6427
6428    Remove --with-opt and --with-debug options from configure.ac; use CFLAGS instead
6429    
6430    configure supports using custom CFLAGS, so remove the --with-opt and
6431    --with-debug options from configure.ac, and the corresponding usage of
6432    COPTFLAGS and CDEBUGFLAGS in src/Makefile.am.
6433
6434commit 1aade6a15f767c49db5c1b69b97a5b1bf8cdce57
6435Author: Josh Triplett <josh@freedesktop.org>
6436Date:   Wed Nov 22 21:49:52 2006 -0800
6437
6438    Check for getaddrinfo rather than gethostbyname in configure.ac
6439    
6440    Since the addition of IPv6 support, we now use getaddrinfo rather than
6441    gethostbyname; update configure.ac accordingly.
6442
6443commit 183c2ba4cca71f4b12f106f801f0bd30c4c527f1
6444Author: Jamey Sharp <jamey@minilop.net>
6445Date:   Wed Nov 22 21:27:41 2006 -0800
6446
6447    xcb-proto has no libraries or headers, so don't use XCBPROTO_CFLAGS or XCBPROTO_LIBS.
6448
6449commit e54dfd73a9652bbc103f3f874abc303aed4b5d76
6450Author: Jamey Sharp <jamey@minilop.net>
6451Date:   Wed Nov 22 20:29:04 2006 -0800
6452
6453    Use pthread-stubs as needed, and list xdmcp in Requires.private when XCB is built to use it.
6454
6455commit 3de6ab7b786775d9e7df1523c27cdfe3ffd3a25f
6456Author: Josh Triplett <josh@freedesktop.org>
6457Date:   Tue Nov 21 20:29:34 2006 -0800
6458
6459    Replace uses of "long" with uint32_t, and similar; fixes 64-bit bugs
6460
6461commit 7fbfebaa3fb3a5ca2d2a307a0a5c40c015e18115
6462Author: Ian Osgood <iano@quirkster.com>
6463Date:   Tue Nov 21 09:04:37 2006 -0800
6464
6465    Fix IP6 work for FreeBSD/Mac.
6466
6467commit 684b8271a4539527daa15da82ad0cc302fb44727
6468Author: Josh Triplett <josh@freedesktop.org>
6469Date:   Tue Nov 21 00:57:49 2006 -0800
6470
6471    Support XDM-AUTHORIZATION-1 on IPv6
6472    
6473    Commit by Jamey Sharp and Josh Triplett.
6474
6475commit d6abe93b06c421b78e92d76ceb5ca181e3adff31
6476Author: Josh Triplett <josh@freedesktop.org>
6477Date:   Mon Nov 20 23:25:41 2006 -0800
6478
6479    Refactor auth code to get display number from xcb_connect
6480    
6481    Change xcb_connect to pass the display number to _xcb_get_auth_info, which
6482    passes it to get_authptr.  This allows get_authptr to stop hacking the display
6483    number out of the sockaddrs of various address families, such as
6484    port - X_TCP_PORT, or the number after the last X in the UNIX socket path. This
6485    also removes a portability bug introduced during the IPv6 changes: relying on
6486    '\0'-termination of the UNIX socket path in a sockaddr_un.
6487    
6488    Commit by Jamey Sharp and Josh Triplett.
6489
6490commit 4a928de402a6e69886921fe428bbffb909c6405e
6491Author: Josh Triplett <josh@freedesktop.org>
6492Date:   Mon Nov 20 22:26:40 2006 -0800
6493
6494    Only use AI_NUMERICSERV if defined.
6495
6496commit 907f8c8c4906cbd2352f3bbddfe144ad81a2cfa8
6497Author: Josh Triplett <josh@freedesktop.org>
6498Date:   Mon Nov 20 22:16:24 2006 -0800
6499
6500    Support authentication for IPv6 connections
6501    
6502    Support AF_INET6 in get_authptr, and refactor to use common code for IPv4 and
6503    v4-mapped IPv6 addresses.
6504    
6505    Commit by Jamey Sharp and Josh Triplett.
6506
6507commit 48776ce233bf77fbaddbe972d2356bca69094239
6508Author: Josh Triplett <josh@freedesktop.org>
6509Date:   Mon Nov 20 17:53:30 2006 -0800
6510
6511    Support displays with IPv6 addresses or hosts which resolve to IPv6 addresses
6512    
6513    xcb_parse_display already correctly handled IPv6 displays.  Now, _xcb_open_tcp
6514    uses getaddrinfo, and correctly connects to IPv6 displays.  Displays can use
6515    bare IPv6 addresses, square-bracketed IPv6 addresses, or hostnames which
6516    resolve to IPv6 addresses.
6517    
6518    Since IPv6 addresses may include colons, including at the end, revise the
6519    DECnet display parsing code to avoid triggering on IPv6 addresses.
6520    
6521    Authorization may not work with IPv6 connections yet.
6522    
6523    This commit brought to you by the (display) number ::1:1.1, the letter X,
6524    the Gobby collaborative editor, Josh Triplett, and Jamey Sharp.
6525
6526commit 6c3a8db3f69aebf2b5c1330ec6c11b3138f9bdfe
6527Author: Josh Triplett <josh@freedesktop.org>
6528Date:   Mon Nov 20 17:48:49 2006 -0800
6529
6530    Add new xcb_parse_display test cases, most related to IPv6 addresses
6531
6532commit 3d9bb02012b35fd51ed3352d0bd974a0f6e439d8
6533Author: Jamey Sharp <jamey@minilop.net>
6534Date:   Sun Nov 19 21:48:27 2006 -0800
6535
6536    Bug #5958: Also zero out the implicit pad byte in empty requests.
6537
6538commit 3fa50020d2d9b78e22cf3597f4688c4e22df5121
6539Author: Jamey Sharp <jamey@minilop.net>
6540Date:   Sun Nov 19 18:31:48 2006 -0800
6541
6542    Remove support for the <localfield> tag: nothing needs it.
6543    
6544    Only one use of <localfield> remained, for a list length expression in
6545    xv.xml. List length parameters that don't actually appear in the
6546    protocol should be left implicit: if no length expression is given, then
6547    a localfield will be automatically created by c-client.xsl.
6548
6549commit 16e7328b4a366b29d54696aafffe62c92e0e0130
6550Author: Jamey Sharp <jamey@minilop.net>
6551Date:   Sun Nov 19 17:56:32 2006 -0800
6552
6553    Bug #5958: zero out padding bytes in requests.
6554
6555commit 6eee1707ea7bc303620e39911706d5fb71fa16b0
6556Author: Jamey Sharp <jamey@minilop.net>
6557Date:   Sun Nov 19 00:30:37 2006 -0800
6558
6559    htmldir was introduced in autoconf 2.59c: set a reasonable value for earlier versions.
6560
6561commit da4d56ef5a880eb24014a141e6e16668ab51f180
6562Author: Jamey Sharp <jamey@minilop.net>
6563Date:   Sun Nov 12 15:30:10 2006 -0800
6564
6565    Provide xcb_prefetch_maximum_request_length counterpart to xcb_get_maximum_request_length.
6566
6567commit aedfa1fe1d91a10ccfe3ee6ac6b7a25885623dc6
6568Author: Jamey Sharp <jamey@minilop.net>
6569Date:   Sat Nov 18 21:30:12 2006 -0800
6570
6571    Fix Doxygen warnings.
6572    
6573    Specifically, fixes these two warnings which were emitted for every
6574    generated source file:
6575    
6576    * Warning: end of file while inside a group
6577    * Warning: group XCB_BigRequests_API already documented. Skipping documentation.
6578
6579commit 44a2160c95f6c90d461d66c0278348c62d2d77fa
6580Author: Jamey Sharp <jamey@minilop.net>
6581Date:   Sat Nov 18 19:36:48 2006 -0800
6582
6583    Make doc installation work and simplify doc/Makefile.am.
6584    
6585    Don't override the default htmldir with an unquoted copy.
6586    
6587    Don't suppress `make` echoing on Doxygen commands.
6588    
6589    Ensure the tutorial is always installed even if Doxygen isn't
6590    available.
6591    
6592    Take better advantage of the automake installation infrastructure.
6593
6594commit 142fe9a6f2b13d42b0e4baaa25155067f45d065b
6595Author: Jamey Sharp <jamey@minilop.net>
6596Date:   Sat Nov 18 16:51:01 2006 -0800
6597
6598    Turn off Doxygen warnings for undocumented API, for now.
6599    
6600    But warn about absent parameter documentation for functions that are
6601    otherwise documented.
6602
6603commit 3634299ecc5223e8505d19468cf04770ac08ec14
6604Author: Jamey Sharp <jamey@minilop.net>
6605Date:   Thu Nov 16 17:02:17 2006 -0800
6606
6607    Switch from the old AM_PATH_CHECK macro to pkg-config.
6608    
6609    check 0.9.4 is now required to build XCB's unit tests.
6610    
6611    The version that we were requiring was not actually new enough to let
6612    our unit tests compile, and the AM_PATH_CHECK macro is now considered
6613    deprecated. We know that versions of check using pkg-config are new
6614    enough to work, and the check dependency was optional anyway, so we've
6615    dropped support for older versions.
6616
6617commit 82762555a1af0b170d01e2042eefe34086e6c132
6618Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6619Date:   Sun Nov 5 18:24:30 2006 +0100
6620
6621    add Makefile.am for building/installing doxygen doc
6622
6623commit f32c7131cc89c64e09e9a88c5f859824c0dfa54f
6624Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6625Date:   Sun Nov 5 18:22:55 2006 +0100
6626
6627    add xcb.doxygen to EXTRA_DIST
6628
6629commit e3b7009dc0be6515d23fbda9c744082506df4cb0
6630Merge: 0262cd6 32cd2d2
6631Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6632Date:   Sun Nov 5 18:21:34 2006 +0100
6633
6634    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6635
6636commit 0262cd63a2c10e52daa3bd71cb21a6a37017f234
6637Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6638Date:   Sun Nov 5 18:20:45 2006 +0100
6639
6640    add AM_CHECK_DOXYGEN macro
6641
6642commit 32cd2d20477ab17704f0a2696c2b3bc2ff8f42de
6643Author: Jamey Sharp <jamey@minilop.net>
6644Date:   Thu Nov 2 18:04:16 2006 -0800
6645
6646    Release libxcb 1.0 RC3 (0.9.93).
6647
6648commit 9d155160546798052533c277c877879e83f85979
6649Author: Jeremy Kolb jkolb@brandeis.edu <jkolb@glorfindel.(none)>
6650Date:   Sun Oct 29 19:31:09 2006 -0500
6651
6652    Added initial doxygen generation stuff.  This should probably be cleaned up later.
6653    
6654    Added support for major/minor-version attributes in the xcb tag.  This is just to give some reference and help people when using *query_version.
