ChangeLog revision 8ffb90f1
1commit c2c4a2cd1947e559718acdba19ef6e7db731dbeb
2Author: Matt Turner <mattst88@gmail.com>
3Date:   Tue May 3 15:09:54 2022 -0700
4
5    libxcb 1.15
6    
7    Signed-off-by: Matt Turner <mattst88@gmail.com>
8
9commit ddafdba11f6919e6fcf977c09c78b06f94de47aa
10Author: Hodong <hodong@yozmos.com>
11Date:   Sat Jan 15 02:32:04 2022 +0900
12
13    Fix a memory leak
14    
15    Signed-off-by: Hodong <hodong@yozmos.com>
16
17commit 43fbf03e549bf6da8d1d8522e0ceddc4d49c37c6
18Author: Demi Marie Obenour <demiobenour@gmail.com>
19Date:   Thu Sep 23 16:46:32 2021 -0400
20
21    Fix integer overflows in xcb_in.c
22    
23    This fixes an integer overflow security vulnerability in xcb_in.c, which
24    may allow for memory corruption.
25
26commit 233d7b7f1f03ef18bf3955eb1f20421e745d22f0
27Author: Thomas Anderson <thomasanderson@google.com>
28Date:   Wed Dec 2 00:25:42 2020 +0000
29
30    Fix hang in xcb_request_check()
31    
32    This fixes https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/53
33    
34    The issue was that libxcb expected to get a reply based on the request_expected
35    variable, but a reply would never arrive because the request was never actually
36    written.  To resolve this, a separate request_expected_written variable is
37    added.
38
39commit dc2811874729ee83fa2aef110f60808c450f9a5a
40Author: Ran Benita <ran@unusedvar.com>
41Date:   Tue Nov 17 23:18:53 2020 +0200
42
43    Avoid request counter truncation in replies map after 2**32 requests
44    
45    The c->in request counters are uint64_t, and can realistically go over
46    2**32 over a lifetime of a client. The c->in->replies map however uses
47    unsigned int keys and the passed request numbers are silently truncated.
48    
49    I haven't analyzed in depth what happens what it wraps around but it's
50    probably nothing good.
51    
52    The only user of the xcb_list.c map code is c->in->replies, so just
53    change it to use uint64_t keys.
54    
55    Reviewed-by: Uli Schlachter <psychon@znc.in>
56    Signed-off-by: Ran Benita <ran@unusedvar.com>
57
58commit 26396bf156cfa00ecd655ec6a5f2f46626d674c0
59Author: Julien Cristau <jcristau@debian.org>
60Date:   Tue Feb 2 12:15:10 2021 +0100
61
62    Add newline when printing auth/connection failure string to stderr
63    
64    The reason strings returned by the server don't all include a newline,
65    so make sure we add one to avoid confusing clients.  Xlib used to do
66    this before it delegated that work to libxcb.
67    
68    Fixes #34
69    
70    Signed-off-by: Julien Cristau <jcristau@debian.org>
71
72commit a503167f751ba77e6819df568b7f5042d2baa9c9
73Author: Uli Schlachter <psychon@znc.in>
74Date:   Sat May 22 21:43:22 2021 +0200
75
76    Improve/fix docs for reply fds functions
77    
78    Fixes: https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/56
79    Signed-off-by: Uli Schlachter <psychon@znc.in>
80
81commit 3c76c0579ffa521af41f78ee937b1337581da985
82Author: Povilas Kanapickas <povilas@radix.lt>
83Date:   Fri Jul 30 22:58:47 2021 +0300
84
85    c_client.py: Implement handling of <length> element
86    
87    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
88
89commit bdc3f21a5205293852bbaa173f43389cbf66f52b
90Author: Povilas Kanapickas <povilas@radix.lt>
91Date:   Fri Jul 30 22:58:46 2021 +0300
92
93    c_client: Extract _c_get_field_mapping_for_expr()
94
95commit 068af21cb376eb824fa1dee29a6539feadb93587
96Author: Povilas Kanapickas <povilas@radix.lt>
97Date:   Fri Jul 30 22:58:45 2021 +0300
98
99    c_client.py: Use get_expr_field_names directly to resolve list fields
100    
101    Using get_expr_fields() is only needed in case we are doing things that
102    can span multiple types easily, e.g. when deciding what data to pass via
103    function parameters and so on.
104    
105    In _c_serialize_helper_list_field() we are building function body, so
106    acquiring field names via get_expr_field_names() is enough.
107    
108    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
109
110commit 4d678b162bf8a3b10e5bdf76df2be63d33c23381
111Author: Povilas Kanapickas <povilas@radix.lt>
112Date:   Fri Jul 30 22:58:44 2021 +0300
113
114    c_client.py: Extract get_expr_field_names()
115    
116    Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
117
118commit 21414e7c447f18224c577ed5e32bd5d6e45c44f9
119Author: Peter Harris <pharris@opentext.com>
120Date:   Mon Feb 1 19:45:28 2021 -0500
121
122    Fix writev emulation on Windows
123    
124    There are at least two bugs in the previous implementation:
125    
126    - If an early iovec is partially written, there can be a gap of missing
127      data (as a later iovec will be started before the early iovec is
128      completed).
129    - If a late iovec returns WSAEWOULDBLOCK, *vector and *count are not
130      updated, leading to a re-send of the entire request.
131    
132    Move the *vector update into the send() loop to update piecemeal as
133    individual iovecs are sent.
134    
135    Example program that demonstrates the issue (this program should run
136    forever after these bugs have been fixed):
137    
138    #include <stdio.h>
139    #include <stdlib.h>
140    #include "xcb.h"
141    
142    // Non-cryptographic random number generator from http://burtleburtle.net/bob/rand/smallprng.html
143    // because Microsoft's random number generators either have a too small RAND_MAX or are too slow
144    typedef struct ranctx { uint32_t a; uint32_t b; uint32_t c; uint32_t d; } ranctx;
145    
146    static uint32_t ranval(ranctx *x);
147    static void raninit(ranctx *x, uint32_t seed);
148    
149    
150    #define MAX_PROP_LEN (128 * 1024)
151    
152    int main(int argc, char *argv[]) {
153        uint32_t seed = 0x12345678;
154        if (argc > 1) {
155            seed = strtoul(argv[1], NULL, 0);
156        }
157        ranctx ran;
158        raninit(&ran, seed);
159    
160        xcb_connection_t *c = xcb_connect(NULL, NULL);
161        if (!c || xcb_connection_has_error(c)) {
162            printf("Cannot connect to $DISPLAY\n");
163            return 1;
164        }
165        const xcb_setup_t *setup = xcb_get_setup(c);
166        char *buf = malloc(MAX_PROP_LEN + 8); // plus a bit of slack so we can run random values off the end
167        if (!buf) {
168            printf("oom\n");
169            return 1;
170        }
171        for (uint32_t i=0; i < (MAX_PROP_LEN + 3) / 4; i++) {
172            ((uint32_t *)buf)[i] = ranval(&ran);
173        }
174    
175        xcb_window_t win = xcb_generate_id(c);
176        xcb_create_window(c, 0, win, xcb_setup_roots_iterator(setup).data[0].root, 0, 0, 1, 1, 0,
177                XCB_WINDOW_CLASS_INPUT_ONLY, 0, 0, NULL);
178        printf("Created window 0x%X\n", win);
179    
180        for (;;) {
181            xcb_flush(c);
182            xcb_generic_event_t *ev = xcb_poll_for_event(c);
183            if (ev) {
184                if (ev->response_type == 0) {
185                    xcb_generic_error_t *err = (xcb_generic_error_t *)ev;
186                    printf("Unexpected X Error %d\n", err->error_code);
187                    printf("   Sequence %d\n", err->sequence);
188                    printf("   Resource ID 0x%X\n", err->resource_id);
189                    printf("   Opcode: %d.%d\n", err->major_code, err->minor_code);
190                    return 1;
191                }
192                printf("Unexpected X Event %d\n", ev->response_type);
193                return 1;
194            }
195    
196            uint32_t siz = ranval(&ran) % MAX_PROP_LEN + 1;
197            xcb_change_property(c, XCB_PROP_MODE_REPLACE, win, XCB_ATOM_STRING, XCB_ATOM_STRING, 8, siz, buf);
198        }
199    
200        return 0;
201    }
202    
203    
204    #define rot(x,k) (((x)<<(k))|((x)>>(32-(k))))
205    static uint32_t ranval(ranctx *x) {
206        uint32_t e = x->a - rot(x->b, 27);
207        x->a = x->b ^ rot(x->c, 17);
208        x->b = x->c + x->d;
209        x->c = x->d + e;
210        x->d = e + x->a;
211        return x->d;
212    }
213    
214    static void raninit(ranctx *x, uint32_t seed) {
215        uint32_t i;
216        x->a = 0xf1ea5eed, x->b = x->c = x->d = seed;
217        for (i = 0; i<20; ++i) {
218            (void)ranval(x);
219        }
220    }
221    
222    Signed-off-by: Peter Harris <pharris@opentext.com>
223
224commit 4b0d9d3868aad8d5f4266821e9eda586e6c2bfa7
225Author: Peter Harris <pharris@opentext.com>
226Date:   Mon Feb 1 17:43:52 2021 -0500
227
228    Fix build on Windows
229    
230    Notable changes: Protect include of unistd.h (and other POSIX headers).
231    Use SOCKET (which is larger than int) and closesocket (because close is
232    not compatible) for sockets. Use <stdint.h>'s intptr_t instead of the
233    non-portable ssize_t.
234    
235    Signed-off-by: Peter Harris <pharris@opentext.com>
236
237commit cd0fba98a2d0867d505ff1a7ca8d7a7c757acfa2
238Author: Alan Coopersmith <alan.coopersmith@oracle.com>
239Date:   Tue Jun 1 18:36:17 2021 -0700
240
241    xcb_auth: Quiet -Wimplicit-fallthrough warning in get_authptr()
242    
243    xcb_auth.c:135:14: warning: this statement may fall through [-Wimplicit-fallthrough=]
244             addr += 12;
245             ~~~~~^~~~~
246    xcb_auth.c:138:5: note: here
247         case AF_INET:
248         ^~~~
249    
250    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
251
252commit 2ef8655987b9862cdddb72223c9f259a860d5777
253Author: Julien Cristau <jcristau@debian.org>
254Date:   Mon Feb 1 19:23:02 2021 +0100
255
256    Increment libtool version info for libxcb-dri3
257    
258    Somewhat belatedly given the last update was in xcb-proto 1.13 in 2017...
259    
260    Quoting @smcv from https://bugs.debian.org/921069:
261    >>>
262    libxcb-dri3 version 1.13 appears to have added new symbols without increasing
263    the minor ABI version in its -version-info. This will break anything that
264    compares libraries by their version info to decide which one is newer.
265    
266    The Steam Runtime uses libraries' major/minor/micro ABI version info (in this
267    case 0.0.0) to decide whether to use the system copy of a library or the copy
268    in the Steam Runtime, depending on which one is newer (#921026). We can
269    work around this by adding a versioned dependency on libxcb-dri3-0 and
270    deleting the copy from the Steam Runtime, but this isn't a particularly
271    scalable solution.
272    >>>
273
274commit 4cbcac4eca967abfbff7cf1ea473777c5e8c375c
275Author: Ran Benita <ran@unusedvar.com>
276Date:   Tue Nov 17 23:45:14 2020 +0200
277
278    gitignore: add files generated by make check
279    
280    Signed-off-by: Ran Benita <ran@unusedvar.com>
281
282commit f01f3c378eb0168fbb055c7be1c2d08a7acd3752
283Author: Ran Benita <ran@unusedvar.com>
284Date:   Tue Nov 17 23:43:06 2020 +0200
285
286    tests: don't use deprecated fail_unless check API
287    
288    It causes errors like this when running make check:
289    
290    check_public.c:40:24: error: too many arguments for format [-Werror=format-extra-args]
291       40 |   fail_unless(success, "unexpected parse failure %sfor '%s'", test_string[test_type], name);
292    
293    Closes: https://gitlab.freedesktop.org/xorg/lib/libxcb/-/issues/49
294    Tested-by: Matt Turner <mattst88@gmail.com>
295    Signed-off-by: Ran Benita <ran@unusedvar.com>
296
297commit 704e0a91b1dece9a4cce7cfe8beaeb86a5e14267
298Author: Eduardo Sánchez Muñoz <esm@eduardosm.net>
299Date:   Mon Mar 2 18:59:45 2020 +0100
300
301    Use the 'present' field to properly check that the XC-MISC
302    extension is available in xcb_generate_id.