6655
6656commit 8eba8690adac2c4425f4cac490c77f9f3b1da6f7
6657Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6658Date:   Fri Oct 20 07:00:15 2006 +0200
6659
6660    fix doxy doc
6661
6662commit a5529a4f2f1a9ed117e3465f0e422b707db21c68
6663Author: Jamey Sharp <jamey@minilop.net>
6664Date:   Wed Oct 18 23:28:27 2006 -0700
6665
6666    Add doxygen config file.
6667    
6668    Thanks to Jeremy Kolb <jkolb@brandeis.edu> for the initial version.
6669
6670commit 094248405fd9c0957b92f13acb92461ab61b2c3f
6671Author: Jamey Sharp <jamey@minilop.net>
6672Date:   Wed Oct 18 23:23:00 2006 -0700
6673
6674    Quit treating xproto specially: handle it like all the extensions.
6675
6676commit 6bc0b37303f26faf6fbdcbbee444c227e83a329c
6677Author: Jamey Sharp <jamey@minilop.net>
6678Date:   Wed Oct 18 23:21:39 2006 -0700
6679
6680    Ignore generated xinerama files.
6681
6682commit 5abb10d63ffebbe34054ae17bbdfd142e0b36396
6683Author: Josh Triplett <josh@freedesktop.org>
6684Date:   Mon Oct 16 05:11:57 2006 -0700
6685
6686    Also test xcb_parse_display with NULL argument and display in $DISPLAY
6687
6688commit 9d619d14e347b9dd1d544627060023a4bca2452b
6689Author: Jamey Sharp <jamey@minilop.net>
6690Date:   Sun Oct 15 12:34:30 2006 -0700
6691
6692    Match only XCB-namespaced XID generators when converting to xcb_generate_id.
6693    
6694    Thanks to Vincent Torri for pointing out that I screwed up. :-)
6695
6696commit 5ec55dc9c0c51137d89dd870ede05cc3268f75ae
6697Author: Zephaniah E. Hull <warp@aehallh.com>
6698Date:   Sat Oct 14 19:52:23 2006 -0700
6699
6700    Define and use constants for opcode numbers.
6701    
6702    Hard coding the opcode numbers in the function just makes it harder to figure
6703    out what's going on, but much more to the point, not defining the opcodes in
6704    the header makes it impossible to use the generated headers instead of the
6705    x11proto headers in the server.
6706    
6707    The name I settled on is very simple, for an extension by the name of xconf,
6708    and a request by the name of list_devices, we get XCB_XCONF_LIST_DEVICES.  If
6709    this somehow causes problems, we can probably add a _OP somewhere in there,
6710    but.
6711    
6712    Acked-by: Jamey Sharp <jamey@minilop.net>
6713    Closes: #8641
6714
6715commit c1504691ec2786594e8ffc33f3962f460f041971
6716Author: Jeremy Kolb <jkolb@brandeis.edu>
6717Date:   Thu Oct 12 23:52:59 2006 -0400
6718
6719    Add library support for xcb-xinerama.
6720    
6721    This is version 1.1 of the Xinerama (PanoramiX) extension.
6722
6723commit 7f74dd6b643f7769701c31587bf3cfa9c5a8e980
6724Author: Josh Triplett <josh@freedesktop.org>
6725Date:   Sun Oct 8 18:16:14 2006 -0700
6726
6727    Add note to xcbxlib.h that nothing except Xlib/XCB should use it.
6728
6729commit 431f210bbb20e39a24ac879af5f2a8a7744e6a6c
6730Author: Jamey Sharp <jamey@minilop.net>
6731Date:   Sat Oct 7 19:50:55 2006 -0700
6732
6733    Release libxcb 1.0 RC2 (0.9.92).
6734
6735commit e0fac22caaf27b3e461807f8c563d0457938baa6
6736Author: Jamey Sharp <jamey@minilop.net>
6737Date:   Sat Oct 7 18:35:53 2006 -0700
6738
6739    Bugfix: make Plan 7 'checked' requests work correctly.
6740    
6741    The initial implementation of Plan 7 dumped all X errors into the event
6742    queue, because the record of a pending reply was pruned too early if an
6743    error occurred in place of the expected reply.
6744
6745commit 7b84d8b650a611d4d76083340e50cb7ea815014d
6746Author: Ian Osgood <iano@quirkster.com>
6747Date:   Sat Oct 7 14:09:29 2006 -0700
6748
6749    Document xcb_generate_id.
6750
6751commit 9e10819a678970928c9f9a1fffb4ba00f5ac7c57
6752Author: Josh Triplett <josh@freedesktop.org>
6753Date:   Sat Oct 7 13:16:43 2006 -0700
6754
6755    Move xcb_generate_id from xcbext.h to xcb.h
6756    
6757    Since extensions no longer provide type-specific XID-generation functions,
6758    xcb_generate_id now forms part of the xcb client API, rather than the
6759    extension API; move it from xcbext.h to xcb.h accordingly.
6760
6761commit b9e49b2a47e7388e20f8a5c009cdaf0b369e117c
6762Author: Josh Triplett <josh@freedesktop.org>
6763Date:   Sat Oct 7 01:27:26 2006 -0700
6764
6765    Stop implicitly importing xproto; goes with proto change to explicitly import it
6766
6767commit dd5ece606587d3e41058df803ab1804654195bbb
6768Author: Jamey Sharp <jamey@minilop.net>
6769Date:   Sat Oct 7 01:27:03 2006 -0700
6770
6771    Declare "struct foo" or "enum foo" as well as the typedef'd name "foo".
6772
6773commit 16ec51397e0711526457709605cf6b5ebf00b5e2
6774Author: Josh Triplett <josh@freedesktop.org>
6775Date:   Sat Oct 7 00:14:50 2006 -0700
6776
6777    Handle "xidunion" instead of "union" for XID unions like DRAWABLE and FONTABLE
6778
6779commit 7b786bd8f34e67617beda4f952a5a636a370396a
6780Author: Jamey Sharp <jamey@minilop.net>
6781Date:   Fri Oct 6 23:32:18 2006 -0700
6782
6783    api_conv.pl: replace xcb_*_new with calls directly to xcb_generate_id, now that we have no type-safety XID wrappers.
6784
6785commit c73ff37b907c8af6d03ab69cca27bc837c1cc189
6786Author: Josh Triplett <josh@freedesktop.org>
6787Date:   Fri Oct 6 20:11:19 2006 -0700
6788
6789    Refer to the "xlib lock" rather than the "IOLock" in xcbxlib.h
6790
6791commit 90eeb461be4975ff35d187ba64fedbb16ff13d15
6792Author: Josh Triplett <josh@freedesktop.org>
6793Date:   Fri Oct 6 17:44:53 2006 -0700
6794
6795    Remove XID wrapper structures and replace them with uint32_t typedefs
6796    
6797    After positive feedback from several people, we have decided to remove the XID
6798    wrapper structures that attempted to provide C type safety, and replace them
6799    with uint32_t typedefs.  Feedback has indicated that these type-safety hacks
6800    generated more trouble than help.
6801    
6802    We will bump the libxcb soname at the next release.
6803
6804commit 9bd2c0c58fbe4c994e76619c2d67e46ed8b43030
6805Author: Jamey Sharp <jamey@minilop.net>
6806Date:   Fri Oct 6 16:47:50 2006 -0700
6807
6808    Remove xcb_get_io_lock from the Xlib-specific API: it is no longer used.
6809
6810commit 34168ab549fdf7c6c8a6fbe39824116698d01b12
6811Author: Jamey Sharp <jamey@minilop.net>
6812Date:   Fri Oct 6 16:12:04 2006 -0700
6813
6814    Remove the 'int *error' out-parameter for xcb_poll_for_event.
6815
6816commit 40589db8124b8c72894deb86a825c6117b0a2cd2
6817Author: Jamey Sharp <jamey@minilop.net>
6818Date:   Wed Oct 4 15:01:00 2006 -0700
6819
6820    Add xcb_xlib_lock and xcb_xlib_unlock, a special-purpose two-level recursive lock just for libX11.
6821
6822commit 57b0cd8fea498a32ff2322583c7278d5e86aa4e8
6823Author: Jamey Sharp <jamey@minilop.net>
6824Date:   Wed Oct 4 12:23:45 2006 -0700
6825
6826    Factor out pthread_mutex_lock and unlock calls for the iolock.
6827
6828commit e7f473afbd02c87cc6b1fc9c7c240d6c5cc26763
6829Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6830Date:   Sat Sep 30 13:06:58 2006 +0200
6831
6832    fix some indentation, remove args of the main function (not needed), xcb_sync -> xcb_aux_sync. Jamey, it seems that your script transforms XCBGetSetup into a type, instead of a function. Same for XCBSetupVendor. Maybe there are other fuctions that have the same problem.
6833
6834commit 408e74538676230e289d79c6b0195b464bfc6606
6835Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6836Date:   Sat Sep 30 12:45:12 2006 +0200
6837
6838    update tutorial to the new api. Add some doc about cursors. Add some colors to section 6 (Opening and closing the connection to an X server). Tell me if you want to keep colors or not (colors can be modified in xcb.css)
6839
6840commit 15993fccd0b0cba870b793345fb5c5ec13ece0c8
6841Author: Jamey Sharp <jamey@minilop.net>
6842Date:   Mon Sep 25 05:12:07 2006 -0700
6843
6844    Release libxcb 1.0 RC1 (0.9.91).
6845
6846commit f7cd80142fa83d531850df2c2a3b4a514129170b
6847Author: Josh Triplett <josh@freedesktop.org>
6848Date:   Mon Sep 25 05:10:37 2006 -0700
6849
6850    Stop installing the protocol descriptions for extensions to an extensions/
6851    subdirectory
6852
6853commit 0d7beff580232d4b7adfc840bcd1cae894b57682
6854Author: Josh Triplett <josh@freedesktop.org>
6855Date:   Mon Sep 25 04:43:04 2006 -0700
6856
6857    Actually install xcb-xlib.pc, and ship xcb-xlib.pc.in in distribution tarballs
6858
6859commit d04e7777d3547de2cece0734c0b9be49d3ddb88c
6860Author: Jamey Sharp <jamey@minilop.net>
6861Date:   Mon Sep 25 04:14:26 2006 -0700
6862
6863    Ensure api_conv.pl is idempotent on names of constants: XCB_[A-Z0-9_]+.
6864
6865commit 0bccf7d2eed45ba15a012d090b9b950075e75803
6866Author: Jamey Sharp <jamey@minilop.net>
6867Date:   Mon Sep 25 04:03:01 2006 -0700
6868
6869    Move header files to $includedir/xcb (generally $prefix/include/xcb).
6870
6871commit 9100981d359f6366ab92b456e6069bac89afeb15
6872Author: Josh Triplett <josh@freedesktop.org>
6873Date:   Mon Sep 25 02:02:16 2006 -0700
6874
6875    Split the Xlib compatibility functions into a separate library libxcb-xlib
6876    
6877    We don't want to have to change the libxcb soname if we later manage to remove
6878    the Xlib compatibility functions, and nothing except Xlib should ever use
6879    them, so split them into a separate library.