303    
304    Also document the returned value when xcb_generate_id fails.
305
306commit 4b40b44cb6d088b6ffa2fb5cf3ad8f12da588cef
307Author: Matt Turner <mattst88@gmail.com>
308Date:   Sat Feb 22 12:10:53 2020 -0800
309
310    Release libxcb 1.14
311    
312    Signed-off-by: Matt Turner <mattst88@gmail.com>
313
314commit 78c492deaa7ae3aad0b0eeb4b126bb291cc637a7
315Author: Matt Turner <mattst88@gmail.com>
316Date:   Sat Feb 22 12:19:58 2020 -0800
317
318    Build xz tarballs instead of bzip2
319    
320    Signed-off-by: Matt Turner <mattst88@gmail.com>
321
322commit 8f7e4c4e9fc6ee6b5b5498a7d8af79150be0d26a
323Author: Matt Turner <mattst88@gmail.com>
324Date:   Sat Feb 22 11:24:11 2020 -0800
325
326    configure.ac: Depend on pthread-stubs only on not-Linux
327    
328    Signed-off-by: Matt Turner <mattst88@gmail.com>
329
330commit f9f4b00aad69ff36e81c63089b1b16660eaca900
331Author: Sam Varshavchik <mrsam@courier-mta.com>
332Date:   Sat Jan 4 10:43:59 2020 -0500
333
334    Implement xcb_total_read() and xcb_total_written().
335    
336    Returns raw byte counts that have been read or written to the
337    xcb_connection_t.
338    
339    I found it very useful when developing a high level widget toolkit, to
340    track down inefficient/sub-optimum code that generates a lot of X
341    protocol traffic.
342    
343    Signed-off-by: Sam Varshavchik <mrsam@courier-mta.com>
344
345commit 59e271e15bcecf0c461cd5c6c59081fb86b96c22
346Author: A. Wilcox <AWilcox@Wilcox-Tech.com>
347Date:   Mon Dec 23 21:49:29 2019 -0600
348
349    tests: Support Check 0.13.0 API
350    
351    [mattst88]: Keep compatibility  with old API via preprocessor
352    
353    Fixes: #43
354
355commit 21324989b7e121c008a2c4fdf98547541cbf7b83
356Author: Martin Dørum <martid0311@gmail.com>
357Date:   Sun May 19 16:05:08 2019 +0200
358
359    Handle EINTR from recvmsg in _xcb_in_read
360    
361    I have a GTK application which occasionally crashes with an "interrupted
362    system call" g_message from gdk. After a lot of debugging, I've found
363    that the call to recvmsg in _xcb_in_read occasionally fails with EINTR,
364    and instead of retrying the system call, xcb would just shut down the
365    connection.
366    
367    This change makes _xcb_in_read treat EINTR the same as it would treat
368    EAGAIN; it returns 1 and libX11 ends up calling xcb_poll_for_event
369    again (from what I have understood).
370    
371    I have spoken with a few people who think recvmsg failing with EINTR in
372    this case shouldn't ever happen, and I don't know enough to agree or
373    disagree with that. In case anyone wants to dig further and try to
374    figure out why the recvmsg call sometimes fails with EINTR, here's the
375    backtrace from inside of _xcb_in_read where that happened:
376    
377    Thread 1 "beanbar" hit Breakpoint 1, _xcb_in_read (c=c@entry=0x55ecbe4aba80) at xcb_in.c:1059
378    1059                fprintf(stderr, "Hello World am %s:%i, errno is %s\n", __FILE__, __LINE__, strerror(errno));
379    (gdb) bt
380    0  0x00007fa48fa48639 in _xcb_in_read (c=c@entry=0x55ecbe4aba80) at xcb_in.c:1059
381    1  0x00007fa48fa489d8 in poll_for_next_event (c=0x55ecbe4aba80, queued=queued@entry=0) at xcb_in.c:352
382    2  0x00007fa48fa48a3d in poll_for_next_event (queued=0, c=<optimized out>) at xcb_in.c:722
383    3  0x00007fa48fa48a3d in xcb_poll_for_event (c=<optimized out>) at xcb_in.c:722
384    4  0x00007fa4908d1b7e in poll_for_event (dpy=dpy@entry=0x55ecbe4a9730, queued_only=queued_only@entry=0) at xcb_io.c:245
385    5  0x00007fa4908d1cf0 in poll_for_response (dpy=dpy@entry=0x55ecbe4a9730) at xcb_io.c:303
386    6  0x00007fa4908d1fed in _XEventsQueued (mode=2, dpy=0x55ecbe4a9730) at xcb_io.c:363
387    7  0x00007fa4908d1fed in _XEventsQueued (dpy=dpy@entry=0x55ecbe4a9730, mode=mode@entry=2) at xcb_io.c:344
388    8  0x00007fa4908c3d47 in XPending (dpy=0x55ecbe4a9730) at Pending.c:55
389    9  0x00007fa493cadbc7 in  () at /usr/lib/libgdk-3.so.0
390    10 0x00007fa49234d08a in g_main_context_prepare () at /usr/lib/libglib-2.0.so.0
391    11 0x00007fa49234d6e6 in  () at /usr/lib/libglib-2.0.so.0
392    12 0x00007fa49234d8ae in g_main_context_iteration () at /usr/lib/libglib-2.0.so.0
393    13 0x00007fa4938b920e in g_application_run () at /usr/lib/libgio-2.0.so.0
394    14 0x000055ecbc820af4 in main (argc=1, argv=0x7ffd06238098) at src/main.c:190
395    
396    Signed-off-by: Martin Dørum <martid0311@gmail.com>
397
398commit 656c08c5429a3cf53e7abd7fc56cd3c3a79c0f64
399Author: Jon Turney <jon.turney@dronecode.org.uk>
400Date:   Sun Sep 14 00:05:27 2014 +0100
401
402    Include time.h before using time()
403    
404    Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
405
406commit be1745c8eb00defcb31d336ccc142de056e92bd8
407Author: Alan Coopersmith <alan.coopersmith@oracle.com>
408Date:   Sun Feb 17 12:06:10 2019 -0800
409
410    Add README.md to EXTRA_DIST
411    
412    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
413
414commit 58f37377c851960cbf21fac1caf38ab1e66c7d6c
415Author: Eduardo Sánchez Muñoz <esm@eduardosm.net>
416Date:   Sun Feb 17 13:33:12 2019 +0100
417
418    Add "ge.*" to src/.gitignore
419
420commit 7bac366953005233fc3c8736e6f1cd5798e65e28
421Author: Alan Coopersmith <alan.coopersmith@oracle.com>
422Date:   Sat Feb 16 13:41:28 2019 -0800
423
424    Update README for gitlab migration
425    
426    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
427
428commit 02ff3eadf48e6affe3b59ef688312b4dab538e8b
429Author: Alan Coopersmith <alan.coopersmith@oracle.com>
430Date:   Sat Feb 16 13:20:45 2019 -0800
431
432    Update configure.ac bug URL for gitlab migration
433    
434    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
435
436commit 542befe40a3a7c6a5d1dcb7f38fb9eb261b96b24
437Author: Alan Coopersmith <alan.coopersmith@oracle.com>
438Date:   Mon Jan 7 14:42:53 2019 -0800
439
440    c_client: fix "adress" typo
441    
442    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
443
444commit 8287ebd7b752c33b0cabc4982606fe4831106f7e
445Author: Uli Schlachter <psychon@znc.in>
446Date:   Thu Sep 27 14:04:17 2018 +0200
447
448    Release libxcb 1.13.1
449
450commit bbda345a718ff73086437e51f03fcbb73e4365b9
451Author: Erik Kurzinger <ekurzinger@nvidia.com>
452Date:   Mon Aug 20 12:06:25 2018 -0700
453
454    don't flag extra reply in xcb_take_socket
455    
456    If any flags are specified in a call to xcb_take_socket,
457    they should only be applied to replies for requests sent
458    after that function returns (and until the socket is
459    re-acquired by XCB).
460    
461    Previously, they would also be incorrectly applied to the
462    reply for the last request sent before the socket was taken.
463    For instance, in this example program the reply for the
464    GetInputFocus request gets discarded, even though it was
465    sent before the socket was taken. This results in the
466    call to retrieve the reply hanging indefinitely.
467    
468    static void return_socket(void *closure) {}
469    
470    int main(void)
471    {
472        Display *dpy = XOpenDisplay(NULL);
473        xcb_connection_t *c = XGetXCBConnection(dpy);
474    
475        xcb_get_input_focus_cookie_t cookie = xcb_get_input_focus_unchecked(c);
476        xcb_flush(c);
477    
478        uint64_t seq;
479        xcb_take_socket(c, return_socket, dpy, XCB_REQUEST_DISCARD_REPLY, &seq);
480    
481        xcb_generic_error_t *err;
482        xcb_get_input_focus_reply(c, cookie, &err);
483    }
484    
485    In practice, this has been causing intermittent KWin crashes when
486    used in combination with the proprietary NVIDIA driver such as
487    https://bugs.kde.org/show_bug.cgi?id=386370 since when Xlib fails to
488    retrieve one of these incorrectly discarded replies it triggers
489    an IO error.
490    
491    Signed-off-by: Erik Kurzinger <ekurzinger@nvidia.com>
492    Signed-off-by: Uli Schlachter <psychon@znc.in>
493
494commit 7e0f166579672d71efd819c81f0c932b0acd542c
495Author: Daniel Stone <daniels@collabora.com>
496Date:   Wed Feb 28 01:26:55 2018 +0000
497
498    Release libxcb 1.13
499    
500    Signed-off-by: Daniel Stone <daniels@collabora.com>
501
502commit a3e9821bec08a69b355b46b0655562b9df61bb21
503Author: Daniel Stone <daniels@collabora.com>
504Date:   Wed Apr 26 17:55:54 2017 +0200
505
506    c_client: Add support for lists of FDs
507    
508    Matching xcbgen changes, add support having a ListType which contains
509    file descriptors. Use this to send a variable number of FDs to the
510    server, including when the list size is not fixed.
511    
512    Signed-off-by: Daniel Stone <daniels@collabora.com>
513
514commit c7aa4e682fdc7f0035f928af6eafd052e38cc15a
515Author: Daniel Stone <daniels@collabora.com>
516Date:   Wed Apr 26 17:54:01 2017 +0200
517
518    c_client: Don't serialise non-wire fields
519    
520    For when we have a variable-sized field followed by a fixed field, make
521    sure we do not serialise non-wire fields.
522    
523    Signed-off-by: Daniel Stone <daniels@collabora.com>
524
525commit d10194a321c3db851b3ede9a98cdc95e951943aa
526Author: Christian Linhart <chris@demorecorder.com>
527Date:   Sat Mar 11 22:03:34 2017 +0100
528
529    enable xinput by default
530    
531    Support for the xinput extension is complete now,
532    as far as I can tell.
533    
534    According to our discussion on the list, we enable it now.
535    
536    Signed-off-by: Christian Linhart <chris@demorecorder.com>
537
538commit fad81b63422105f9345215ab2716c4b804ec7986
539Author: David McFarland <corngood@gmail.com>
540Date:   Tue Apr 18 23:58:59 2017 -0300
541
542    read from connection when polling special events and replies
543    
544    Using the mesa vulkan driver, if you acquire an image from a
545    swapchain using a finite timeout (x11_acquire_next_image_poll_x11),
546    it will occasionally lock, calling xcb_poll_for_special_event in
547    a loop until the timeout expires.
548    
549    Call _xcb_in_read() once from the polling functions for special
550    events and replies, in the same way as xcb_poll_for_event.
551    
552    Signed-off-by: David McFarland <corngood@gmail.com>
553    Signed-off-by: Uli Schlachter <psychon@znc.in>
554
555commit f830eb93c9c38b2c6c7ea2971af3bc6a61e92277
556Author: Tobias Stoeckmann <tobias@stoeckmann.org>
557Date:   Sun Mar 26 15:41:12 2017 +0200
558
559    Check strdup for NULL return value.
560    
561    _xcb_open does not check strdup's return value for NULL if launchd suport
562    was configured.
563    
564    Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
565    Signed-off-by: Uli Schlachter <psychon@znc.in>
566
567commit ee9dfc9a7658e7fe75d27483bb5ed1ba4d1e2c86
568Author: Christian Linhart <chris@demorecorder.com>
569Date:   Wed Jan 25 10:21:05 2017 +0100
570
571    add support for eventstruct
572    
573    eventstruct allows to use events as part of requests.
574    This is, e.g., needed by xcb_input_send_extension_event.
575    
576    Signed-off-by: Christian Linhart <chris@demorecorder.com>
577
578commit 0c2c5d50f8670da3e7601feb6a29b53509513da5
579Author: Christian Linhart <chris@demorecorder.com>
580Date:   Fri Jan 20 20:14:57 2017 +0100
581
582    optionally build the GE extension
583    
584    xcb contains an xml-definition for the GenericEvent extension
585    but this extension was neither generated nor built.
586    
587    This patch enables optional building of the GenericEvent extension
588    with configure option --enable-ge
589    
590    By default, the GenericEvent extension is not built.
591    Normally this is not needed by application programs
592    because there is implicit support for the GE-extension
593    for the specific events built with this extension.
594    
595    But it may be useful for X-protocol analyzers and stuff like that.
596    
597    Signed-off-by: Christian Linhart <chris@demorecorder.com>
598
599commit 9bce1f72e329cb407b7a95589b9675a08129b65d
600Author: Christian Linhart <chris@demorecorder.com>
601Date:   Fri Jan 20 14:40:25 2017 +0100
602
603    move symbol lookup of sumof expr to the parser
604    
605    replace the complicated symboltable lookup for sumof expr
606    by accessing the lenfield of the expr-object.