6880
6881commit 810dfb886247dc4484886aff2e2e7d80ea945de4
6882Author: Jamey Sharp <jamey@minilop.net>
6883Date:   Mon Sep 25 02:10:44 2006 -0700
6884
6885    Move debian/ directory to new branch "debian".
6886
6887commit 46b7dbdd9306235ae04da9d0e3431a1b58f89048
6888Author: Jamey Sharp <jamey@minilop.net>
6889Date:   Mon Sep 25 02:00:23 2006 -0700
6890
6891    We no longer need xproto: do not list it as a dependency.
6892
6893commit cf80bab8e6410915ad27a61cef3bb5f456c7d80f
6894Author: Jamey Sharp <jamey@minilop.net>
6895Date:   Mon Sep 25 01:21:42 2006 -0700
6896
6897    Lowercase library names to correspond with the new API.
6898
6899commit a1a7646d5a791c2e918bc1a9f46d64334d5ab575
6900Author: Ian Osgood <iano@quirkster.com>
6901Date:   Sun Sep 24 14:10:20 2006 -0700
6902
6903    Add XCBAllPlanes from xcb-image util library.
6904
6905commit dfbde9a4e972ed9bbd701fd6f89c3a6e6f641740
6906Author: Josh Triplett <josh@freedesktop.org>
6907Date:   Sun Sep 24 03:15:39 2006 -0700
6908
6909    Integrate top-level .gitignore into .gitignore for each subdirectory
6910    
6911    In preparation for the repository split, move the relevant contents of the
6912    top-level .gitignore into the .gitignore for each immediate subdirectory.
6913
6914commit 905379e4474c0137e5dd535798ae0afb07070df6
6915Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6916Date:   Sun Sep 24 08:51:20 2006 +0200
6917
6918    now, I know how to use api_conv.pl :)
6919
6920commit 16516a5e94205edffddde546abc833106c68068e
6921Author: Josh Triplett <josh@freedesktop.org>
6922Date:   Sat Sep 23 14:44:15 2006 -0700
6923
6924    Remove the xcb_[extension]_init functions; use xcb_get_extension_data directly
6925
6926commit 9691890529fddc051c15e191a8a5f06017514c1c
6927Author: Josh Triplett <josh@freedesktop.org>
6928Date:   Sat Sep 23 14:17:52 2006 -0700
6929
6930    More fixups for incorrect API conversions by api_conv.pl
6931
6932commit 0a867d72337afa88f4ab453b6d3fbb519368c308
6933Author: Josh Triplett <josh@freedesktop.org>
6934Date:   Sat Sep 23 13:59:40 2006 -0700
6935
6936    Convert the XCB test suite to the new API
6937
6938commit 63b38f31974b0a61c923a62a2d224e7024a0fc2b
6939Author: Jamey Sharp <jamey@minilop.net>
6940Date:   Sat Sep 23 13:51:10 2006 -0700
6941
6942    Apply const-names.xsl to the un-renamed XML; replace the XSL with the result.
6943
6944commit 86a4c0cc284366bbb01898a77df360278d5a8194
6945Author: Josh Triplett <josh@freedesktop.org>
6946Date:   Sat Sep 23 13:33:27 2006 -0700
6947
6948    We do not conflict with Xmd.h anymore; remove the include hack from xcb_auth.c
6949
6950commit 4ff12a1a94895837dc85ac9e37fd1a5f40819cf3
6951Author: Josh Triplett <josh@freedesktop.org>
6952Date:   Sat Sep 23 13:31:13 2006 -0700
6953
6954    Fix some mis-conversions by api_conv.pl, and remove the now-unused Xmd types
6955
6956commit a3bd6f4760b5b3f5f360a690920839646e2b9d06
6957Author: Josh Triplett <josh@freedesktop.org>
6958Date:   Sat Sep 23 12:22:22 2006 -0700
6959
6960    The Great XCB Renaming
6961    
6962    Rename API to follow a new naming convention:
6963    * XCB_CONSTANTS_UPPERCASE_WITH_UNDERSCORES
6964    * xcb_functions_lowercase_with_underscores
6965    * xcb_types_lowercase_with_underscores_and_suffix_t
6966    * expand all abbreviations like "req", "rep", and "iter"
6967    
6968    Word boundaries for the names in the protocol descriptions fall:
6969    * Wherever the protocol descriptions already have an underscore
6970    * Between a lowercase letter and a subsequent uppercase letter
6971    * Before the last uppercase letter in a string of uppercase letters followed
6972      by a lowercase letter (such as in LSBFirst between LSB and First)
6973    * Before and after a string of digits (with exceptions for sized types like
6974      xcb_char2b_t and xcb_glx_float32_t to match the stdint.h convention)
6975    
6976    Also fix up some particular naming issues:
6977    * Rename shape_op and shape_kind to drop the "shape_" prefix, since otherwise
6978      these types end up as xcb_shape_shape_{op,kind}_t.
6979    * Remove leading underscores from enums in the GLX protocol description,
6980      previously needed to ensure a word separator, but now redundant.
6981    
6982    This renaming breaks code written for the previous API naming convention.  The
6983    scripts in XCB's tools directory will convert code written for the old API to
6984    use the new API; they work well enough that we used them to convert the
6985    non-program-generated code in XCB, and when run on the old program-generated
6986    code, they almost exactly reproduce the new program-generated code (modulo
6987    whitespace and bugs in the old code generator).
6988    
6989    Authors: Vincent Torri, Thomas Hunger, Josh Triplett
6990
6991commit ca72e777740f917753f3c180fdfdb55df91c09c4
6992Author: Jamey Sharp <jamey@minilop.net>
6993Date:   Sat Sep 23 01:33:45 2006 -0700
6994
6995    Special cases to agree with the conventions Josh and I have hashed out.
6996
6997commit 4168ddc13dff7bf2479c2229f42b114d75098112
6998Author: Jamey Sharp <jamey@minilop.net>
6999Date:   Fri Sep 22 22:53:17 2006 -0700
7000
7001    Handle remaining incorrect API translations in api_conv.pl.
7002    
7003    The big change is treating numbers as separate "words" in the
7004    translation, which leads to funny-looking names like xcb_char_2_b_t, but
7005    makes more sense than the alternative in other situations and was
7006    suggested on the mailing list.
7007    
7008    This version still disagrees with Vincent's last proposed XSL, but I
7009    think my output is now preferable. Hopefully Josh has been thinking
7010    along the same lines.
7011
7012commit 91568d7070e2c2400dfdbdff515f51da74ca3701
7013Author: Jamey Sharp <jamey@minilop.net>
7014Date:   Fri Sep 22 15:37:49 2006 -0700
7015
7016    Make xcb_conn.c agree that XCBSetupReq is now XCBSetupRequest.
7017
7018commit 2538acd8c89d73e8c074849c63857cb899a29bbb
7019Author: Alan Coopersmith <alan.coopersmith@sun.com>
7020Date:   Fri Sep 22 11:53:21 2006 -0700
7021
7022    List xproto.xml path explicitly instead of relying on GNU Make $< expansion
7023    
7024    (Some non-gnu makes, such as Solaris make, only recognize $< in implicit
7025     suffix rules, not explicit ones like this.)
7026
7027commit 2c8b5994b3fbba343199ef555594a32e29d8bcee
7028Author: Jamey Sharp <jamey@minilop.net>
7029Date:   Fri Sep 15 01:51:05 2006 -0700
7030
7031    Shut down the connection in all "fatal" error cases.
7032
7033commit 79e3227022ae3d66f9f3806d231fdeec2a06cc6b
7034Author: Jamey Sharp <jamey@minilop.net>
7035Date:   Fri Sep 15 01:57:53 2006 -0700
7036
7037    Add a private connection shutdown method for error cases.
7038
7039commit 7f71bf9c0f30536e85907b2c991cb7001861e1d3
7040Author: Jamey Sharp <jamey@minilop.net>
7041Date:   Fri Sep 15 01:09:27 2006 -0700
7042
7043    Make all public functions do nothing on an error connection.
7044
7045commit 0aa96bfc7abe18889cd85bfaa05b05d53e572bb1
7046Author: Jamey Sharp <jamey@minilop.net>
7047Date:   Fri Sep 15 00:39:51 2006 -0700
7048
7049    Convert connection functions to return error objects.
7050
7051commit ac17ae62fe1a3a29991e36e35eeee838ac4acb00
7052Author: Jamey Sharp <jamey@minilop.net>
7053Date:   Fri Sep 15 00:29:39 2006 -0700
7054
7055    Provide a "has error" property for XCBConnection.
7056
7057commit df7fb77d6e22be76ca73f111c586db99a60178ae
7058Author: Jamey Sharp <jamey@minilop.net>
7059Date:   Thu Sep 21 12:13:37 2006 -0700
7060
7061    Refactor XCBPollForEvent with a shorter critical section.
7062    
7063    This simplifies the patch for bug #8208 later.
7064
7065commit b08ca2b4b451a94ece20207766cd5262fd55179b
7066Author: Jamey Sharp <jamey@minilop.net>
7067Date:   Mon Sep 18 16:52:00 2006 -0700
7068
7069    XCB*Id is a variable, not a type: adjust API conventions accordingly.
7070
7071commit 3cc0ddf0e6b72ba553ebfc2b9126d62ef4013bd3
7072Author: Jamey Sharp <jamey@minilop.net>
7073Date:   Wed Sep 13 12:30:11 2006 -0700
7074
7075    Fix bug #7261: events do not signal the end of replies for that sequence number.
7076
7077commit c912187f15c560c46768acb88aae30e67b0f78ce
7078Author: Jamey Sharp <jamey@minilop.net>
7079Date:   Mon Sep 18 00:25:26 2006 -0700
7080
7081    Handle names of constants during API conversion.
7082    
7083    Use an XSLT stylesheet to get a list of all the constant names.
7084
7085commit 06fba014435cfbdd1ff284d6d513d114503d02c2
7086Author: Jamey Sharp <jamey@minilop.net>
7087Date:   Sun Sep 17 23:38:57 2006 -0700
7088
7089    Complete rewrite of api_conv.pl.
7090    
7091    Now handles all API changes except constant names, which are treated like type names.
7092
7093commit bf41af718d6a83532d1c2f63ac16e6484e8e8b68
7094Author: Jamey Sharp <jamey@minilop.net>
7095Date:   Thu Sep 14 00:04:47 2006 -0700
7096
7097    Simplify the API conversion tool without functionality changes.
7098
7099commit cdffbdd7ef9dee58b3c36ca46bb88aa187b46526
7100Author: Thomas Coppi <thisnukes4u@gmail.com>
7101Date:   Wed Sep 13 23:50:23 2006 -0700
7102
7103    Prototype API conversion tool for upcoming lowercased XCB API.
7104
7105commit a92716f1da2741fca850b3c37299e80032726276
7106Author: Jamey Sharp <jamey@minilop.net>
7107Date:   Wed Sep 13 12:15:23 2006 -0700
7108
7109    Finish removing deprecated functions. Fixes build failure (oops).
7110
7111commit 476ccc1ba3d20f3a545b84089b6fbd40576e7bf5
7112Author: Jamey Sharp <jamey@minilop.net>
7113Date:   Tue Sep 12 13:43:17 2006 -0700
7114
7115    Ignore xproto.xml that now gets copied in while building.
7116
7117commit 0d7fb3afdd308d714a97144125a69a5f4976916a
7118Author: Jamey Sharp <jamey@minilop.net>
7119Date:   Tue Sep 12 13:25:49 2006 -0700
7120
7121    Remove deprecated public API functions in preparation for 1.0 release. Xlib-specific deprecation remains.
7122
7123commit d4e768fc32ff9ce79a4259d252f4a4a4b11d5db8
7124Author: Bart Massey <bart@cs.pdx.edu>
7125Date:   Tue Sep 12 00:42:11 2006 -0700
7126
7127    Cleaned up previous fix for GNU Make stuff.
7128    Got "make distcheck" to work (whatever that does).
7129
7130commit 0c3e528563f8e8db45c340e7d050a1b42e65cb73
7131Author: Bart Massey <bart@bartfan.localdomain>
7132Date:   Mon Sep 11 23:29:10 2006 -0700
7133
7134    Removed GNU make dependency.
7135    Made extensions directory persist.
7136
7137commit 81d6fd64c09255e90104b3c6e37c9023bd0aa6ec
7138Author: Jamey Sharp <jamey@minilop.net>
7139Date:   Sat Sep 9 15:52:37 2006 -0700
7140
7141    Handle XC-MISC request failure when allocating new XIDs. (fixes #7001)
7142
7143commit f6d4fc007fc0f4bdf8f06ab0798aadce87d468d1
7144Author: Thomas Hunger <hto@arcor.de>
7145Date:   Mon Aug 14 06:59:46 2006 -0700
7146
7147    Tutorial corrections.
7148
7149commit dbd2d9689305ce5bf24aebd894551bb6b590f6e4
7150Author: Josh Triplett <josh@freedesktop.org>
7151Date:   Sun Jul 30 22:02:37 2006 -0700
7152
7153    Implement error handling plan 7.
7154    
7155    Needs improvement: should not duplicate the code of every request function.
7156
7157commit 65ed274f05ba670eb02a55b098aed141fa9611ec
7158Author: Josh Triplett <josh@freedesktop.org>
7159Date:   Sun Jul 30 20:28:16 2006 -0700
7160
7161    Parameterize context in cookie-type and request-matching template.
7162
7163commit f74487e34fd61d11a501c07030f8fd7ed8caeb1c
7164Author: Jamey Sharp <jamey@minilop.net>
7165Date:   Sun Jul 30 14:55:14 2006 -0700
7166
7167    Fix Keith's 32-bit wrap fix.
7168    
7169    Issue 0, 1, or 2 syncs as needed and do not handle 16-bit wrap until
7170    absolutely necessary.
7171
7172commit 13896d8f658b917e891aa18e0ac4906d630881bb
7173Author: Josh Triplett <josh@freedesktop.org>
7174Date:   Sun Jul 30 15:09:03 2006 -0700
7175
7176    Fix typo: s/request/sequence/
7177
7178commit 685d8c76c25fd9f236fa1a74bae61699eaa78579
7179Author: Josh Triplett <josh@freedesktop.org>
7180Date:   Sun Jul 30 14:41:19 2006 -0700
7181
7182    Add XCBRequestCheck function needed for Plan 7.
7183
7184commit b3a2f83f1e4d3567fcd4494e7bae31e99645ef85
7185Author: Eric Anholt <anholt@FreeBSD.org>
7186Date:   Fri Jul 28 15:29:05 2006 -0700
7187
7188    Restore netinet/in.h include that was removed in the DECnet commit.
7189
7190commit b7c96681b2bfc968b198058122c93050feb7927e
7191Author: Eric Anholt <anholt@FreeBSD.org>
7192Date:   Tue Jul 25 09:13:59 2006 -0700
7193
7194    Replace HAVE_* tests with just checking for (__solaris__)
7195    
7196    The installed headers can't be relying on the presence of the internal
7197    config.h defines, and it was breaking the xcb build for me as well due to
7198    config.h not being included early enough.
7199
7200commit 14faffc326b9ab1f738edb3ec7febdbe0f5bc079
7201Author: TORRI Vincent <torri@doursse.(none)>
7202Date:   Tue Jul 25 08:24:59 2006 +0200
7203
7204    fix closing <b> tag
7205
7206commit 8ba93ebb757706ae8dcf5af57c6e3812f72de80e
7207Author: Ian Osgood <iano@quirkster.com>
7208Date:   Fri Jul 7 07:18:41 2006 -0700
7209
7210    Fix build on Solaris (use inttypes.h)
7211
7212commit e0574a617b9133cd728be8ea3618ef1312eeebbf
7213Author: Ian Osgood <iano@quirkster.com>
7214Date:   Sat Jul 1 00:55:08 2006 -0700
7215
7216    Finally remove X.h from xcb.h, fix broken image tests.
7217
7218commit db2504130bc450bd328830060cb3a243dd06b52f
7219Author: Keith Packard <keithp@neko.keithp.com>
7220Date:   Sat Jul 1 00:16:07 2006 -0700
7221
7222    Switch sequence comparisons to handle 32-bit sequence number wrap.
7223    
7224    Create a macro, XCB_SEQUENCE_COMPARE, that accepts two sequence numbers and
7225    a comparison operator and correctly handles 32-bit wrap around.
7226    Rewrite all ordered sequence number comparisons to use this macro.
7227    Also, caught one error where a sequence was stored in a signed int variable.
7228    Push out a GetInputFocus request when the sequence number does wrap at 32
7229    bits so that applications cannot see sequence 0 (as that is an error
7230    indicator).
7231
7232commit 75fead5b868a0dfdc9e6fd5ef0dd37eb71761312
7233Author: Josh Triplett <josh@freedesktop.org>
7234Date:   Thu Jun 15 03:03:13 2006 -0700
7235
7236    Remove arbitrary division between xcb_types and xproto by merging
7237    xcb_types.xml into xproto.xml.
7238
7239commit 62749d54fd79b12123607599d58add126ce5de6e
7240Author: TORRI Vincent <torri@doursse.(none)>
7241Date:   Thu Jun 8 06:59:47 2006 +0200
7242
7243    fix compilation with c++ compilers. Remove some trailing spaces
7244
7245commit 829188cdd6274b5b5271dff5612b30c978a61a38
7246Author: Ian Osgood <iano@quirkster.com>
7247Date:   Tue May 30 11:54:25 2006 -0700
7248
7249    Deprecate XCBSync, move to XCBAuxSync.
7250
7251commit 936077cbc87addc914d33ab79a7d066f0f51d3ad
7252Author: Jamey Sharp <jamey@minilop.net>
7253Date:   Sun May 14 22:49:18 2006 -0700
7254
7255    Use correct word offset when testing for GetFBConfigsSGIX VendorPrivate.
7256
7257commit 442730a9a25644e6d09065cdde2f1595ea65caf3
7258Author: Josh Triplett <josh@freedesktop.org>
7259Date:   Sun May 14 22:37:55 2006 -0700
7260
7261    In the GLX workaround, use !strcmp to check for equality with "GLX", not strcmp.
7262
7263commit 2e49f58e4cd670e6bd6a0006833277cfb1da60e6
7264Author: Josh Triplett <josh@freedesktop.org>
7265Date:   Wed May 10 14:22:27 2006 -0700
7266
7267    Stop overwriting CFLAGS in xcb/src/Makefile.am.
7268
7269commit 92456577dbbe3874d6e40a1ef26b63dd405c160c
7270Author: Donnie Berkholz <spyderous@gentoo.org>
7271Date:   Sun May 7 15:44:37 2006 -0700
7272
7273    Fix dependencies on libXCB.la to not walk down from $(top_builddir); this fixes parallel builds.
7274
7275commit 4a891c6f4e49a04866b2cac7bf8bb679a256d3b4
7276Author: Josh Triplett <josh@freedesktop.org>
7277Date:   Sat Apr 29 17:02:43 2006 -0700
7278
7279    Work around the unsetting of CFLAGS in src/Makefile.am, by passing CFLAGS to make via DEB_MAKE_INVOKE in debian/rules.
7280
7281commit 140c30e522c664f0675b63301bea22e4062cd155
7282Author: Josh Triplett <josh@freedesktop.org>
7283Date:   Sat Apr 29 16:28:01 2006 -0700
7284
7285    Remove unnecessary dirs files.
7286
7287commit 105c6d67e6fa3d636068a2a4407bfa6adfafe0cc
7288Author: Josh Triplett <josh@freedesktop.org>
7289Date:   Sat Apr 29 16:09:06 2006 -0700
7290
7291    Stop setting DEB_CONFIGURE_EXTRA_FLAGS = --with-opt in debian/rules; cdbs does the right thing on its own, and correctly handles DEB_BUILD_OPTS.