607    
608    This requires the corresponding patch for xcb/proto
609    which sets the lenfield accordingly.
610    
611    This should be OK because for official releases we define
612    that dependency in the build system.
613    
614    For getting versions off the HEAD of the git repo, it should
615    be obvious that xcb/proto and xcb/libxcb have to be updated together.
616    
617    I have tested this patch and it generates exactly the same code
618    as before.
619    
620    Tested-by: Christian Linhart <chris@demorecorder.com>
621    Signed-off-by: Christian Linhart <chris@demorecorder.com>
622
623commit 65b298c7ca317d7e4316aa2b9e0499e13047c65c
624Author: Alan Coopersmith <alan.coopersmith@oracle.com>
625Date:   Sat Feb 6 12:26:21 2016 -0800
626
627    Correct @param "e" to "error" in xcb_poll_for_reply*()
628    
629    Found by clang -Wdocumentation:
630    
631    ./xcbext.h:271:11: warning: parameter 'e' not found in the function
632          declaration [-Wdocumentation]
633     * @param e Location to store errors in, or NULL. Ignored for un...
634              ^
635    ./xcbext.h:271:11: note: did you mean 'error'?
636     * @param e Location to store errors in, or NULL. Ignored for un...
637              ^
638              error
639    
640    ./xcbext.h:283:11: warning: parameter 'e' not found in the function
641          declaration [-Wdocumentation]
642     * @param e Location to store errors in, or NULL. Ignored for un...
643              ^
644    ./xcbext.h:283:11: note: did you mean 'error'?
645     * @param e Location to store errors in, or NULL. Ignored for un...
646              ^
647              error
648    
649    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
650    Signed-off-by: Uli Schlachter <psychon@znc.in>
651
652commit 32a9084546add979115a686f9a167e70b8967149
653Author: Alan Coopersmith <alan.coopersmith@oracle.com>
654Date:   Sat Feb 6 12:26:20 2016 -0800
655
656    Remove : from @param names in manually written headers
657    
658    Makes style match the @param names in autogenerated headers and makes
659    clang -Wdocumentation stop complaining about all of them:
660    
661    ./xcb.h:523:11: warning: parameter 'display:' not found in the function
662          declaration [-Wdocumentation]
663     * @param display: A pointer to the display number.
664              ^~~~~~~~
665    ./xcb.h:523:11: note: did you mean 'display'?
666     * @param display: A pointer to the display number.
667              ^~~~~~~~
668              display
669    
670    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
671    Signed-off-by: Uli Schlachter <psychon@znc.in>
672
673commit 8740a288ca468433141341347aa115b9544891d3
674Author: Thomas Klausner <wiz@NetBSD.org>
675Date:   Thu May 19 17:31:18 2016 +0200
676
677    Fix inconsistent use of tabs vs. space.
678    
679    Needed for at least python-3.5.x.
680    
681    Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
682    Signed-off-by: Uli Schlachter <psychon@znc.in>
683
684commit d34785a34f28fa6a00f8ce00d87e3132ff0f6467
685Author: Uli Schlachter <psychon@znc.in>
686Date:   Sat May 14 10:36:54 2016 +0200
687
688    Release libxcb 1.12
689    
690    Signed-off-by: Uli Schlachter <psychon@znc.in>
691
692commit b11fca06f75b26b94e3f1d3e3c3954d365f80759
693Author: Uli Schlachter <psychon@znc.in>
694Date:   Sat May 14 10:33:54 2016 +0200
695
696    Bump xcb-proto requirement to 1.12
697    
698    This is needed due to various changes that were done to the XML schema.
699    
700    Signed-off-by: Uli Schlachter <psychon@znc.in>
701
702commit 095353ff1a4f611922dfc4c98b0c4bd55d9f6d4f
703Author: Mark Kettenis <kettenis@openbsd.org>
704Date:   Sat Jan 23 17:29:32 2016 +0100
705
706    Increase unix socket send buffer to at least 64KB
707    
708    Some systems (e.g. OpenBSD) have a rather small default socket send buffer
709    size of 4KB.  The result is that sending requests with a largish payload
710    requires serveral writev(2) system calls.  Make sure the socket send buffer
711    is at least 64KB such that we're likely to succeed with a single system
712    call for most requests.  A similar change was made to the xtrans code
713    some time ago.
714    
715    Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
716    Reviewed-by: Matthieu Herrb <matthieu@herrb.eu>
717
718commit b3516102b4469df99db39e0e354deae496867f35
719Author: Christian Linhart <chris@demorecorder.com>
720Date:   Mon Jan 18 06:56:39 2016 +0100
721
722    do not serialize pads by default anymore
723    
724    Pads should not be serialized/deserialized to maintain
725    ABI compatibility when adding explicit align pads.
726    
727    Therefore this pad switches off serialization of pads
728    unless it is enforced by serialize=true in the xml-definition
729    of that pad
730    
731    Signed-off-by: Christian Linhart <chris@demorecorder.com>
732
733commit c03388ff9e253b86e4b41fb77034ca6007fe47e9
734Author: Jaya Tiwari <tiwari.jaya18@gmail.com>
735Date:   Wed Nov 11 01:02:09 2015 +0100
736
737    calculate lengthless list
738    
739    Some rework done by Christian Linhart
740    
741    Signed-off-by: Jaya Tiwari <tiwari.jaya18@gmail.com>
742    Signed-off-by: Christian Linhart <chris@demorecorder.com>
743
744commit 775825756714eb6b8467e099cda73a03b782ea0e
745Author: Christian Linhart <chris@demorecorder.com>
746Date:   Tue Nov 10 12:53:04 2015 +0100
747
748    Fix handling of align-pads in end-iterators
749    
750    If a list is preceded by an align-pad, then
751    accessor for the end-iterator returned a wrong
752    value.
753    
754    Reason: the length of the align-iterator was added
755    to a pointer of list-member type. Therefore, the length
756    was multiplied by the size of the list-member type,
757    due to C pointer arithmetic rules.
758    
759    This has looked like the following, e.g., in
760    xcb_randr_get_crtc_transform_pending_params_end:
761    
762    i.data = ((xcb_render_fixed_t *) prev.data) + ((-prev.index) & (4 - 1)) + (R->pending_nparams);
763    
764    This bug was introduced with the following commit:
765    http://cgit.freedesktop.org/xcb/libxcb/commit/?id=4033d39d4da21842bb1396a419dfc299591c3b1f
766    
767    The fix handles this by casting to char* before adding the align,
768    and then casting the result to the member type.
769    
770    Signed-off-by: Christian Linhart <chris@demorecorder.com>
771
772commit 32a2189183696e942b002efcbca823a416fe5f6a
773Author: Christian Linhart <chris@demorecorder.com>
774Date:   Sun Nov 1 18:35:35 2015 +0100
775
776    set the align-offset as provided by proto
777    
778    instead of using the lower bits of the pointer address.
779    This fixes a bug reported by Peter Hutterer in off-list communication
780    back in June 2015.
781    
782    This requires the alignment-checker patches in xcb/proto.
783    
784    Signed-off-by: Christian Linhart <chris@demorecorder.com>
785
786commit 6e0378ebbf5f842992e37c93ad60b960427849ac
787Author: Adam Jackson <ajax@redhat.com>
788Date:   Mon Sep 21 15:27:52 2015 -0400
789
790    Bump version to 1.11.90
791    
792    We've released 1.11.1 and new libX11 wants that or better.  git master
793    will suffice, so bump the version number ahead of 1.11 branch.
794    
795    Signed-off-by: Adam Jackson <ajax@redhat.com>
796
797commit 4033d39d4da21842bb1396a419dfc299591c3b1f
798Author: Christian Linhart <chris@demorecorder.com>
799Date:   Fri Sep 19 13:44:38 2014 +0200
800
801    make lists after align-pads work
802    
803    Handle align-pads when generating an end-function
804    in the same way as handling them when generating
805    an accessor or iterator function.
806    
807    Signed-off-by: Christian Linhart <chris@demorecorder.com>
808
809commit b15c96f9507119e5d38a61d92b4dbcd479ea2099
810Author: Christian Linhart <chris@demorecorder.com>
811Date:   Thu Jun 11 18:58:38 2015 +0200
812
813    make support for server side stuff optional
814    
815    and make it disabled by default with an EXPERIMENTAL warning
816    
817    reason: this feature is unfinished and we want to have flexibility for
818    ABI/API changes, while still being able to make a release soon
819    
820    Signed-off-by: Christian Linhart <chris@demorecorder.com>
821
822commit c5d923d8ff4a9b9fc1aef1c6c6918bab15098d34
823Author: Jon TURNEY <jon.turney@dronecode.org.uk>
824Date:   Tue Mar 17 17:49:14 2015 +0000
825
826    Link with winsock library for socket functions on MinGW
827    
828    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
829
830commit 5b40681c887192307f3ae147d2158870aa79c05f
831Author: Uli Schlachter <psychon@znc.in>
832Date:   Fri Jun 12 15:13:05 2015 +0200
833
834    Fix a thread hang with xcb_wait_for_special_event()
835    
836    Consider the following:
837    
838    - Two threads are calling xcb_wait_for_special_event() and xcb_wait_for_reply()
839      concurrently.
840    - The thread doing xcb_wait_for_reply() wins the race and poll()s the socket for
841      readability.
842    - The other thread will be put to sleep on the special_event_cond of the special
843      event (this is done in _xcb_conn_wait() via the argument
844      xcb_wait_for_special_event() gives it).
845    - The first thread gets its reply, but does not yet receive any special event.
846    
847    In this case, the first thread will return to its caller. On its way out, it
848    will call _xcb_in_wake_up_next_reader(), but that function cannot wake up
849    anything since so far it did not handle xcb_wait_for_special_event().
850    
851    Thus, the first thread stays blocked on the condition variable and no thread
852    tries to read from the socket.
853    
854    A test case demonstrating this problem is available at the bug report.
855    
856    Fix this similar to how we handle this with xcb_wait_for_reply():
857    
858    The function wait_for_reply() adds an entry into a linked list of threads that
859    wait for a reply. Via this list, _xcb_in_wake_up_next_reader() can wake up this
860    thread so that it can call _xcb_conn_wait() again and then poll()s the socket.
861    
862    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84252
863    Signed-off-by: Uli Schlachter <psychon@znc.in>
864    Tested-by: Michel Dänzer <michel.daenzer@amd.com>
865
866commit f85661c3bca97faa72431df92a3867be39a74e23
867Author: Michel Dänzer <michel.daenzer@amd.com>
868Date:   Mon Jun 1 11:04:18 2015 +0900
869
870    Call _xcb_wake_up_next_reader from xcb_wait_for_special_event
871    
872    All functions calling _xcb_conn_wait() must make sure that waiting
873    readers are woken up when we read a reply or event that they are waiting
874    for. xcb_wait_for_special_event() did not do so. This adds the missing
875    call to_xcb_in_wake_up_next_reader().
876    
877    Fixes deadlock when waiting for a special event and concurrently
878    processing the display connection queue in another thread.
879    
880    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84252
881    Tested-by: Thomas Daede <bztdlinux@gmail.com>
882    Tested-by: Clément Guérin <geecko.dev@free.fr>
883    Reviewed-by: Uli Schlachter <psychon@znc.in>
884    Signed-off-by: Michel Dänzer <michel@daenzer.net>
885    Signed-off-by: Uli Schlachter <psychon@znc.in>
886
887commit 8584c0e09573a29d8ba7050e3d5afd925b4d8d80
888Author: Uli Schlachter <psychon@znc.in>
889Date:   Thu May 14 09:44:05 2015 +0200
890
891    send_fds(): Handle too many outstanding FDs to send
892    
893    Before this patch, the following code caused an endless loop in send_fds(),
894    because the queue of FDs to send was eventually full, but _xcb_out_flush_to()
895    didn't make any progress, since there was no request to send:
896    
897       while (1) { xcb_send_fd(conn, dup(some_fd)); }
898    
899    Fix this by sending a sync when flushing didn't make any progress. That way we
900    actually have something to send and can attach the pending FDs.
901    
902    Because send_fds() can now send requests, the code in
903    xcb_send_request_with_fds64() has to be changed. It has to call send_fds()
904    before it establishes a good sequence number for the request it wants to send.
905    
906    Signed-off-by: Uli Schlachter <psychon@znc.in>
907
908commit 658fb4a5f0050db68fdf092936afe596412ef5f7
909Author: Uli Schlachter <psychon@znc.in>
910Date:   Wed Apr 22 09:26:05 2015 +0200
911
912    Code generator: Use xcb_send_request_with_fds()
913    
914    Signed-off-by: Uli Schlachter <psychon@znc.in>
915
916commit b15aa6bd4efde784e546d168bb23b8a8e816e85b
917Author: Uli Schlachter <psychon@znc.in>
918Date:   Wed Apr 22 09:23:47 2015 +0200
919
920    Add xcb_send_request_with_fds() and *_with_fds64()
921    
922    Doing xcb_send_fd(), xcb_send_request() is racy. If two threads do this at the
923    same time, they could mix up their file descriptors. This commit makes it
924    possibly to fix this race by providing a single function which does everything
925    that is needed.