7292
7293commit 911ae3fbdd12bcfa39e6bbd85a7327ddd1cb9287
7294Author: Josh Triplett <josh@freedesktop.org>
7295Date:   Sat Apr 29 15:57:28 2006 -0700
7296
7297    Set DEB_MAKE_CHECK_TARGET=check in debian/rules.
7298
7299commit 3b7cf7ecf8ae614d7cb833ac4ef9fbcde3df05ae
7300Author: Josh Triplett <josh@freedesktop.org>
7301Date:   Sat Apr 29 15:34:11 2006 -0700
7302
7303    Add a Section field for the source package in debian/control.
7304
7305commit 4e26eefcd9a035c4f7ee1d5f267095edc7df7c89
7306Author: Josh Triplett <josh@freedesktop.org>
7307Date:   Sat Apr 29 11:15:47 2006 -0700
7308
7309    Drop library version from source package name.
7310
7311commit eb3521283f91a84ce149fcc416720289ca37e29f
7312Author: Josh Triplett <josh@freedesktop.org>
7313Date:   Sat Apr 29 10:58:23 2006 -0700
7314
7315    Set distribution to experimental.
7316
7317commit 2a18842f0f93516ff726a2dfa44348410bb55626
7318Author: Josh Triplett <josh@freedesktop.org>
7319Date:   Sat Apr 29 10:56:56 2006 -0700
7320
7321    Remove Bugs field in debian/control, so bugs go to the Debian BTS.
7322
7323commit e6132aff2fdcd0ab1ef5d7ac84a3c597dc07d370
7324Author: Ian Osgood <iano@quirkster.com>
7325Date:   Fri Apr 28 15:27:09 2006 -0700
7326
7327    Move the remainder of the constants in X.h into XML enumerations.
7328    Fix xcb_auth to use one of the new enumerations.
7329
7330commit 162c7593adf1577ca8aecfcd53fd5644b6182609
7331Author: Josh Triplett <josh@freedesktop.org>
7332Date:   Fri Apr 28 00:51:53 2006 -0700
7333
7334    Change libxcbxvmc0-dev Depends to libxcbxv0-dev, not libxv0-dev.
7335
7336commit 4f9b6556e2b68d1f05ed4df9d2cc82edb9868872
7337Author: Josh Triplett <josh@freedesktop.org>
7338Date:   Thu Apr 27 23:40:55 2006 -0700
7339
7340    Update debian packaging to create library, -dev, and -dbg packages for each new extension library.
7341
7342commit b7d77a8de79493919281c4f742b1a65c1328b1c4
7343Author: Josh Triplett <josh@freedesktop.org>
7344Date:   Thu Apr 27 17:32:20 2006 -0700
7345
7346    Fix the year in COPYING.
7347
7348commit 5c35ea63db8dcc30bd84eb1ae8f12f0969bb8a63
7349Author: Josh Triplett <josh@freedesktop.org>
7350Date:   Thu Apr 27 17:20:52 2006 -0700
7351
7352    Add an explanation of libtool -version-info to src/Makefile.am, and add an explicit -version-info 0:0:0 for all libraries.
7353
7354commit 08f5cc389275a71821a62c1674622ca85792b3fe
7355Author: Josh Triplett <josh@freedesktop.org>
7356Date:   Thu Apr 27 16:24:35 2006 -0700
7357
7358    Fix tutorial to use XCBSetup rather than the previous deprecated name XCBConnSetupSuccessRep.
7359
7360commit 34016bcdab9b8373b4a8f9520b01d3179d35a315
7361Author: Josh Triplett <josh@freedesktop.org>
7362Date:   Thu Apr 27 16:15:50 2006 -0700
7363
7364    Remove execute bit on tutorial.
7365
7366commit 135cda6e9dafeef36efc8c4be49ae99b417034f2
7367Author: Josh Triplett <josh@freedesktop.org>
7368Date:   Thu Apr 27 13:47:21 2006 -0700
7369
7370    Fix a comment for the renaming of XCBConnSetupSuccessRep to XCBSetup, and fix another comment which had a copy-paste error.
7371
7372commit 17c3448f58c1aaf1d89b706dfa00e697061115c7
7373Author: Josh Triplett <josh@freedesktop.org>
7374Date:   Thu Apr 27 12:33:41 2006 -0700
7375
7376    Rename xcb-xfree86dri.pc.in to xcb-xf86dri.pc.in to match libXCBxf86dri, change its Libs line to use -lXCBxf86dri, and change configure.ac and Makefile.am accordingly.
7377
7378commit ef18582d83fac146b1a689bd60d580819dba0ffd
7379Author: Josh Triplett <josh@freedesktop.org>
7380Date:   Thu Apr 27 12:04:52 2006 -0700
7381
7382    Fix typo and rephrase checking message.
7383
7384commit 3e213a1becd3081e905e3711ff3fe3e4865cd7dd
7385Author: Jamey Sharp <jamey@minilop.net>
7386Date:   Thu Apr 27 10:34:22 2006 -0700
7387
7388    Declare arguments const in the various structure accessors, so XCBGetSetup is not so painful to use.
7389
7390commit dd932e025b2de0fa33b65971288c63679238e08a
7391Author: Jamey Sharp <jamey@minilop.net>
7392Date:   Wed Apr 26 23:19:16 2006 -0700
7393
7394    Rename ConnSetup* to Setup*, Setup*Rep to Setup*, and SetupSuccess* to Setup*. Provide deprecated backwards-compatability functions and typedefs for the old names, to be removed before 1.0.
7395
7396commit b825f3385becccfcde34626f872721301e82f960
7397Author: Josh Triplett <josh@freedesktop.org>
7398Date:   Wed Apr 26 11:23:32 2006 -0700
7399
7400    Add .pc.in files for all the new extension libraries.  Generate .pc files from the new .pc.in files in configure.ac.  Install the new .pc files in Makefile.am, and add the new .pc.in files to EXTRA_DIST.
7401    
7402    Based on a patch by Vincent Torri.  Changes from that patch: add Requires to the .pc.in files based on the <import>s in the extensions, add a .pc.in file for libXCBxtest, change Name and Description fields to match extension names.
7403
7404commit 1476dcecfd3e817236ae996314c10482aae8e3e9
7405Author: Ian Osgood <iano@quirkster.com>
7406Date:   Wed Apr 26 10:50:31 2006 -0700
7407
7408    Changes to makefile to build libXCBxtest
7409
7410commit bb8cf58015bae083dd77f8679f9a8299603c58a6
7411Author: Patrick Caulfield <patrick@tykepenguin.com>
7412Date:   Mon Apr 24 08:29:18 2006 -0700
7413
7414    Add support for DECnet. Still needs configure-script options to enable.
7415
7416commit 771761ccaad31d029d470dde84279e94494310b6
7417Author: Jamey Sharp <jamey@minilop.net>
7418Date:   Thu Apr 20 11:51:01 2006 -0700
7419
7420    Minor performance improvement: do not call _xcb_in_expect_reply unless it is needed. It is not often needed.
7421
7422commit 8953a14f2b1518042ed0745574e22fa2adfb6cac
7423Author: Jamey Sharp <jamey@minilop.net>
7424Date:   Wed Apr 19 22:48:10 2006 -0700
7425
7426    Add <sys/select.h> to xcb_in.c to fix bug #6122.
7427
7428commit 05a66af895442b9fceb96c9130e77694927eabaf
7429Author: Alan Coopersmith <alan.coopersmith@sun.com>
7430Date:   Wed Apr 19 21:51:33 2006 -0700
7431
7432    Bugfix: xcb_conn.c included <sys/fcntl.h> instead of the POSIX-standard <fcntl.h>.
7433
7434commit f705456744fe4beb193d27eb64fa9157102db753
7435Author: Alan Coopersmith <alan.coopersmith@sun.com>
7436Date:   Wed Apr 19 21:40:42 2006 -0700
7437
7438    Only set CWARNFLAGS to gnu flags if $GCC is set. Otherwise check for Sun compiler and use its enhanced warning flag.
7439
7440commit 6f369fde3c676e2b1a67bd71923b61942991d726
7441Merge: 922cb61 f090da9
7442Author: Ian Osgood <iano@quirkster.com>
7443Date:   Wed Apr 19 20:45:31 2006 -0700
7444
7445    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7446
7447commit f090da98f367ed869fd9277d2fef22555be0f91d
7448Author: Jamey Sharp <jamey@minilop.net>
7449Date:   Wed Apr 19 20:31:20 2006 -0700
7450
7451    Remove the last goto in XCB: XCBWaitForReply now permits multiple threads to force the same cookie.
7452
7453commit d5347485a55e58381781d803e19bfdd982a4685b
7454Author: Jamey Sharp <jamey@minilop.net>
7455Date:   Wed Apr 19 20:23:37 2006 -0700
7456
7457    Restructure XCBWaitForReply to eliminate two gotos.
7458
7459commit d5ab03b4b71648bf0a06e42e3c288a68c46ba497
7460Author: Jamey Sharp <jamey@minilop.net>
7461Date:   Wed Apr 19 20:15:15 2006 -0700
7462
7463    Fixed poll_for_reply, added comments, and refactored XCBWaitForReply to call poll_for_reply.
7464
7465commit 7667adbc631119ec39f3ef5a316aec42dbf5f393
7466Author: Jamey Sharp <jamey@minilop.net>
7467Date:   Wed Apr 19 16:49:32 2006 -0700
7468
7469    Add XCBPollForReply and deprecate XCBGetRequestRead and XCBGetQueuedRequestRead.
7470
7471commit 53971ea183d9d1dcfbaec18b135e49c9c118fabb
7472Merge: 71de16f 8275ac3
7473Author: Josh Triplett <josh@freedesktop.org>
7474Date:   Sun Apr 16 11:23:52 2006 -0700
7475
7476    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7477
7478commit 71de16fac2a145d5ef8069d2d28d7c32cea603cf
7479Author: Josh Triplett <josh@freedesktop.org>
7480Date:   Sun Apr 16 11:05:50 2006 -0700
7481
7482    Improve package descriptions.
7483
7484commit e92bde6e5152c6d0c4efa9240604e75178c1a3db
7485Author: Josh Triplett <josh@freedesktop.org>
7486Date:   Sun Apr 16 09:59:13 2006 -0700
7487
7488    Change Depends on x-dev to x11proto-core-dev.
7489
7490commit eca61f6b5e9df7321222499a1f660f6eb7e6112e
7491Author: Josh Triplett <josh@freedesktop.org>
7492Date:   Sun Apr 16 09:25:36 2006 -0700
7493
7494    Add Build-Depends on libxdmcp-dev.
7495
7496commit a0057d7a48b90b8f11fc9d5c82b5b8d800c34db5
7497Author: Josh Triplett <josh@freedesktop.org>
7498Date:   Sun Apr 16 09:24:01 2006 -0700
7499
7500      * Debian X11R7 transition:
7501        * Change Build-Depends on x-dev to x11proto-core-dev.
7502        * Install headers to /usr/include/X11, not /usr/X11R6/include/X11.
7503        * Pre-Depends: x11-common (>= 1:1.09).
7504
7505commit 11c62f7d9d65c10c796c2199c73c8f167e53f234
7506Author: Josh Triplett <josh@freedesktop.org>
7507Date:   Sat Apr 15 22:50:33 2006 -0700
7508
7509    Stop installing the libtool .la files.
7510
7511commit 057ae541a3a73cffd58533029292c1c721fa3162
7512Author: Josh Triplett <josh@freedesktop.org>
7513Date:   Sat Apr 15 20:44:11 2006 -0700
7514
7515    Update -dbg package handling to work with debhelper compat level 5.
7516    Increase minimum version on debhelper Build-Depends to 5.0.0.
7517
7518commit 8eedb4a487dcede0e52849e36c4da13bdf0c8b51
7519Author: Josh Triplett <josh@freedesktop.org>
7520Date:   Sat Apr 15 20:22:05 2006 -0700
7521
7522    Stop running autoreconf in debian/rules, and remove Build-Depends for
7523    autoconf, automake, and libtool.
7524
7525commit cc075990f4fc2ed09c708036569049ddd24605ac
7526Author: Josh Triplett <josh@freedesktop.org>
7527Date:   Sat Apr 15 20:21:22 2006 -0700
7528
7529    Add xcbint.h to noinst_HEADERS, so it gets distributed.
7530
7531commit cb6e1849b66c17f96d79598adb740ed16325a9c1
7532Author: Josh Triplett <josh@freedesktop.org>
7533Date:   Sat Apr 15 14:17:52 2006 -0700
7534
7535    Use screensaver.xml
7536
7537commit 91aeea2a3e72af16733ddcba037541440c0c4739
7538Author: Josh Triplett <josh@freedesktop.org>
7539Date:   Sat Apr 15 12:52:05 2006 -0700
7540
7541    Put EXTHEADERS and EXTSOURCES in order.
7542
7543commit ff38c17c48c271847d12c81cbf80142c6918dc78
7544Author: Josh Triplett <josh@freedesktop.org>
7545Date:   Sat Apr 15 12:26:03 2006 -0700
7546
7547    Split all non-essential extensions into their own separate libraries, named
7548    libXCBextname.  To use extension extname, include extname.h and link with
7549    -lXCBextname.  This allows extensions to change without bumping the main
7550    libXCB version.
7551    
7552    bigreq and xc_misc remain in libXCB, because XCB uses them internally to make
7553    big requests and to allocate XIDs, respectively.
7554
7555commit 8275ac3a4a23220b5c3b4f191a45befe2d34d6bd
7556Author: Eric Anholt <anholt@FreeBSD.org>
7557Date:   Sun Apr 9 19:51:10 2006 -0700
7558
7559    Retry a select() if it returns with EINTR.  Fixes IO errors in Xephyr, which is
7560    often interrupted by timers.
7561
7562commit 66a88ed0e556ca869ddc9df5a35e3d6446d12b02
7563Author: Eric Anholt <anholt@FreeBSD.org>
7564Date:   Sun Apr 9 19:19:12 2006 -0700
7565
7566    Remove unnecessary include.  Noticed by jamey.
7567
7568commit 922cb6137a12982ecd9e1c73ecefdcbc9e193eae
7569Author: Ian Osgood <iano@quirkster.com>
7570Date:   Wed Mar 22 17:57:57 2006 -0800
7571
7572    Work on the tutorial, and update xproto.xml to match:
7573    * Fixed grammar
7574    * Answered some TODO's and added some more
7575    * Updated X.h constants to those in xproto.h
7576    * Added enumerations used in tutorial to xproto.xml
7577    * Prefered XCBFlush to XCBSync
7578    * Corrected and refactored the "events" example
7579    
7580    I extracted the examples to test them. Where should I put them?
7581
7582commit fc577b81bfeb79bb78ee529278ed52d59d489f89
7583Author: Josh Triplett <josh@freedesktop.org>
7584Date:   Tue Mar 21 14:22:21 2006 -0800
7585
7586    Remove outdated fd.o-* entries from */debian/.gitignore (obsolete since before they came from .cvsignore).
7587
7588commit df5d8adc1f18776e4417a03b465dae9273511fb1
7589Author: Jamey Sharp <jamey@minilop.net>
7590Date:   Thu Mar 16 11:27:06 2006 -0800
7591
7592    The typedefs replacing Xmd.h conflict with Xmd.h. Here is a hacky workaround: FIXME!
7593
7594commit 5ccf7216d266f22755e40a8b9858ebd60fd48f40
7595Author: Ian Osgood <iano@quirkster.com>
7596Date:   Wed Mar 15 22:41:08 2006 -0800
7597
7598    Change <bit> syntax to <bit>n</bit> as Jamey and Josh suggested
7599     and add more button, key, and graphics enumerations.
7600
7601commit 35ecaf45be75cb78fe18dd4ea9d564b03f6cea67
7602Author: Ian Osgood <iano@quirkster.com>
7603Date:   Wed Mar 15 10:09:47 2006 -0800
7604
7605    Remove proto/X11 dependencies from xcb-demo.
7606
7607commit 4142e34695310c20a8a30af64e0b40842bfb5dbe
7608Author: Ian Osgood <iano@quirkster.com>
7609Date:   Tue Mar 14 18:23:37 2006 -0800
7610
7611    Remove xcb-util dependency on proto/X11
7612    by moving many defs from X.h to <enum>s in xproto.xml
7613
7614commit f5c4956a81ed5aee3ea46e91785ec14f82e205d5
7615Author: Ian Osgood <iano@quirkster.com>
7616Date:   Tue Mar 14 11:53:01 2006 -0800
7617
7618    Fix XCBGetSetup to match prototype.
7619
7620commit 9472c251ae426bb496a3112d0ae390f1f1c35515
7621Merge: 71ddf29 0d648ac
7622Author: Ian Osgood <iano@quirkster.com>
7623Date:   Tue Mar 14 11:44:43 2006 -0800
7624
7625    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7626
7627commit 71ddf29d00ad8ff2323faf2cc6c354cbf3eca16c
7628Author: Ian Osgood <iano@quirkster.com>
7629Date:   Tue Mar 14 10:18:22 2006 -0800
7630
7631    Add an expression construct <bit bit="n"/>
7632    for mask enumerations such as CW flags.
7633    Replaces the C-specific use of 1<<n in xproto.xml
7634
7635commit 0d648ac0ab3a2d457284644e677fd2ed612f3e7c
7636Author: TORRI Vincent <torri@doursse.(none)>
7637Date:   Tue Mar 14 07:40:39 2006 +0100
7638
7639    doxygen documentation of the API
7640
7641commit a810d1ffe4825b9359f13ebb395f5f681961fc89
7642Author: Ian Osgood <iano@quirkster.com>
7643Date:   Mon Mar 13 17:49:17 2006 -0800
7644
7645    Remove dependencies on Xmd.h and X.h
7646    (Still including X.h until defs are moved to xproto.xml
7647     and xcb-util and xcb-demo are fixed to use them.)
7648
7649commit be1302b6efb33967bce5356af58e3e0ae3b19363
7650Author: Ian Osgood <iano@quirkster.com>
7651Date:   Mon Mar 13 10:36:13 2006 -0800
7652
7653    Remove last deprecation warning.
7654
7655commit d8de2c7c2a9b8bf59c8f1fcece5faed0b6cbf6f4
7656Merge: 5e0cfa8 e757673
7657Author: Ian Osgood <iano@quirkster.com>
7658Date:   Sun Mar 12 23:03:24 2006 -0800
7659
7660    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7661
7662commit e7576738c33e73fb4f29c1426c2ec49257564129
7663Author: Ian Osgood <iano@quirkster.com>
7664Date:   Sun Mar 12 23:02:45 2006 -0800
7665
7666    Restructure to remove most deprecation warnings.
7667
7668commit 5e0cfa84bec3e04f3a1991baaa29da09c7c4a02c
7669Author: Jamey Sharp <jamey@minilop.net>
7670Date:   Sun Mar 12 17:15:50 2006 -0800
7671
7672    Use libXdmcp, if available with XdmcpWrap, for XDM-AUTHORIZATION-1. Closes bug #6106.
7673
7674commit d69c403cba9bdebd1bd41b62ae7e28f5852248d4
7675Merge: b83f18a 5cdc02e
7676Author: Jamey Sharp <jamey@minilop.net>
7677Date:   Sun Mar 12 13:36:33 2006 -0800
7678
7679    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7680
7681commit b83f18a4cc2303dfda59807d56e16bbc5c18b09d
7682Author: Jamey Sharp <jamey@minilop.net>
7683Date:   Sun Mar 12 13:20:29 2006 -0800
7684
7685    Only _xcb_conn_wait calls _xcb_out_write now, so move it to xcb_conn.c and make it static.
7686
7687commit fb61c94d685a254ef0702a2e2093b8cdda02d514
7688Author: Jamey Sharp <jamey@minilop.net>
7689Date:   Sun Mar 12 12:40:34 2006 -0800
7690
7691    Remove c->out.vec. Pass iovecs directly down the call tree. Add _xcb_out_flush_to, refactor other functions, make write_block static.
7692
7693commit 5cdc02e3441dc623e91ebc5d7f735565b83f8435
7694Author: Jamey Sharp <jamey@minilop.net>
7695Date:   Sat Mar 11 20:32:04 2006 -0800
7696
7697    Portability fixes. Should help DragonFly and Solaris, and should not hurt anything else. Tested only on linux so far though.
7698
7699commit 83e652f566671f96ffc53a3c0099a84a1606c695
7700Author: Jamey Sharp <jamey@minilop.net>
7701Date:   Thu Mar 9 00:02:42 2006 -0800
7702
7703    Move c->out.vec refs out of _xcb_conn_wait up to _xcb_out_flush.
7704
7705commit 5b0ae3421dd373a8575b7a0d60989edfc056cf71
7706Merge: fd1f9cb 1261a41
7707Author: Jamey Sharp <jamey@minilop.net>
7708Date:   Wed Mar 8 22:50:48 2006 -0800
7709
7710    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7711
7712commit fd1f9cb13f9ab9a6bb6aa0c894d4891809bece8a
7713Author: Jamey Sharp <jamey@minilop.net>
7714Date:   Wed Mar 8 22:49:59 2006 -0800
7715
7716    Unify autofoo .gitignore bits into one top-level file, and remove resulting redundancies.
7717
7718commit 1261a4150c8bd39a6fb6887df52abbbb446f7ffb
7719Author: Ian Osgood <iosgood@Titania.local>
7720Date:   Wed Mar 8 22:26:20 2006 -0800
7721
7722    Add symbols for error and reply response_types,
7723    and use XCBKeymapNotify from xcb-proto.