926    
927    Signed-off-by: Uli Schlachter <psychon@znc.in>
928
929commit cc04cfb41bece6ec239f57d83822286b507f4482
930Author: Uli Schlachter <psychon@znc.in>
931Date:   Mon May 18 21:40:34 2015 +0200
932
933    send_fds(): Make sure no other thread interrupts us
934    
935    Two threads trying to send fds at the same time could interfere. To guarantee a
936    correct ordering, we have to use correct locking. The code in send_fds() missed
937    one case: If there was another thread already writing requests, we slept on the
938    "done with writing" condition variable (c->out.cond). This would allow other
939    threads to re-acquire the iolock before us and could cause fds to be sent out of
940    order.
941    
942    To fix this, at the beginning of send_fds() we now make sure that no other
943    thread is already writing requests. This is what prepare_socket_request() does.
944    Additionally, it gets the socket back in case xcb_take_socket() was called,
945    which is a good thing, too, since fds are only sent with corresponding requests.
946
947commit 25f9e7e45a7652b35b71c7941beef774a39f0d86
948Author: Uli Schlachter <psychon@znc.in>
949Date:   Wed Apr 22 09:20:38 2015 +0200
950
951    xcb_send_fd(): Always close fds
952    
953    The API docs for xcb_send_fd() says "After this function returns, the file
954    descriptor given is owned by xcb and will be closed eventually".
955    
956    Let the implementation live up to its documentation. We now also close fds if fd
957    passing is unavailable (!HAVE_SENDMSG) and when the connection is in an error
958    state.
959    
960    (This also does sneak in some preparatory functions for follow-up commits and
961    thus does things in a more complicated way than really necessary.)
962    
963    Signed-off-by: Uli Schlachter <psychon@znc.in>
964
965commit bbdf1d133f7bd979c6ff3bf44ec3d0c2d2b9dbfe
966Author: Ran Benita <ran234@gmail.com>
967Date:   Wed Mar 18 12:27:32 2015 +0200
968
969    c_client.py: don't generate useless empty /** < */ comments
970    
971    (This does not change doxygen's output or warnings).
972    
973    Signed-off-by: Ran Benita <ran234@gmail.com>
974    Reviewed-by: Christian Linhart <chris@demorecorder.com>
975
976commit ff6cb3913b64a4aa29cb5e65168ea49d77195296
977Author: Ran Benita <ran234@gmail.com>
978Date:   Wed Mar 18 12:27:31 2015 +0200
979
980    c_client.py: use pattern matching with enumerate()
981    
982    Signed-off-by: Ran Benita <ran234@gmail.com>
983
984commit cb621341a62e6d2233db3e337611f6fdd4f675a6
985Author: Christian Linhart <chris@demorecorder.com>
986Date:   Wed Apr 29 09:11:37 2015 +0200
987
988    expose 64-bit sequence numbers for XLib
989    
990    While XCB uses 64-bit sequence number internally, it only exposes
991    "unsigned int" so that, on 32-bit architecture, Xlib based applications
992    may see their sequence number wrap which causes the connection to the X
993    server to be lost.
994    
995    Expose 64-bit sequence number from XCB API so that Xlib and others can
996    use it even on 32-bit environment.
997    
998    This implies the following API addition:
999    
1000      xcb_send_request64()
1001      xcb_discard_reply64()
1002      xcb_wait_for_reply64()
1003      xcb_poll_for_reply64()
1004    
1005    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71338
1006    
1007    Reviewed-by: Uli Schlachter <psychon@znc.in>
1008    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1009    Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
1010
1011commit c49aa985941112be05599032b9bb45b2652301ce
1012Author: Alan Coopersmith <alan.coopersmith@oracle.com>
1013Date:   Wed Apr 29 23:23:05 2015 -0700
1014
1015    Escape \n to display properly in xcb-requests man page
1016    
1017    In nroff, \n is a macro that "Interpolates number register x" (where x
1018    is the character following the \n sequence), thus the man page currently
1019    prints 0 instead of \n" in several lines, leading to output such as:
1020    
1021     printf("The _NET_WM_NAME atom has ID %u0, reply-⁠>atom);
1022    
1023    It needs to be escaped here, as \\n, as is done in other examples in
1024    this man page already.
1025    
1026    https://bugs.freedesktop.org/show_bug.cgi?id=90231
1027    
1028    Reported-by: Stefan Merettig
1029    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
1030
1031commit a90be9955d2c5a635f791d44db1154633b9d3322
1032Author: Ran Benita <ran234@gmail.com>
1033Date:   Sun Oct 12 21:58:20 2014 +0300
1034
1035    c_client.py: make condition easier to follow in _c_complex()
1036    
1037    Signed-off-by: Ran Benita <ran234@gmail.com>
1038    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1039
1040commit f9f925107e37e0c9a0ed8220ee3a23f584e3b2ec
1041Author: Ran Benita <ran234@gmail.com>
1042Date:   Sun Oct 12 21:58:14 2014 +0300
1043
1044    c_client.py: don't add /* <name> */ before references to 'S'
1045    
1046    The name can be understood from the type of S already.
1047    
1048    For examples, look for 'S->' in xkb.c or xinput.c.
1049    
1050    Signed-off-by: Ran Benita <ran234@gmail.com>
1051    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1052    Reviewed-by: Rémi Cardona <remi@gentoo.org>
1053
1054commit 17f9bda6c291a8b9ccc98b22c241b64880b80621
1055Author: Ran Benita <ran234@gmail.com>
1056Date:   Sun Oct 12 21:58:19 2014 +0300
1057
1058    c_client.py: remove duplicated `cookie_type` argument for requests
1059    
1060    It is implied already inside the function by the `void` argument.
1061    
1062    Signed-off-by: Ran Benita <ran234@gmail.com>
1063    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1064
1065commit c65005e9d0ce60524d2e883c13c027a5f1f7914c
1066Author: Ran Benita <ran234@gmail.com>
1067Date:   Sun Oct 12 21:58:18 2014 +0300
1068
1069    c_client.py: spell out keyword arguments in c_request() for clarity
1070    
1071    Signed-off-by: Ran Benita <ran234@gmail.com>
1072    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1073
1074commit 6872e925828e615e2c1d4a0bbcc6be6a32fa4e7a
1075Author: Ran Benita <ran234@gmail.com>
1076Date:   Sun Oct 12 21:58:17 2014 +0300
1077
1078    c_client.py: simplify _c_reply_has_fds()
1079    
1080    Signed-off-by: Ran Benita <ran234@gmail.com>
1081    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1082
1083commit 8bf8b62316a3066b61243d797d22c2100fd173f5
1084Author: Ran Benita <ran234@gmail.com>
1085Date:   Sun Oct 12 21:58:16 2014 +0300
1086
1087    c_client.py: remove commented debug statements
1088    
1089    Signed-off-by: Ran Benita <ran234@gmail.com>
1090    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1091
1092commit ec435aebd6ed0523de2d341e6aff5ae66f230b10
1093Author: Ran Benita <ran234@gmail.com>
1094Date:   Sun Oct 12 21:58:15 2014 +0300
1095
1096    c_client.py: use C99 initializers instead of comments
1097    
1098    Signed-off-by: Ran Benita <ran234@gmail.com>
1099    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1100
1101commit 89498d1d450d0e6e476f6b24908ecae857863386
1102Author: Ran Benita <ran234@gmail.com>
1103Date:   Sun Oct 12 21:58:11 2014 +0300
1104
1105    c_client.py: remove end-of-function comments
1106    
1107    Signed-off-by: Ran Benita <ran234@gmail.com>
1108    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1109
1110commit 2871d4b1b8736044ba50df2f7dacfcc65d820fa9
1111Author: Ran Benita <ran234@gmail.com>
1112Date:   Sun Oct 12 21:58:10 2014 +0300
1113
1114    c_client.py: no need to compare bools to True/False
1115    
1116    Signed-off-by: Ran Benita <ran234@gmail.com>
1117    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1118
1119commit 30976e5255a25117d17d4d8162e29b84913953cd
1120Author: Ran Benita <ran234@gmail.com>
1121Date:   Sun Oct 12 21:58:09 2014 +0300
1122
1123    c_client.py: use "foo".join() instead of reduce
1124    
1125    Signed-off-by: Ran Benita <ran234@gmail.com>
1126    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1127
1128commit 0ab52cbcc630ff8ddfa6f6b245c6b3071867b291
1129Author: Ran Benita <ran234@gmail.com>
1130Date:   Sun Oct 12 21:58:08 2014 +0300
1131
1132    c_client.py: fix indentation
1133    
1134    (Also remove unnecessary parens around the condition).
1135    
1136    Signed-off-by: Ran Benita <ran234@gmail.com>
1137    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1138
1139commit 80341d5df3e30e2cfc3066d05256af9e513e4500
1140Author: Ran Benita <ran234@gmail.com>
1141Date:   Sun Oct 12 21:58:07 2014 +0300
1142
1143    c_client.py: use comprehensions instead of map/filter
1144    
1145    Signed-off-by: Ran Benita <ran234@gmail.com>
1146    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1147
1148commit 86ea6645d9e388112670cf88da1fe3b350953789
1149Author: Ran Benita <ran234@gmail.com>
1150Date:   Sun Oct 12 21:58:06 2014 +0300
1151
1152    c_client.py: use print as a function for python3 compatibility
1153    
1154    This works for all python>=2.6, which is what configure requires.
1155    
1156    Signed-off-by: Ran Benita <ran234@gmail.com>
1157    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1158
1159commit 70d32ce7d8ef3ce9490d9a51b189829c8bcdb5c4
1160Author: Ran Benita <ran234@gmail.com>
1161Date:   Sun Oct 12 21:58:05 2014 +0300
1162
1163    c_client.py: fix pyflakes errors
1164    
1165    c_client.py:2: 'from xml.etree.cElementTree import *' used; unable to detect undefined names
1166    c_client.py:3: 'basename' imported but unused
1167    c_client.py:9: 'time' imported but unused
1168    c_client.py:1437: local variable 'list_obj' is assigned to but never used
1169    c_client.py:1745: local variable 'varfield' is assigned to but never used
1170    c_client.py:2050: local variable 'length' is assigned to but never used
1171    c_client.py:2416: local variable 'R_obj' is assigned to but never used
1172    c_client.py:2441: local variable 'S_obj' is assigned to but never used
1173    
1174    Signed-off-by: Ran Benita <ran234@gmail.com>
1175    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1176
1177commit 12d23b934f189607ad77667834205dfcf89a41f0
1178Author: Ran Benita <ran234@gmail.com>
1179Date:   Sun Oct 12 21:58:04 2014 +0300
1180
1181    c_client.py: simplify maximum expression
1182    
1183    Signed-off-by: Ran Benita <ran234@gmail.com>
1184    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1185
1186commit 1b37d6ad3a37a9d0201e3b2caee6cd53b7b4b3dc
1187Author: Ran Benita <ran234@gmail.com>
1188Date:   Sun Oct 12 21:58:03 2014 +0300
1189
1190    c_client.py: remove unneeded call to get_serialize_params()
1191    
1192    The results are not used, and the function doesn't have side effects.
1193    
1194    Signed-off-by: Ran Benita <ran234@gmail.com>
1195    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1196
1197commit e3ec1f74637237ce500dfd0ca59f2e422da4e019
1198Author: Jaya Tiwari <tiwari.jaya18@gmail.com>
1199Date:   Thu Jan 22 12:16:33 2015 -0500
1200
1201    Adding accessors for requests
1202    
1203    Added accessor functions for requests the same way they were added for
1204    structs,events and replies.
1205    Lists for replies have accessor functions now.
1206    
1207    Signed-off-by: Jaya Tiwari <tiwari.jaya18@gmail.com>
1208    Reviewed-by: Christian Linhart <chris@demorecorder.com>
1209    
1210    Comment from the Reviewer Christian Linhart:
1211    I have tested your patch after fixing the issues with the patch-format.
1212    It looks good:
1213    * only adds new functions, and does not modify existing functions.
1214      Therefore it is API and ABI compatible.
1215    
1216    * adds accessors for varsized-stuff in requests.
1217      This is needed for server-side XCB and may be useful for implementing X11-protocol proxies.
1218
1219commit 5353c0216e091b64d01a43e6580e6d69b2ac16c7
1220Merge: bbca7b8 c6f3fb2
1221Author: Christian Linhart <chris@demorecorder.com>
1222Date:   Tue Feb 10 10:13:04 2015 +0100
1223
1224    Merge http://git.demorecorder.com/git/free-sw/xcb/libxcb
1225    branch 'ParametrizedStruct-V7'
1226
1227commit c6f3fb2529a6211221e8254f58c85fd67c1d8844
1228Author: Christian Linhart <chris@DemoRecorder.com>
1229Date:   Mon Nov 3 09:57:59 2014 +0100
1230
1231    generator: support parametrized structs
1232    
1233    Parametrized structs contain paramref expressions which
1234    refer to the value of a field defined in the context
1235    where the struct is used.