7724
7725commit 6fb661f3ff2fca342b4ea76d5a583c5e4f53e076
7726Author: Ian Osgood <iosgood@Titania.local>
7727Date:   Wed Mar 8 21:56:57 2006 -0800
7728
7729    Bugfix: null-terminate the path in sockaddr_un before using it.
7730    This may fix itermittant connect failures.
7731
7732commit 621f891c49cbf4beba1e20fb9b6fb1be576d42f3
7733Author: Jamey Sharp <jamey@minilop.net>
7734Date:   Wed Mar 8 14:21:16 2006 -0800
7735
7736    Move c->out.vec refs out of _xcb_out_write up to _xcb_conn_wait.
7737
7738commit c491eeb9a9f670f7d4869d7dae7a5adce4565998
7739Author: Jamey Sharp <jamey@minilop.net>
7740Date:   Tue Mar 7 21:19:58 2006 -0800
7741
7742    Fix off-by-one error that kept the last byte(s) of the output queue from being used.
7743
7744commit 522a6e0eac9adeaac533a5b700f42d85d46e2dd7
7745Author: TORRI Vincent <Vincent.Torri@iecn.u-nancy.fr>
7746Date:   Tue Mar 7 10:25:23 2006 -0800
7747
7748    Use the GCC 4 visibility extension to mark everything in xcbint.h hidden.
7749
7750commit 5437032c7dee85da99b612a707fa94d012d40282
7751Merge: 2ad5450 d1cfd4d
7752Author: Jamey Sharp <jamey@minilop.net>
7753Date:   Tue Mar 7 00:26:38 2006 -0800
7754
7755    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7756
7757commit d1cfd4d0a21feaa5ccf0a1fd693327b8c5718abf
7758Author: Jamey Sharp <jamey@minilop.net>
7759Date:   Mon Mar 6 01:10:20 2006 -0800
7760
7761    Off-by-one error in the sequence-wrapping proof, and therefore in the corresponding code.
7762
7763commit 2ad5450ff3566a6e378769d7ff68f148c992b468
7764Author: TORRI Vincent <torri@doursse.(none)>
7765Date:   Sun Mar 5 09:49:02 2006 +0100
7766
7767    wrong spaces
7768
7769commit 81ea81c6d9a67fdf664fc502fe3610b37aea8354
7770Author: TORRI Vincent <torri@doursse.(none)>
7771Date:   Sun Mar 5 09:41:11 2006 +0100
7772
7773    fix some font sizes
7774
7775commit 66364da372dc1a2e1d75729a1265d632ae642ad6
7776Author: TORRI Vincent <torri@doursse.(none)>
7777Date:   Sun Mar 5 09:40:15 2006 +0100
7778
7779    fixes + valid css
7780
7781commit 01656220ef275a811a284420121d37bf9aa85f70
7782Author: TORRI Vincent <torri@doursse.(none)>
7783Date:   Sun Mar 5 09:22:03 2006 +0100
7784
7785    fixes + valid css
7786
7787commit e5458e477db95e1e064e46ca28245ecd51b5b524
7788Author: Jamey Sharp <jamey@minilop.net>
7789Date:   Sun Mar 5 00:20:50 2006 -0800
7790
7791    Implement provably-correct sequence wrap handling. Add flag XCB_REQUEST_DISCARD_REPLY.
7792
7793commit 6659c8c63b82329468b249cc99181d72ec26c698
7794Author: TORRI Vincent <torri@doursse.(none)>
7795Date:   Sun Mar 5 09:05:21 2006 +0100
7796
7797    lots of fixes. Thanks to Indan Zupancic
7798
7799commit 6e4745bbd0c924d846550496a072b0f2b0d4482c
7800Author: TORRI Vincent <torri@doursse.(none)>
7801Date:   Sun Mar 5 07:54:20 2006 +0100
7802
7803    lots of fixes. Thanks to Indan Zupancic
7804
7805commit df9c7cb4f9b05caf07900f3c15be379d3f9979fb
7806Author: Josh Triplett <josh@freedesktop.org>
7807Date:   Sat Mar 4 18:35:53 2006 -0800
7808
7809    Add *.tar.{bz2,gz} to .gitignore for the benefit of "make distcheck".
7810
7811commit 29f9fe0fc805a1ec6860f167a45664cc1cf0c769
7812Author: Jamey Sharp <jamey@minilop.net>
7813Date:   Fri Mar 3 11:08:10 2006 -0800
7814
7815    API/ABI change: XCBSendRequest returns the sequence number instead of using an out-parameter. Now 0 is a special sequence number indicating failure.
7816
7817commit 255c21b17f61147388bab6e1d42623a008a4a8d2
7818Author: Jamey Sharp <jamey@minilop.net>
7819Date:   Fri Mar 3 01:45:00 2006 -0800
7820
7821    Add 32-bit full_sequence fields to generic errors and events, so callers can always correlate events and errors with requests correctly.
7822
7823commit 622b599c8fcf5d677f76ca03f3241a23dba58712
7824Author: Jamey Sharp <jamey@minilop.net>
7825Date:   Thu Mar 2 23:39:38 2006 -0800
7826
7827    Tweak to previous API change: Require that spare iovecs fall before vector[0]. Leave vector in well-defined state.
7828
7829commit 87905f0579e749ac6d92843780af246160318eca
7830Author: Jamey Sharp <jamey@minilop.net>
7831Date:   Thu Mar 2 23:31:35 2006 -0800
7832
7833    assert() that XCBSendRequest was handed enough space to set opcodes and short length fields.
7834
7835commit 45e1cc0935ca1d7a138e1c700db229f50205a556
7836Author: Jamey Sharp <jamey@minilop.net>
7837Date:   Thu Mar 2 23:12:21 2006 -0800
7838
7839    Bugfix: The fixed-length part needs two iovecs reserved, just like all the other parts.
7840
7841commit b46953f46c7d0fd8cbb73f922524968c1ac86fa4
7842Author: Jamey Sharp <jamey@minilop.net>
7843Date:   Thu Mar 2 23:05:08 2006 -0800
7844
7845    Use XPath position() function instead of xsl:number, to evaluate array indices while generating code for easier readability.
7846
7847commit 5e115e2441ed32f5fa495370b36b01c03bbff66d
7848Author: Jamey Sharp <jamey@minilop.net>
7849Date:   Thu Mar 2 15:35:31 2006 -0800
7850
7851    API/ABI change: XCBSendRequest callers must pad to 4-byte boundaries now. When not in RAW mode, a null pointer for iov_base is replaced by up to 3 padding bytes.
7852
7853commit ed823bf65192a72f8c3060698c9bded9f77d49c2
7854Author: Jamey Sharp <jamey@minilop.net>
7855Date:   Thu Mar 2 13:49:00 2006 -0800
7856
7857    Separate notion of request-completed from current-request, and mark requests completed more aggressively. Detects some usage errors that are otherwise undetectable.
7858
7859commit c05ae15b661bff6b95deb3abc7f48abe16892ac8
7860Author: Jamey Sharp <jamey@minilop.net>
7861Date:   Mon Feb 27 12:12:33 2006 -0800
7862
7863    Buffer a couple CARD32s on the stack instead of using an extra iovec. Also a bugfix: do not hold a reference to longlen after it goes out of scope.
7864
7865commit 8f991bdd389f6c419cb18cdaea966304529de165
7866Author: Jamey Sharp <jamey@minilop.net>
7867Date:   Mon Feb 27 11:03:13 2006 -0800
7868
7869    Add XCB_REQUEST_RAW flag for XCBSendRequest.
7870
7871commit 5e350126a728f3c0c3bc8d8673e5ad67dc174f79
7872Author: Jamey Sharp <jamey@minilop.net>
7873Date:   Mon Feb 27 10:53:08 2006 -0800
7874
7875    Bugfix: how about *not* leaking all pending_replies when no reply matches, as often happens with Xlib?
7876
7877commit 731c85762d8994119f4eaf572cc59c9adbb7abd9
7878Author: Jamey Sharp <jamey@minilop.net>
7879Date:   Mon Feb 27 02:57:36 2006 -0800
7880
7881    Move test for other writing threads *before* allocating a sequence number.
7882
7883commit 8ab4dcde9db8175ef0093123bd04c72471e205ff
7884Author: Jamey Sharp <jamey@minilop.net>
7885Date:   Mon Feb 27 02:14:48 2006 -0800
7886
7887    _xcb_list is no longer used, so remove it. Simplify _xcb_map.
7888
7889commit 76ad79a7adbb315f7c7d5f08d6f42553210868cd
7890Author: Jamey Sharp <jamey@minilop.net>
7891Date:   Mon Feb 27 01:43:07 2006 -0800
7892
7893    Replace readers generic list with hand-written typesafe version.
7894
7895commit fa71b9b835da6360c374dcb4ab93be3a27dbd737
7896Author: Jamey Sharp <jamey@minilop.net>
7897Date:   Mon Feb 27 01:06:15 2006 -0800
7898
7899    Rename struct XCBReplyData to reader_list to follow my naming conventions and better reflect its purpose.
7900
7901commit ff665b57266b9e7e9b9a366272c2115bbd516173
7902Author: Jamey Sharp <jamey@minilop.net>
7903Date:   Mon Feb 27 00:54:59 2006 -0800
7904
7905    In struct XCBReplyData, change void *data to pthread_cond_t *data. That was dumb.
7906
7907commit 50acfeae36f5f67e8b92fd7610141c489ee348c5
7908Author: Jamey Sharp <jamey@minilop.net>
7909Date:   Sun Feb 26 23:43:44 2006 -0800
7910
7911    Delete unused xcb_list functions and refactor others.
7912
7913commit 86ce18c22c8daebe8796d3c38e8a1d496ab6e8de
7914Author: Jamey Sharp <jamey@minilop.net>
7915Date:   Sun Feb 26 22:58:17 2006 -0800
7916
7917    Replace current_reply generic queue with hand-written version. No generic queues remain so delete their implementation.
7918
7919commit 0f130b4d945a27fd2b4655c351ebe70d61cac598
7920Author: Jamey Sharp <jamey@minilop.net>
7921Date:   Sun Feb 26 18:28:50 2006 -0800
7922
7923    Replace events generic queue with hand-written typesafe version.
7924
7925commit ff7b6c9124e7caf26381cc7a10fba9eaf1875652
7926Author: Jamey Sharp <jamey@minilop.net>
7927Date:   Sun Feb 26 15:45:08 2006 -0800
7928
7929    API/ABI break: Add flags to XCBSendRequest, first for error-checked requests.
7930    There's no more race condition between event and reply handling.
7931    The *RequestBlind and *RequestChecked functions are not yet implemented.
7932
7933commit 7875040fa179d8ca3fa4ec046c1b8c2a351a6621
7934Author: Jamey Sharp <jamey@minilop.net>
7935Date:   Sun Feb 26 02:00:03 2006 -0800
7936
7937    Replace pending_replies generic queue with a hand-implemented typesafe version.
7938
7939commit a1eff0c49a2dbb328a79edbeb2b851fa1dfe9e61
7940Author: Jamey Sharp <jamey@minilop.net>
7941Date:   Sun Feb 26 01:27:01 2006 -0800
7942
7943    Replace my old generic map ADT with a growable array for the extension cache.
7944
7945commit 46a754998149c5f4a1670787b3ea36731caf6506
7946Author: Jamey Sharp <jamey@minilop.net>
7947Date:   Sat Feb 25 23:27:47 2006 -0800
7948
7949    XCB has not had tracing features for a long time: remove the remnants.
7950
7951commit b6cbe837748df5819d00682330d34e2c7d24d349
7952Author: Jamey Sharp <jamey@minilop.net>
7953Date:   Sat Feb 25 23:26:55 2006 -0800
7954
7955    Rearrange an if statement that's been bothering me.
7956
7957commit f27166f49b9ef6bdcce78429bffc724d1e4fb360
7958Author: Jamey Sharp <jamey@minilop.net>
7959Date:   Fri Feb 24 02:17:36 2006 -0800
7960
7961    Coalesce _xcb_writev into _xcb_out_write and simplify.
7962
7963commit 9463653b1e6dc0a9054266aa3eecb0839129b991
7964Author: Jamey Sharp <jamey@minilop.net>
7965Date:   Fri Feb 24 01:56:18 2006 -0800
7966
7967    Take advantage of Requires and *.private fields for a more accurate pkg-config file.
7968
7969commit bae98d36040d0cda1862839410cf92bac3927280
7970Author: Jamey Sharp <jamey@minilop.net>
7971Date:   Fri Feb 24 01:50:48 2006 -0800
7972
7973    Move _xcb_set_fd_flags to xcb_conn.c and make it static. xcb_util.c now has only public functions.
7974
7975commit 67b2649dc4b6726c6d11fb0e41429ae5de82b0e8
7976Author: Jamey Sharp <jamey@minilop.net>
7977Date:   Fri Feb 24 01:40:45 2006 -0800
7978
7979    Move _xcb_read_block to xcb_in.c and make it static. Change calls in xcb_conn.c to _xcb_in_read_block instead.
7980
7981commit 838317f4d305e07f35ddd99e2ebccfa8fac680c6
7982Author: Jamey Sharp <jamey@minilop.net>
7983Date:   Fri Feb 24 01:17:03 2006 -0800
7984
7985    Remove XCB_CEIL and use a simpler definition for XCB_PAD.
7986
7987commit 1b50d2ee1ef532429674126eace88ac73d51ec23
7988Author: Jamey Sharp <jamey@minilop.net>
7989Date:   Fri Feb 24 01:04:48 2006 -0800
7990
7991    Quit using "-include config.h": use #ifdef HAVE_CONFIG_H etc. like everyone else.
7992
7993commit e1c2777abac0479ebdf06f2005120d25fff53935
7994Author: Jamey Sharp <jamey@minilop.net>
7995Date:   Fri Feb 24 00:59:08 2006 -0800
7996
7997    Coalesce readn into _xcb_in_read and simplify.
7998
7999commit 3f8d0bd5322749132626e8f203017b6da6448fd0
8000Author: Jamey Sharp <jamey@minilop.net>
8001Date:   Fri Feb 24 00:48:18 2006 -0800
8002
8003    Move _xcb_readn to xcb_in.c and make it static. Minor change to _xcb_read_block to not depend on _xcb_readn.
8004
8005commit cdf362f33a7bc4d412231de32e37ffb65a06d3f1
8006Author: Jamey Sharp <jamey@minilop.net>
8007Date:   Fri Feb 24 00:25:34 2006 -0800
8008
8009    Bugfix: protect the output queue from being written while another thread is flushing it.
8010
8011commit 4e91ae275e46e76de64ab089d6f13c1e9b6ddee6
8012Author: Jamey Sharp <jamey@minilop.net>
8013Date:   Fri Feb 24 00:02:43 2006 -0800
8014
8015    Simplify: Always use writev. (In _xcb_out_flush, convert the output queue to a single iovec if needed.)
8016
8017commit 7f0bc778c88ab2f565cc05d5d3d5ee4c8d1388a1
8018Author: Jamey Sharp <jamey@minilop.net>
8019Date:   Thu Feb 23 22:41:59 2006 -0800
8020
8021    Factor padding out of _xcb_out_write_block and into its callers, XCBSendRequest and write_setup.
8022    This requires dynamically allocating memory in XCBSendRequest, but this
8023    malloc/free pair turns out to cause a 30% speed hit for the 'x11perf -noop'
8024    test -- so for the moment I use alloca where available and fall back to malloc
8025    on other platforms. Later I think I'll change the contract of XCBSendRequest
8026    so the caller is responsible for memory allocation, because the caller ought
8027    to always be able to stack-allocate here.
8028
8029commit 6e29e5f2ee2e6158f1a9480a83e4f906ab9c04d1
8030Author: Jamey Sharp <jamey@minilop.net>
8031Date:   Thu Feb 23 17:50:53 2006 -0800
8032
8033    Add XCBGetQueuedRequestRead for Xlib that does no syscalls, just returns whatever XCB already knows about.
8034
8035commit 26ac6292ba0535ac3747d23f2f4d284c4b5f42ba
8036Author: Jamey Sharp <jamey@minilop.net>
8037Date:   Thu Feb 23 15:29:40 2006 -0800
8038
8039    Minor performance fix: Change the contract on XCBGetRequestSent so that it does not waste time re-locking. This is for Xlib, and Xlib has already locked.
8040
8041commit a736674943295245ccb90865c13d2b75dc6ecccb
8042Author: Jamey Sharp <jamey@minilop.net>
8043Date:   Thu Feb 23 14:57:46 2006 -0800
8044
8045    Minor performance fix: Only rearrange buffers for BIG-REQUESTs.
8046
8047commit e866bed9348bf3ffd480fd6ec329fd8b9581c328
8048Author: Jamey Sharp <jamey@minilop.net>
8049Date:   Thu Feb 23 14:32:11 2006 -0800
8050
8051    Move request_written update back where it was for now: doing it early can cause XCBWaitForReply to wrongly believe that the request has been flushed. Eventually, we should fix bug #6021.
8052
8053commit 1b83f8f8f326eca9d8852c82dd36696f81a720dc
8054Author: Jamey Sharp <jamey@minilop.net>
8055Date:   Thu Feb 23 14:30:08 2006 -0800
8056
8057    Remove _xcb_assert_valid_sequence. One test is trivially true, and the other may be temporarily violated without anything bad happening.
8058
8059commit 55c1842686d2e668708cd106b5e08847df0184c3
8060Author: Jamey Sharp <jamey@minilop.net>
8061Date:   Thu Feb 23 12:48:27 2006 -0800
8062
8063    Move _xcb_write and _xcb_writev to xcb_out.c and make them static, since only _xcb_out_write calls them.
8064
8065commit 213b5725928ccf8aedc807cc40a261b2d5431247
8066Author: Jamey Sharp <jamey@minilop.net>
8067Date:   Thu Feb 23 12:34:08 2006 -0800
8068
8069    Ignore test byproducts.
8070
8071commit 6149c7a6b57a193bc132fbc35b1772b75e5c7ca7
8072Author: Jamey Sharp <jamey@minilop.net>
8073Date:   Thu Feb 23 12:15:09 2006 -0800
8074
8075    More return value changes, and make _xcb_in_read_packet static since it is not called from outside xcb_in.c.
8076
8077commit 5b1d39e27b1a966df537ead248da5a57e7d9de97
8078Author: Jamey Sharp <jamey@minilop.net>
8079Date:   Thu Feb 23 11:50:12 2006 -0800
8080
8081    More return value changes.
8082
8083commit 662479760c42fc38c458381ee3eaed92e2c8b733
8084Author: Jamey Sharp <jamey@minilop.net>
8085Date:   Thu Feb 23 10:17:40 2006 -0800
8086
8087    Make the return value of _xcb_conn_wait boolean, instead of syscall-like.
8088
8089commit f8a8b465e080faf5479c0b38dc67cc9b9ae24927
8090Author: Jamey Sharp <jamey@minilop.net>
8091Date:   Wed Feb 22 23:38:16 2006 -0800
8092
8093    Bugfix: move request_written update *before* _xcb_conn_wait in _xcb_out_flush. Otherwise a reply may be read before we record that we have sent the request, and then XCB gets... confused.
8094
8095commit e0a35783837e0d4974cfa20c6a7a073698f3b0a5
8096Author: Jamey Sharp <jamey@minilop.net>
8097Date:   Wed Feb 22 23:11:36 2006 -0800
8098
8099    Remove USE_THREAD_ASSERT compile option. _xcb_conn_wait can no longer be re-entered, since we have no callback hooks now.
8100
8101commit a7d749ec3fd3303a4e7ace9d4f0f1672f9310ef2
8102Author: Jamey Sharp <jamey@minilop.net>
8103Date:   Wed Feb 22 23:09:29 2006 -0800
8104
8105    Remove vestiges of adjacent request combining implementation. It is very dead.
8106
8107commit 06c788932bb635da79076728927b7b93f64037c1
8108Author: Jamey Sharp <jamey@minilop.net>
8109Date:   Mon Feb 20 00:41:39 2006 -0800
8110
8111    Control debugging, optimization, and warning flags all independently.
8112
8113commit 3f98dae4a8274f2430c4967f8354356ef7c45024
8114Merge: e346f8c f825a32
8115Author: Jamey Sharp <jamey@minilop.net>
8116Date:   Sat Feb 18 23:57:18 2006 -0800
8117
8118    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
8119
8120commit e346f8c8efe6f931aa43026087d0ea406258c094
8121Author: Jamey Sharp <jamey@minilop.net>
8122Date:   Sat Feb 18 23:57:14 2006 -0800
8123
8124    Update .gitignores for .o files and autofoo stuff.
8125
8126commit f825a32135fa613dd6b6dd8553e366d8125bd7a4
8127Author: Eric Anholt <anholt@FreeBSD.org>
8128Date:   Sat Feb 18 23:05:32 2006 -0800
8129
8130    Add missing entries to .gitignore files.
8131
8132commit 9e013b564cc4479802c51f79ea5054d5f25edacc
8133Merge: d16efb5 7474473
8134Author: Eric Anholt <anholt@FreeBSD.org>
8135Date:   Sat Feb 18 22:57:42 2006 -0800
8136
8137    Merge branch 'master' of git+ssh://anholt@git.freedesktop.org/git/xcb
8138
8139commit d16efb51915d9661788c96140635514cf6314d46
8140Author: Eric Anholt <anholt@FreeBSD.org>
8141Date:   Sat Feb 18 22:56:03 2006 -0800
8142
8143    Move .cvsignore to .gitignore.
8144
8145commit 7474473f3312597f6de736b8f25cedc433427092
8146Author: Jamey Sharp <jamey@minilop.net>
8147Date:   Sat Feb 18 20:03:35 2006 -0800
8148
8149    Quit defining _XOPEN_SOURCE. We never needed it...
8150
8151commit 1d977c7b2d519e22f319075563e7faf4bbd31d58
8152Author: Jamey Sharp <jamey@id.minilop.net>
8153Date:   Sat Feb 18 19:23:33 2006 -0800
8154
8155    Remove the old ChangeLog file. GIT makes that obsolete.
8156
8157commit 5b7182c659391160239467f1041a1d755db45bd3
8158Author: Jamey Sharp <jamey@minilop.net>
8159Date:   Sat Feb 18 18:12:40 2006 -0800
8160
8161    Workaround X server bug, fd.o #3210: if a GLXGetFBConfigs request is sent, fix the length field in the reply.
8162
8163commit 47ceed78612d48dcda62cc2686fc527d61abe38b
8164Author: Josh Triplett <josh@freedesktop.org>
8165Date:   Sat Feb 18 16:49:41 2006 -0800
8166
8167    Remove xcl and CVSROOT.
8168