1236    
1237    Implementing the parametrized structs turned out
1238    to be somewhat easier than previously thought
1239    because the generator already had some support for type-parametrization
1240    because this is needed when case or bitcase refers to fields outside
1241    of the switch.
1242    
1243    So I decided to go with the flow and to implement the solution
1244    which best fits the current implementation.
1245    
1246    I did the following:
1247    * I provided a way to specify fieldref with an explicitely given type:
1248      This resulted in <paramref type="CARD8>fieldname</paramref>
1249      A paramref is just a fieldref with an explicit type.
1250      The type is necessary because there is no local field of that
1251      name where the type can be derived from.
1252    
1253    * then I tested it and made several changes in the generator
1254      such that it really works.
1255    
1256    Basically the generated code is as follows:
1257    * The parameter appears on the parameter list of the
1258      sizeof-function of the parametrized struct.
1259      When that function gets called, an appropriate argument is supplied.
1260    
1261    * The parameter also appears as an additional member of the iterator-struct
1262      for the iterator of lists of that parametrized struct.
1263      This way, the next-function can get the value of that parameter from the iterator.
1264      When the iterator is created, this iterator-member is set accordingly.
1265    
1266    * When the paramref appears in the length-expression of a list, then
1267      the parameter appears on the parameterlist of the "length" and "end" functions.
1268      When these functions get called, an appropriate argument is supplied.
1269    
1270    Some comments:
1271    * I did not implement inline structs.
1272      This would probably have been more complicated, and at least some additional effort.
1273      But that can be implemented later if needed.
1274      (Inline structs could probably use some code from switch-case/bitcase which is already kind of
1275      an inlined struct but one has to be careful not to break the functionality
1276      of switch-case/bitcase. Support for inline structs inside lists must probably
1277      be implemented from scratch...)
1278    
1279    * The paramref expression refers to a field of the same name in the struct/request/...
1280      where it is used.
1281      So it is not possible to pass the value of arbitrary fields or even expressions
1282      to the parametrized struct.
1283      This would have been possible with the previously discussed <typearg>.
1284      That can be added later, if needed.
1285      ( Wont be too complicated )
1286    
1287    * So this is pretty much like the proposal from Ran Benita.
1288    
1289    changes for V2 of this patch, according to suggestions from Ran Benita:
1290    * replace map with list comprehension
1291      because map returns an iterator instead of a list from Python 3 on,
1292      so it cannot be added to a list anymore.
1293    
1294    * removed "self" parameter of function additional_params_to_str
1295      and accessed the variable additional_params from the outer
1296      function directly.
1297    
1298    changes for V2 of this patch:
1299    * adapt to revision 2 of patchset ListInputDevices
1300    * style fixes for similar things that Ran Benita has found in my previous patches
1301    
1302    Message-ID: <54574397.4060000@DemoRecorder.com>
1303    Patch-Thread-Subject: [Xcb] parametrized structs implemented
1304    Patch-Set: ParametrizedStruct
1305    Patch-Number: libxcb 1/1
1306    Patch-Version: V3
1307    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1308
1309commit 912cd97a6dd019e9e7ecf09c82c4577dd2ad7529
1310Author: Christian Linhart <chris@DemoRecorder.com>
1311Date:   Mon Nov 3 09:58:08 2014 +0100
1312
1313    generator: support listelement-ref
1314    
1315    Support for listelement-ref needs the following three changes
1316    (in the order as they appear in the patch):
1317    
1318    * making the current list-element accessible with the variable
1319      xcb_listelement which is a pointer to the list-element
1320    
1321    * supporting lists of simple-type for sumof with a nested expression
1322    
1323    * using the variable for resolving a listelement-ref expression
1324    
1325    Changes for V2 of this patch:
1326    - adapt to removal of patch "libxcb 2/6" from patchset "ListInputDevices".
1327    
1328    Changes for V3 of this patch:
1329    - adapt to V2 of patch "libxcb 5/6" from patchset "ListInputDevices"
1330    
1331    Changes for V4 of this patch:
1332    - adapt to revision 2 of the patchset "ListInputDevices"
1333    
1334    Message-ID: <545743A0.50907@DemoRecorder.com>
1335    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1336    Patch-Set: PopcountList
1337    Patch-Number: libxcb 4/4
1338    Patch-Version: V4
1339    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1340
1341commit 422458b66380e4103c4937f0e2e8bb93e31f273a
1342Author: Christian Linhart <chris@demorecorder.com>
1343Date:   Thu Sep 4 17:50:50 2014 +0200
1344
1345    generator: _c_accessor_get_length: remove buggy special case
1346    
1347    The function _c_accessor_get_length had a special case handling
1348    for intermixed var and fixed size fields.
1349    
1350    However:
1351    * The implementation of that special case was buggy:
1352      It tried to call a python-dict as a function which causes
1353      Python to abort the program with a stacktrace and error message.
1354      So this code was never used.
1355    
1356    * The case it tried to handle is handeled elsewhere in the
1357      meantime: in _c_helper_absolute_name by previous patches
1358      made by me.
1359    
1360    Message-ID: <1409845851-38950-3-git-send-email-chris@demorecorder.com>
1361    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1362    Patch-Set: PopcountList
1363    Patch-Number: libxcb 3/4
1364    Patch-Version: V1
1365    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1366
1367commit b1e4a3bbd8194d12d7fcd9705fcbbe0deb59bcba
1368Author: Christian Linhart <chris@demorecorder.com>
1369Date:   Thu Sep 4 17:50:49 2014 +0200
1370
1371    generator: generate accessors for events, too
1372    
1373    Accessors are generally needed for var-sized fields
1374    and fields after var-sized fields.
1375    
1376    Generic events can have ver-sized fields.
1377    Therefore they need accessors.
1378    
1379    Message-ID: <1409845851-38950-2-git-send-email-chris@demorecorder.com>
1380    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1381    Patch-Set: PopcountList
1382    Patch-Number: libxcb 2/4
1383    Patch-Version: V1
1384    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1385
1386commit 6234225b4be862c5882bf547d066c5a3885906dd
1387Author: Christian Linhart <chris@demorecorder.com>
1388Date:   Thu Sep 4 17:50:48 2014 +0200
1389
1390    generator: no type-setup for eventcopies anymore
1391    
1392    _c_type_setup is not called for eventcopies anymore:
1393    Reasons:
1394    * the type-setup of an eventcopy would overwrite members of the original
1395      event object such as c_type, ...
1396    * it is needed for the next patch, i.e., generating accessors:
1397      type_setup would create sizeof-etc funtions which called
1398      undefined accessor functions.
1399    
1400    Sizeof-functions are generated for compatibility:
1401    Reason:
1402    * Type-setup of eventcopies has previously generated
1403      sizeof-functions for eventcopies.
1404      So, we still need to generate these functions.
1405      These new sizeof-functions simply call the sizeof-function
1406      of the defining event of the eventcopy.
1407    
1408    Message-ID: <1409845851-38950-1-git-send-email-chris@demorecorder.com>
1409    Patch-Thread-Subject: [Xcb] support popcount of a list and associated xml changes
1410    Patch-Set: PopcountList
1411    Patch-Number: libxcb 1/4
1412    Patch-Version: V1
1413    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1414
1415commit 18ff453edd42712ea4d1e7218bbe8829f9f4caba
1416Author: Christian Linhart <chris@DemoRecorder.com>
1417Date:   Sun Nov 2 13:46:58 2014 +0100
1418
1419    _c_helper_fieldaccess_expr: remove handling for empty sep
1420    
1421    The loop-variable "sep" is never empty in function
1422    "_c_helper_fieldaccess_expr", after a fix elsewhere.
1423    Therefore I removed the handling of the case of "sep" being empty.
1424    
1425    Thanks to Ran Benita for the hint that this can be removed.
1426    
1427    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1428    Reviewed-by: Ran Benita <ran234@gmail.com>
1429    
1430    Message-ID: <545627C2.3050608@DemoRecorder.com>
1431    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1432    Patch-Set: ListInputDevices
1433    Patch-Number: libxcb 9/9
1434    Patch-Version: V1
1435
1436commit d905b886185fd8e0d9d7f70e0d112cb58ab86f98
1437Author: Christian Linhart <chris@DemoRecorder.com>
1438Date:   Sun Nov 2 13:46:50 2014 +0100
1439
1440    function _c_helper_fieldaccess_expr: improve description
1441    
1442    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1443    Reviewed-by: Ran Benita <ran234@gmail.com>
1444    
1445    Message-ID: <545627BA.1000909@DemoRecorder.com>
1446    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1447    Patch-Set: ListInputDevices
1448    Patch-Number: libxcb 8/9
1449    Patch-Version: V1
1450
1451commit 17f6e04493c93014beaf3a704c8cc4d050cb13bf
1452Author: Christian Linhart <chris@DemoRecorder.com>
1453Date:   Sun Nov 2 13:46:38 2014 +0100
1454
1455    rename _c_helper_absolute_name to _c_helper_fieldaccess_expr
1456    
1457    The function _c_helper_absolute_name was named in
1458    a misleading way.
1459    It computes a C-expression for accessing a field of an xcb-type.
1460    
1461    Therefore the name _c_helper_fieldaccess_expr is more appropriate.
1462    
1463    Note: Patch 6 of this series has been removed during the review process.
1464    
1465    Signed-off-by: Christian Linhart <chris@demorecorder.com>
1466    Reviewed-by: Ran Benita <ran234@gmail.com>
1467    
1468    Message-ID: <545627AE.2040200@DemoRecorder.com>
1469    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1470    Patch-Set: ListInputDevices
1471    Patch-Number: libxcb 7/9
1472    Patch-Version: V1
1473
1474commit 51a0d57acc6ec0c9487d2dbc2dda806f05c49884
1475Author: Christian Linhart <chris@DemoRecorder.com>
1476Date:   Sun Nov 2 13:46:16 2014 +0100
1477
1478    generator: sumof with nested expression
1479    
1480    Support sumof with a nested expression.
1481    The nested expression is computed for every list-element
1482    and the result of the computation is added to the sum.
1483    
1484    This way, sumof can be applied to a list of structs,
1485    and, e.g., compute the sum of a specific field of that struct.
1486    
1487    example:
1488    <struct name="SumofTest_Element">
1489       <field type="CARD16" name="foo" />
1490       <field type="CARD16" name="bar" />
1491    </struct>
1492    
1493    <struct name="SumofTest_FieldAccess">
1494       <field type="CARD32" name="len" />
1495       <list type="SumofTest_Element" name="mylist1">
1496               <fieldref>len</fieldref>
1497       </list>
1498       <list type="CARD16" name="mylist2">
1499            <sumof ref="mylist1">
1500                    <fieldref>bar</fieldref>
1501            </sumof>
1502       </list>
1503    </struct>
1504    
1505    generated tmpvar:
1506        int xcb_pre_tmp_1; /* sumof length */
1507        int xcb_pre_tmp_2; /* sumof loop counter */
1508        int64_t xcb_pre_tmp_3; /* sumof sum */
1509        const xcb_input_sumof_test_element_t* xcb_pre_tmp_4; /* sumof list ptr */
1510    
1511    generated code:
1512        /* mylist2 */
1513        /* sumof start */
1514        xcb_pre_tmp_1 = _aux->len;
1515        xcb_pre_tmp_3 = 0;
1516        xcb_pre_tmp_4 = xcb_input_sumof_test_field_access_mylist_1(_aux);
1517        for ( xcb_pre_tmp_2 = 0; xcb_pre_tmp_2 < xcb_pre_tmp_1; xcb_pre_tmp_2++) {
1518            xcb_pre_tmp_3 += xcb_pre_tmp_4->bar;
1519            xcb_pre_tmp_4++;
1520        }
1521        /* sumof end. Result is in xcb_pre_tmp_3 */
1522        xcb_block_len += xcb_pre_tmp_3 * sizeof(uint16_t);
1523    
1524    changes for V2 of this patch:
1525    * explicitely set the member access operator in the prefix-tuple
1526      passed to function _c_helper_field_mapping.
1527      This enables us to simplify function "_c_helper_absolute_name"
1528      (which will be renamed "_c_helper_fieldaccess_expr" soon)
1529    
1530    V3: Changed style and formatting according to suggestions from Ran Benita
1531    
1532    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1533    Reviewed-by: Ran Benita <ran234@gmail.com>
1534    
1535    Message-ID: <54562798.8040500@DemoRecorder.com>
1536    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1537    Patch-Set: ListInputDevices
1538    Patch-Number: libxcb 5/9
1539    Patch-Version: V3
1540
1541commit 4a915c0dbadf326ea61349e29a0029d29f4bd339
1542Author: Christian Linhart <chris@DemoRecorder.com>
1543Date:   Sun Nov 2 13:45:40 2014 +0100
1544
1545    generator: sumof: support any type, generate explicit code
1546    
1547    A sumof-expression now generates explicit code ( for-loop etc )
1548    instead of calling xcb_sumof.
1549    
1550    This way, it supports any type which can be added.
1551    Previously, only uint_8 was supported.
1552    
1553    Here's an example and the generated code:
1554    
1555    xml:
1556    <struct name="SumofTest">
1557       <field type="CARD32" name="len" />
1558       <list type="CARD16" name="mylist1">
1559               <fieldref>len</fieldref>
1560       </list>
1561       <list type="CARD8" name="mylist2">
1562               <sumof ref="mylist1"/>
1563       </list>
1564    </struct>
1565    
1566    declaration of tempvars at the start of enclosing function:
1567        int xcb_pre_tmp_1; /* sumof length */
1568        int xcb_pre_tmp_2; /* sumof loop counter */
1569        int64_t xcb_pre_tmp_3; /* sumof sum */
1570        const uint16_t* xcb_pre_tmp_4; /* sumof list ptr */
1571    
1572    code:
1573        /* mylist2 */
1574        /* sumof start */
1575        xcb_pre_tmp_1 = _aux->len;
1576        xcb_pre_tmp_3 = 0;
1577        xcb_pre_tmp_4 = xcb_input_sumof_test_mylist_1(_aux);
1578        for ( xcb_pre_tmp_2 = 0; xcb_pre_tmp_2 < xcb_pre_tmp_1; xcb_pre_tmp_2++) {
1579            xcb_pre_tmp_3 += *xcb_pre_tmp_4;
1580            xcb_pre_tmp_4++;
1581        }
1582        /* sumof end. Result is in xcb_pre_tmp_3 */
1583        xcb_block_len += xcb_pre_tmp_3 * sizeof(uint8_t);
1584    
1585    This patch is also a preparation for sumof which can access
1586    fields of lists of struct, etc.
1587    
1588    V2: Changed style and formatting according to suggestions from Ran Benita
1589    
1590    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1591    Reviewed-by: Ran Benita <ran234@gmail.com>
1592    
1593    Message-ID: <54562774.8030306@DemoRecorder.com>
1594    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1595    Patch-Set: ListInputDevices
1596    Patch-Number: libxcb 4/9
1597    Patch-Version: V2
1598
1599commit fda1fb4ed47a705744677a0074d83464af7aa4eb
1600Author: Christian Linhart <chris@DemoRecorder.com>
1601Date:   Sun Nov 2 13:45:29 2014 +0100
1602
1603    generator: expressions can generate pre-code
1604    
1605    This patch provides a mechanism for generating
1606    preparatory code for expressions.
1607    
1608    This is e.g. necessary when an expression needs computations
1609    which cannot be done in a C-Expression, like for-loops.
1610    
1611    This will be used for sumof expressions but may be useful
1612    elsewhere.
1613    
1614    Note: Patch 2 of this series has been removed during the review process.
1615    
1616    V2: adapt to changes in previous patches
1617    
1618    V3: some style and formatting changes according to suggestions from Ran Benita.
1619    
1620    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1621    Reviewed-by: Ran Benita <ran234@gmail.com>
1622    
1623    Message-ID: <54562769.3090405@DemoRecorder.com>
1624    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1625    Patch-Set: ListInputDevices
1626    Patch-Number: libxcb 3/9
1627    Patch-Version: V3
1628
1629commit 265d38882cffce597367cc8bb2160b9e2482a80f
1630Author: Christian Linhart <chris@DemoRecorder.com>
1631Date:   Sun Nov 2 13:45:12 2014 +0100
1632
1633    generator: fix absname for fields with only accessor function
1634    
1635    Fix _c_helper_absolute_name for fields which cannot be accessed
1636    as a struct/union member but which can be accessed by an
1637    accessor function.
1638    
1639    The fix generates calls to the accessor function in these cases.
1640    
1641    Example:
1642    <struct name="AbsnameTest">
1643    <field type="CARD32" name="len" />
1644    <list type="CARD8" name="mylist1">
1645       <fieldref>len</fieldref>
1646    </list>
1647    <list type="CARD8" name="mylist2">
1648       <sumof ref="mylist1"/>
1649    </list>
1650    </struct>
1651    
1652    The sumof-expression ( <sumof ref="mylist1"/> ) refers to mylist1
1653    which is only acessible by an accessor function.
1654    
1655    Previously, sumof was only used inside bitcases,
1656    where such lists are accessible by members of the
1657    deserialized parent struct.
1658    (there is a difference between deserialization of switches
1659    and structs.)
1660    
1661    V2 of this patch:
1662    * replaced "!= None" with "is not None" because that's more pythonic.
1663    (according to suggestion from Ran Benita)
1664    
1665    V3 of this patch: simplification:
1666    * fixed the recursion in _c_type_setup
1667      so that _c_helper_absolute_name does not need check
1668      a gazillion things as a workaround anymore.
1669    
1670    * simplified _c_helper_absolute_name
1671      - remove unneeded check for empty string before
1672        append of last_sep to prefix_str
1673    
1674      - removed those if-conditions which are not
1675        needed anymore after fixing the recursion
1676        in _c_type_setup.
1677    
1678      - extract functionality for checking whether a field
1679        needs an accessor ( and which type of accessor )
1680        in functions.
1681        (also extracted from _c_accessors)
1682    
1683      - rearrange the condition branches and actions for
1684        more readability.
1685    
1686    V3 generates exactly the same *.c and *.h files as V2.
1687    
1688    V4 of this patch:
1689    * improve formatting as per suggestions of Ran
1690    
1691    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1692    Reviewed-by: Ran Benita <ran234@gmail.com>
1693    
1694    Message-ID: <54562758.5090107@DemoRecorder.com>
1695    Patch-Thread-Subject: [Xcb] [PATCHSET] ListInputDevices revision 2
1696    Patch-Set: ListInputDevices
1697    Patch-Number: libxcb 1/9
1698    Patch-Version: V4
1699
1700commit bbca7b82f803fa13fd30a2891ec06f2a213a28c2
1701Merge: 382d306 fdb291b
1702Author: Peter Harris <pharris@opentext.com>
1703Date:   Thu Oct 30 11:51:57 2014 -0400
1704
1705    Merge branch 'NestedStructTypenames-V5' of http://infra-srv1.demorecorder.com/git/free-sw/xcb/libxcb
1706
1707commit fdb291b414a7afc2c2326124b8ca11d6846cd7b9
1708Author: Christian Linhart <chris@demorecorder.com>
1709Date:   Wed Sep 3 10:10:49 2014 +0200
1710
1711    no typename for nested structs
1712    
1713    Nested structs which are generated for named case and bitcase
1714    do not get a typename anymore, i.e., they are anonymous structs.
1715    
1716    Reasons for this change:
1717    * Prior typenames have caused nameclashes
1718    * Prior typenames introduced names in the global namespace which
1719      did not start with the xcb prefix.
1720    
1721    This change is safe with respect to API compatibility because:
1722    I have searched for instances of named bitcases and there's only one place
1723    where they are used, and that's in xkb.xml: reply GetKbdByName.
1724    ( no need to search for <case> because it was introduced after the last release )
1725    
1726    The reply GetKbdByName is broken in its current form in the xkb.xml anyways,
1727    so it is most probably not used anywhere.
1728    
1729    So, my conclusion is that we can safely omit named types for nested structs.
1730    No need for an attribute.
1731    
1732    Message-ID: <1409731849-51897-1-git-send-email-chris@demorecorder.com>
1733    Patch-Thread-Subject: Re: [Xcb] names of nested structs of named bitcase/case are prone to nameclashes. Solution?
1734    Patch-Set: NestedStructTypenames
1735    Patch-Number: libxcb 1/1
1736    Patch-Version: V1
1737    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1738    Reviewed-By: Ran Benita <ran234@gmail.com>
1739
1740commit a7c75be5b1e2da32f7ce2c255972178e7d4b7598
1741Author: Christian Linhart <chris@DemoRecorder.com>
1742Date:   Tue Sep 9 23:26:40 2014 +0200
1743
1744    generator: fix align-pads for switches which start at unaligned pos
1745    
1746    Fix the alignment computation inside switches which start at
1747    an unaligned pos.
1748    This affects both explicit and implicit align pads.
1749    
1750    The alignment offset is derived from the lowest 3 bits of
1751    the pointer to the protocol-data at the start of the switch.
1752    This is sufficient for correcting all alignments up to 8-byte alignment.
1753    As far as I know there is no bigger alignment than 8-byte for the
1754    X-protocol.
1755    
1756    Example:
1757    struct InputState, where the switch starts after two 1-byte fields,
1758    which is a 2 byte offset for 4-byte and 8-byte alignment.
1759    
1760    The previous problem can be demonstrated when adding a
1761    <pad align="4"/> at the end of case "key".
1762    
1763    (Or when finding a testcase which reports the case "valuator" not
1764    at the last position of the QueryDeviceState-reply.
1765    I didn't find such a testcase, so I have used the pad align
1766    as described above.)
1767    
1768    V2: patch modified in order to fix bugs which I found when working on the
1769    next issue:
1770    * xcb_padding_offset has to be set 0 when xcb_block_len is set 0
1771    * xcb_padding_offset cannot be "const" therefore
1772    * for unpack and unserialize, the padding_offset must computed
1773      from _buffer instead of from the aux_var.
1774    
1775    V3: patch revised according to suggestion by Ran Benita:
1776    * only create and use xcb_padding_offset for switch
1777    
1778    Message-ID: <1410298000-24734-1-git-send-email-chris@demorecorder.com>
1779    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
1780    Patch-Set: QueryDeviceState
1781    Patch-Number: libxcb 4/4
1782    Patch-Version: V3
1783    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1784    Reviewed-By: Ran Benita <ran234@gmail.com>
1785
1786commit 277ea629def6728c9d826ff88e95b31c3e25915f
1787Author: Christian Linhart <chris@demorecorder.com>
1788Date:   Thu Aug 21 22:35:55 2014 +0200
1789
1790    generator: support lists of structs which contain a switch
1791    
1792    This essentially requires to have a correct sizeof-function
1793    for the struct.
1794    This in turn requires a sizeof-function for the switch, too.
1795    
1796    Making a sizeof-function for the switch is triggered by
1797    replacing "elif" by "if" in the first change of this patch.
1798    This way, c_need_sizeof is also set to True for switches if appropriate.
1799    
1800    The _c_serialize_helper_switch_field function has to support
1801    the context "sizeof":
1802    This is done in the second change of this patch
1803    
1804    The third change of this patch fixes an alignment error:
1805    It does not make sense to base the padding on the struct-type
1806    which is generated for switch because this struct does not
1807    represent the protocol. Rather it is the output of deserialization.
1808    ( The implicit padding for var-sized fields has other issues, IMHO,
1809    but I am not touching these now...)
1810    
1811    The effect on the generated code for the current xml-files
1812    is as follows:
1813    * several additional sizeof-functions are generated
1814    * the fix of the alignment error only changes one place
1815      in the XKB-extension for the GetKbdByName-reply.
1816      This is no problem because that reply in its current form
1817      is broken/unfinished anyways.
1818    
1819    Note:
1820    This patch also fixes a problem in the generator when
1821    a fixed-size list is the last field of a case or bitcase.
1822    
1823    Message-ID: <1408653356-21191-2-git-send-email-chris@demorecorder.com>
1824    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
1825    Patch-Set: QueryDeviceState
1826    Patch-Number: libxcb 2/3
1827    Patch-Version: V1
1828    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1829    Reviewed-By: Ran Benita <ran234@gmail.com>
1830
1831commit 382d306d6c44a9ece5551c210a932773b5cb94a5
1832Author: Emil Velikov <emil.l.velikov@gmail.com>
1833Date:   Fri Sep 5 01:46:40 2014 +0100
1834
1835    Move internal/private dependencies to Requires.private
1836    
1837    Program using the xcb sub-modules has indirect compile and runtime
1838    dependency of core xcb. To ensure this out we currently list xcb in
1839    the Requires field of the pkg-config files. While this provides all
1840    the required dependencies for successful compilation this causes
1841    over-linking and hides potential linking miss-use against the xcb modules.
1842    
1843    By moving to Requires.private we retain the compilation and runtime
1844    compatibility and avoids any runtime problems.
1845    
1846    Cc: Keith Packard <keithp@keithp.com>
1847    Cc: Alan Coopersmith <alan.coopersmith@oracle.com>
1848    References: http://people.freedesktop.org/~dbn/pkg-config-guide.html#faq
1849    References: https://wiki.mageia.org/en/Overlinking_issues_in_packaging
1850    References: http://err.no/personal/blog/2008/Mar/25
1851    Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
1852    Reviewed-by: Julien Cristau <jcristau@debian.org>
1853    Signed-off-by: Uli Schlachter <psychon@znc.in>
1854
1855commit d74d066949dbbbbbcb03bf7764e63f4347f99974
1856Author: Christian Linhart <chris@DemoRecorder.com>
1857Date:   Sat Sep 6 20:06:15 2014 +0200
1858
1859    generator: support fixed size lists in var-sized structs
1860    
1861    V2: patch revised according to suggestions from Ran Benita:
1862    * removed blanks before an after parentheses of function-calls or tuples
1863    * replaced if by elif in "if field.type.is_list". ( this fixes old code )
1864    
1865    Message-ID: <540B4D17.1080908@DemoRecorder.com>
1866    Patch-Thread-Subject: [Xcb] xinput:QueryDeviceState: full-support: generator and xml-changes
1867    Patch-Set: QueryDeviceState
1868    Patch-Number: libxcb 1/3
1869    Patch-Version: V2
1870    Signed-off-by: Christian Linhart <chris@DemoRecorder.com>
1871    Reviewed-By: Ran Benita <ran234@gmail.com>
1872
1873commit b0e6c2de09c7474868dd7185674fa113a5c2e0aa
1874Author: Uli Schlachter <psychon@znc.in>
1875Date:   Mon Aug 18 10:38:48 2014 +0200
1876
1877    xcb_get_setup(): Never return NULL
1878    
1879    The documentation doesn't mention it and it's unlikely that a lot of code out
1880    there handles this case correctly. So, instead of returning NULL, let
1881    xcb_get_setup() return a pointer to a static, invalid, all-zero setup
1882    information structure.
1883    
1884    Signed-off-by: Uli Schlachter <psychon@znc.in>
1885
1886commit c4e40f646b8da4fd112ea54a612c880be5e942a8
1887Author: Uli Schlachter <psychon@znc.in>
1888Date:   Mon Aug 18 10:38:41 2014 +0200
1889
1890    Make some functions also work on error connections
1891    
1892    There is no technical reason why xcb_get_setup() and xcb_get_file_descriptor()
1893    shouldn't work on non-static error connections. They cannot be used for many
1894    useful things, but at least they work.
1895    
1896    This works around bugs in lots of programs out there which assume that
1897    xcb_get_setup() does not return NULL and which just happily dereference the
1898    results. Since xcb_connect() never returns NULL, it's a bit weird that
1899    xcb_get_setup() can do so. xcb_get_file_descriptor() is just modified since this
1900    can be done here equally easily and because the fd isn't closed until the final
1901    xcb_disconnect() on the error connection.
1902    
1903    Non-static error connections are connections which entered an error state after
1904    xcb_connect() succeeded. If something goes wrong in establishing a connection,
1905    xcb_connect() will return a static error connection which doesn't have the
1906    fields used here.
1907    
1908    Signed-off-by: Uli Schlachter <psychon@znc.in>
1909
1910commit 355d4d6ab9f5c12c2ee4a91e8cf6eb4a2854d73c
1911Author: Christian Linhart <chris@demorecorder.com>
1912Date:   Tue Aug 19 15:57:34 2014 +0200
1913
1914    support switch case in the generator
1915    
1916    The implementation is rather simple:
1917    When a <case> is used instead of a <bitcase>
1918    then operator "==" is used instead of "&" in the if-condition.
1919    
1920    So it creates a series of "if" statements
1921    (instead of a switch-case statement in C )
1922    
1923    In practice this does not matter because a good
1924    optimizing compiler will create the same code
1925    as for a switch-case.
1926    
1927    With this simple implementation we get additional
1928    flexibility in the following forms:
1929    * a case value may appear in multiple case branches.
1930      for example:
1931            case C1 will be selected by values 1, 4, or 5
1932            case C2 will be selected by values 3, 4, or 7
1933    
1934    * mixing of bitcase and case is possible
1935            (this will usually make no sense but there may
1936            be protocol specs where this is needed)
1937    
1938    details of the impl:
1939    * replaced "is_bitcase" with "is_case_or_bitcase" in all places
1940      so that cases are treated like bitcases.
1941    
1942    * In function "_c_serialize_helper_switch": write operator "=="
1943      instead of operator "&" if it is a case.
1944
1945commit 966fba6ba4838949d0727dfa45eeb9392d1f85d9
1946Author: Daniel Martin <consume.noise@gmail.com>
1947Date:   Tue Jul 29 22:48:44 2014 +0200
1948
1949    Disable Xevie and Xprint by default
1950    
1951    Both extensions have been dropped from the X-Server in 2008:
1952        http://cgit.freedesktop.org/xorg/xserver/commit/?id=1c8bd31
1953        http://cgit.freedesktop.org/xorg/xserver/commit/?id=f4036f6
1954    
1955    Don't build them by default.
1956    
1957    Reviewed-by: Julien Cristau <jcristau@debian.org>
1958    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
1959    Signed-off-by: Uli Schlachter <psychon@znc.in>
1960
1961commit a5e90ae6a1543a681d95b831dc5c44e9c6e78610
1962Author: Gaetan Nadon <memsize@videotron.ca>
1963Date:   Wed Mar 26 15:40:57 2014 -0400
1964
1965    help text: do not report the insanly long list of Warning flags.
1966    
1967    Originally there was just one. Now that XCB has been integrated with X and
1968    uses the same compiler flags, it is a different story.
1969    
1970      Used CFLAGS:
1971        CPPFLAGS............:
1972        CFLAGS..............: -g -O2
1973        Warning CFLAGS......:  -Wall -Wpointer-arith AND SO ON FOR 8 lines...
1974    
1975    It completely defaces the otherwise excellent output.
1976    
1977    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
1978    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
1979    Signed-off-by: Uli Schlachter <psychon@znc.in>
1980
1981commit 529e3bfc2088dcd35e68a0a8394d0e9bff7c978b
1982Author: Gaetan Nadon <memsize@videotron.ca>
1983Date:   Wed Mar 26 15:40:56 2014 -0400
1984
1985    Add ChangeLog and INSTALL using xorg macros
1986    
1987    Same as all other X modules. The one in libxcb git is removed.
1988    
1989    Those files are created during 'make dist'
1990    
1991    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
1992    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
1993    Signed-off-by: Uli Schlachter <psychon@znc.in>
1994
1995commit 74f552c1b39eb6a1ab64451477a492a2b22d9d7a
1996Author: Gaetan Nadon <memsize@videotron.ca>
1997Date:   Wed Mar 26 15:24:46 2014 -0400
1998
1999    sendmsg: remove --enable-sendfds as it is superceeded by --enable-dri3
2000    
2001    DRI3 requires sendmsg support which is auto-detected. A builder can enable
2002    or disable dri3 feature. If sendmsg function is not available, dri3 cannot
2003    be enabled.
2004    
2005    This reverts af8067cbf4856 which was done at a time where --enable-dri3
2006    had not been added yet.
2007    
2008    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2009    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2010    Signed-off-by: Uli Schlachter <psychon@znc.in>
2011
2012commit 23f57ac8bf0812bbc851ec9a815e50a640b97db5
2013Author: Gaetan Nadon <memsize@videotron.ca>
2014Date:   Wed Mar 26 15:24:45 2014 -0400
2015
2016    config: issue an error if DRI3 is requested, but sendfds is not available
2017    
2018    When a user issues the --enable-dri3 option and sendfds is not available
2019    on the system, the configuration will abort with an error message.
2020    
2021    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2022    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2023    Signed-off-by: Uli Schlachter <psychon@znc.in>
2024
2025commit 7f07b57be587a2ebe0cadceba3fe67ed4a1e79db
2026Author: Gaetan Nadon <memsize@videotron.ca>
2027Date:   Wed Mar 26 15:24:44 2014 -0400
2028
2029    config: default option for enable-dri3 is not implemented
2030    
2031    The first symptom is the help text:
2032    
2033      --enable-dri3     Build XCB DRI3 Extension (default: "$sendfds")
2034    
2035    The implementation variable $sendfds leaked into the user interface.
2036    Testing the various user inputs:
2037      <nothing>                DRI3 is enabled     PASS
2038      --enable-dri3            DRI3 is enabled     PASS
2039      --enable-dri3=yes        DRI3 is enabled     PASS
2040      --enable-dri3=no         DRI3 is disabled    PASS
2041      --disable-dri3           DRI3 is disabled    PASS
2042      --enable-dri3=$sendfds   DRI3 is disabled    FAIL
2043    
2044    This patch implements the usual idiom for features that are enabled by
2045    default if the various conditions are met (sendfds is available).
2046    New help text:
2047    
2048      --enable-dri3     Build XCB DRI3 Extension (default: auto)
2049    
2050    With the additional user input:
2051    
2052      --enable-dri3=auto
2053    
2054    which is equivalent to providing no input at all.
2055    
2056    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2057    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2058    Signed-off-by: Uli Schlachter <psychon@znc.in>
2059
2060commit d1e8ec96fca4862f37ec9f0e9407bb989c4c161a
2061Author: Uli Schlachter <psychon@znc.in>
2062Date:   Fri Aug 1 15:56:52 2014 +0200
2063
2064    Release libxcb 1.11
2065
2066commit c5c6cfa4d2a27d0f9e139a1856d866433b6046bc
2067Author: Uli Schlachter <psychon@znc.in>
2068Date:   Fri Aug 1 16:03:24 2014 +0200
2069
2070    Bump xcb-proto requirement to 1.11
2071    
2072    This is needed for the new direct_imports field that we need from xcb-proto.
2073    
2074    Signed-off-by: Uli Schlachter <psychon@znc.in>
2075
2076commit 70ea5da64b34336bb0916f6c325545cb50746159
2077Author: Alexander Mezin <mezin.alexander@gmail.com>
2078Date:   Sun Jun 29 17:33:48 2014 +0700
2079
2080    xcb.h: add 'struct' before xcb_setup_t, xcb_query_extension_reply_t
2081    
2082    These structs are typedef'ed in xproto.h, so in xcb.h these types
2083    (without 'struct') are actually undefined.
2084    
2085    GCC reports this as error when building precompiled header.
2086    
2087    Signed-off-by: Alexander Mezin <mezin.alexander@gmail.com>
2088    Reviewed-by: Peter Harris <pharris@opentext.com>
2089
2090commit 7e6af51b4e984f661fe4f21596cab5cb8ee15ea0
2091Author: Ran Benita <ran234@gmail.com>
2092Date:   Tue Feb 25 14:11:35 2014 +0200
2093
2094    c_client.py: remove more trailing space from generated files
2095    
2096    Signed-off-by: Ran Benita <ran234@gmail.com>
2097    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2098    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2099
2100commit 8221d249b77131b338e3b35ce2229193f129e514
2101Author: Ran Benita <ran234@gmail.com>
2102Date:   Sun Feb 23 22:55:21 2014 +0200
2103
2104    c_client.py: remove trailing whitespace from generated files
2105    
2106    Signed-off-by: Ran Benita <ran234@gmail.com>
2107    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2108
2109commit e3c728ee3d9a2fd7478d5f57830c3483b774a16e
2110Author: Ran Benita <ran234@gmail.com>
2111Date:   Sun Feb 23 22:55:20 2014 +0200
2112
2113    c_client.py: remove useless generated comments
2114    
2115    They are bloated, don't add anything over the signature, in some cases
2116    duplicate the doxygen comments, and are not integrated with the <doc>
2117    tags in any way. Remove them and cut the generated LOC by half.
2118    
2119    Signed-off-by: Ran Benita <ran234@gmail.com>
2120    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2121
2122commit cae2e398563841c5b814596fd1f1c64354dcac71
2123Author: Ran Benita <ran234@gmail.com>
2124Date:   Sun Feb 23 22:55:19 2014 +0200
2125
2126    c_client.py: make the man page output deterministic
2127    
2128    Some parts of the man pages (SEE ALSO and ERRORS) are generated by
2129    iterating a Python dict. But the iteration order in a dict is random,
2130    so each build the output is ordered differently. Avoid that by iterating
2131    in sorted order.
2132    
2133    Signed-off-by: Ran Benita <ran234@gmail.com>
2134    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2135
2136commit bfbf83b1d8113ac398b57c2738706792946d1c03
2137Author: Ran Benita <ran234@gmail.com>
2138Date:   Sun Feb 23 22:55:18 2014 +0200
2139
2140    c_client.py: prefix all monkey-patched fields with c_
2141    
2142    The script adds many fields to the objects coming from xcbgen. To
2143    distinguish them, a c_ prefix is used, but for some it was missing.
2144    
2145    Signed-off-by: Ran Benita <ran234@gmail.com>
2146    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2147
2148commit 285d566a5c11c6edd3665beb42312e24bde77d16
2149Author: Ran Benita <ran234@gmail.com>
2150Date:   Sun Feb 23 22:55:17 2014 +0200
2151
2152    c_client.py: remove trailing whitespace
2153    
2154    These are extra annoying in python code.
2155    
2156    Signed-off-by: Ran Benita <ran234@gmail.com>
2157    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2158
2159commit 285838cfe41c212a2453d903497757d2392fd0ce
2160Author: Ran Benita <ran234@gmail.com>
2161Date:   Sun Feb 23 22:55:16 2014 +0200
2162
2163    c_client.py: remove useless 'today' variable
2164    
2165    Signed-off-by: Ran Benita <ran234@gmail.com>
2166    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2167
2168commit 49a61c8b459ab19c7f39e653bbb0d0339ea8f00f
2169Author: Alan Coopersmith <alan.coopersmith@oracle.com>
2170Date:   Fri Jul 11 20:41:15 2014 -0700
2171
2172    Fix typos & awkward wording in tutorial
2173    
2174    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2175
2176commit 125135452a554e89e49448e2c1ee6658324e1095
2177Author: Michael Haubenwallner <michael.haubenwallner@salomon.at>
2178Date:   Fri Jun 13 16:18:34 2014 +0200
2179
2180    bug#79986: include system headers early
2181    
2182    AIX <sys/poll.h> does redefine 'events' to 'reqevents' eventually.
2183    To not have this cause compilation errors, need to include the local
2184    header files after any system header file.
2185    
2186    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2187    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2188
2189commit bc5a1047548b578624cfbc44ca192cde7664ed78
2190Author: Alan Coopersmith <alan.coopersmith@oracle.com>
2191Date:   Fri Jun 13 21:26:21 2014 -0700
2192
2193    Document failure modes of xcb_connect*() functions
2194    
2195    Documentation was previously unclear that these always return a non-NULL
2196    pointer, and that callers need to check it for error values, instead of
2197    checking for a NULL return value.
2198    
2199    Triggered by having to dig through code to answer a user's question on
2200    the #xcb irc channel, since neither of us found it covered in the docs.
2201    
2202    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2203    Reviewed-by: Uli Schlachter <psychon@znc.in>
2204
2205commit 72e45969ff71204cee2dde3502841736cfd41c8a
2206Author: Daniel Martin <consume.noise@gmail.com>
2207Date:   Mon Jun 9 17:55:04 2014 +0200
2208
2209    Handle <pad align="n" /> between lists
2210    
2211    Without this patch we end up with invalid C code if we've a
2212    <pad align="n" /> between two variadic lists. Check for such a condition
2213    and take the alignment pad into account.
2214    
2215    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=79808
2216    
2217    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
2218    Signed-off-by: Peter Harris <pharris@opentext.com>
2219
2220commit d978a4f69b30b630f28d07f1003cf290284d24d8
2221Author: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2222Date:   Sat Apr 5 20:54:59 2014 -0700
2223
2224    xcb_open: Improve abstraction for launchd secure sockets
2225    
2226    This changes away from hard-coding the /tmp/launch-* path to now
2227    supporting a generic <path to unix socket>[.<screen>] format for
2228    $DISPLAY.
2229    
2230    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2231
2232commit 29e419c5840a1eeda3336a0802686ee723dcaab3
2233Author: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2234Date:   Sat Apr 5 21:09:42 2014 -0700
2235
2236    xcb_open: Minor code cleanup for better readability
2237    
2238    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
2239    Reviewed-by: Uli Schlachter <psychon@znc.in>
2240
2241commit be0fe56c3bcad5124dcc6c47a2fad01acd16f71a
2242Author: Keith Packard <keithp@keithp.com>
2243Date:   Mon Dec 23 21:15:20 2013 -0800
2244
2245    Ensure xcb owns socket and no other threads are writing before send_request
2246    
2247    send_request may only write to out.queue if no other thread is busy
2248    writing to the network (as that thread may be writing from out.queue).
2249    
2250    send_request may only allocate request sequence numbers if XCB owns
2251    the socket.
2252    
2253    Therefore, send_request must make sure that both conditions are true
2254    when it holds iolock, which can only be done by looping until both
2255    conditions are true without having dropped the lock waiting for the
2256    second condition.
2257    
2258    We choose to get the socket back from Xlib first as get_socket_back
2259    has a complicated test and checking for other threads writing is a
2260    simple in-lined check.
2261    
2262    This also changes the sequence number checks (64k requests with no
2263    reply, 4M request wrapping) to ensure that both conditions are true
2264    before queueing the request.
2265    
2266    Signed-off-by: Keith Packard <keithp@keithp.com>
2267    Reviewed-by: Uli Schlachter <psychon@znc.in>
2268
2269commit e2813e1cde893f384fa620ff3c13493beebabe0c
2270Author: Keith Packard <keithp@keithp.com>
2271Date:   Wed Feb 12 14:15:46 2014 -0800
2272
2273    Update .pc file Requires lines to express full dependencies
2274    
2275    Some xcb libraries depend on others; make these dependencies explicit
2276    in the .pc files that are installed.
2277    
2278    This change was generated automatically by running 'check-pc-requires -fix'
2279    
2280    Signed-off-by: Keith Packard <keithp@keithp.com>
2281    Signed-off-by: Uli Schlachter <psychon@znc.in>
2282
2283commit 32de4c4213a49c61127c6957ea05fef3e5355291
2284Author: Keith Packard <keithp@keithp.com>
2285Date:   Wed Feb 12 14:15:45 2014 -0800
2286
2287    Validate .pc file Requires lines
2288    
2289    This walks through the .pc.in files and makes sure all of the Requires
2290    lines express sufficient dependency information.
2291    
2292    Signed-off-by: Keith Packard <keithp@keithp.com>
2293    Signed-off-by: Uli Schlachter <psychon@znc.in>
2294
2295commit 1f6cd9f1fcd3c07d323b678292c9cb00ae1f7504
2296Author: Keith Packard <keithp@keithp.com>
2297Date:   Wed Feb 12 14:15:44 2014 -0800
2298
2299    Only #include directly referenced module header files
2300    
2301    This avoids having the nested header files also included at the top
2302    level, which is more efficient.
2303    
2304    Signed-off-by: Keith Packard <keithp@keithp.com>
2305    Signed-off-by: Uli Schlachter <psychon@znc.in>
2306
2307commit cb686b576739deea00180c54697c8b62b8419ae0
2308Author: Uli Schlachter <psychon@znc.in>
2309Date:   Tue Feb 25 15:50:50 2014 +0100
2310
2311    Add doxygen documentation to functions in xcbext.h
2312    
2313    Signed-off-by: Uli Schlachter <psychon@znc.in>
2314    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2315
2316commit 2fb14e5883f2ea2f01d248674cfcc26ccb704753
2317Author: Uli Schlachter <psychon@znc.in>
2318Date:   Tue Dec 31 15:18:01 2013 +0100
2319
2320    Make xcb_disconnect(NULL) safe
2321    
2322    Code can be simplified if the deallocation functions can always be called in
2323    cleanup code. So if you have some code that does several things that can go
2324    wrong, one of which is xcb_connect(), after this change, the xcb_connection_t*
2325    variable can be initialized to NULL and xcb_disconnect() can always be called on
2326    the connection object.
2327    
2328    References: http://lists.freedesktop.org/archives/xcb/2013-September/008659.html
2329    
2330    Signed-off-by: Uli Schlachter <psychon@znc.in>
2331    Reviewed-by: Julien Cristau <jcristau@debian.org>
2332
2333commit 4dcbfd77b78ca6b016ce815af26235501f6cd75a
2334Author: Uli Schlachter <psychon@znc.in>
2335Date:   Tue Dec 31 15:05:36 2013 +0100
2336
2337    xcb_disconnect(): Fix leak with error connections
2338    
2339    There are two kind of error connections in XCB. First, if something goes wrong
2340    while the connection is being set up, _xcb_conn_ret_error() is used to return a
2341    static connection in an error state. If something goes wrong later,
2342    _xcb_conn_shutdown() is used to set c->has_error.
2343    
2344    This is important, because the static object that _xcb_conn_ret_error() returns
2345    must not be freed, while the dynamically allocated objects that go through
2346    _xcb_conn_shutdown() must obviously be properly deallocated.
2347    
2348    This used to work correctly, but in 769acff0da8, xcb_disconnect() was made to
2349    ignore all connections in an error state completely. Fix this by only ignoring
2350    the few static error connections that we have.
2351    
2352    This was tested with the following hack:
2353    
2354        xcb_connection_t *c = xcb_connect(NULL, NULL);
2355        close(xcb_get_file_descriptor(c));
2356        xcb_discard_reply(c, xcb_get_input_focus(c).sequence);
2357        xcb_flush(c);
2358        xcb_disconnect(c);
2359    
2360    Valgrind confirms that xcb has a memory leak before this patch that this patch
2361    indeed fixes.
2362    
2363    Signed-off-by: Uli Schlachter <psychon@znc.in>
2364    Reviewed-by: Julien Cristau <jcristau@debian.org>
2365
2366commit d84dd752ef571491b015443fefedca53c7f81282
2367Author: Uli Schlachter <psychon@znc.in>
2368Date:   Mon Sep 9 13:04:11 2013 +0200
2369
2370    Remove tabs and trailing whitespaces
2371    
2372    Signed-off-by: Uli Schlachter <psychon@znc.in>
2373
2374commit 4ffa6f83b92763eb901c7ddb7c20775e24d507ca
2375Author: Ran Benita <ran234@gmail.com>
2376Date:   Sat Jan 18 17:10:53 2014 +0200
2377
2378    Add comments about how _xcb_conn_ret_error() works
2379    
2380    If xcb_connect() fails, it doesn't return NULL. Instead, it always
2381    returns an xcb_connection_t*, and the user should check for errors with
2382    the xcb_connection_has_error() function. What this function does is
2383    check if conn->has_error contains a non-zero error code, and returns it.
2384    
2385    If an error did occur, xcb doesn't actually return a full
2386    xcb_connection_t though, it just returns (xcb_connection_t *)
2387    error_code. Since the 'has_error' field is the first, it is still
2388    possible to check conn->has_error.
2389    
2390    That last trick was not immediately obvious to me, so add some guiding
2391    comments. This also ensures no one obliviously rearranges the struct.
2392    
2393    Signed-off-by: Ran Benita <ran234@gmail.com>
2394    Signed-off-by: Uli Schlachter <psychon@znc.in>
2395
2396commit d7eb0bdf3b5b11ee9f40ee5e73df8fc0bdfa59f3
2397Author: Gaetan Nadon <memsize@videotron.ca>
2398Date:   Tue Jan 7 14:02:21 2014 -0500
2399
2400    generated man pages: use xorg footer and no hard coded extension
2401    
2402    The section number is no longer hard-coded
2403    The left footer is now "X Version 11".
2404    The center footer is the package name with the version, "libxcb 1.9"
2405    The three values above are provided through xorg-macros. They are passed-in
2406    to the python c_client code.
2407    
2408    Example of footer (last line, above dotted line)
2409    
2410    [...]
2411    AUTHOR
2412           Generated from xproto.xml. Contact xcb@lists.freedesktop.org for cor‐
2413           rections and improvements.
2414    
2415    X Version 11                      libxcb 1.9                 xcb_send_event(3)
2416    
2417    ------------------------------------------------------------------------------
2418    
2419    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2420    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2421
2422commit e4061b8f00d301a51d4c9eef4a07f1e87592fe85
2423Author: Gaetan Nadon <memsize@videotron.ca>
2424Date:   Tue Jan 7 14:02:16 2014 -0500
2425
2426    generated man pages: build without hard coded extension
2427    
2428    The automake MAN primary requires a hard coded extension to build
2429    man pages. Let's avoid that as the extension number may vary by platform.
2430    Take advantage of the fact that the man directory only contains man pages.
2431    Wildcards are not supported by Automake but it happens to work
2432    sufficiently well here.
2433    
2434    Normally xorg build man pages by converting a source .man file to a
2435    target file with the extension number. That would be too many files
2436    in this case.
2437    
2438    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2439    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2440
2441commit 3cdd524cadc4352ebd9e17b1f73134bec1838b40
2442Author: Gaetan Nadon <memsize@videotron.ca>
2443Date:   Tue Jan 7 14:02:05 2014 -0500
2444
2445    man: build static man pages using xorg patterns
2446    
2447    The section number is no longer hard-coded, supplied by xorg-macros.
2448    The left footer is now "X Version 11".
2449    The center footer is the package name with the version, "libxcb 1.9"
2450    The man directory is a sibbling to the doc directory. One can build
2451    or clean the man pages without disturbing the library code.
2452    
2453    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2454    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2455
2456commit c056adcd92daa06f4825d5c85a40e140a3e85b42
2457Author: Gaetan Nadon <memsize@videotron.ca>
2458Date:   Thu Jan 9 14:32:18 2014 -0500
2459
2460    autoconf: replace all tabs with spaces
2461    
2462    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2463    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2464
2465commit 412928f113c8a5e5b30c03a294a42b0b1cf5f5d7
2466Author: Gaetan Nadon <memsize@videotron.ca>
2467Date:   Tue Jan 7 14:00:44 2014 -0500
2468
2469    autoconf: use default xorg configuration for doxygen documentation
2470    
2471    No content or form changes for the xcb manual or tutorial.
2472    Only the configuration user visible bits change.
2473    
2474    Xcb will now have the same configuration options as the 30 other
2475    xorg modules.
2476    
2477    Xorg classifies documentation as "user", "developer" or "specifications".
2478    The xcb manual falls under the "developer" category. Developers docs
2479    are never installed under $prefix.
2480    
2481    A builder can selectively turn on/off any or all of the categories. He can
2482    also selectively turn on/off any of the many tools used to generate
2483    documentation such as doxygen, xmlto, etc... Each tool has an environment
2484    variable defined such as DOXYGEN.
2485    
2486    Other features are available, the user interface and the functionality
2487    is the same on all modules.
2488    
2489    --with-doxygen=FILE is replaced with DOXYGEN env variable
2490    --disable-build-docs is replaced with --disable-devel-docs
2491    
2492    The new interface displayed with ./configure --help:
2493    
2494      --enable-devel-docs     Enable building the developer documentation
2495                              (default: yes)
2496      --with-doxygen          Use doxygen to regenerate documentation (default:
2497                              auto)
2498      DOXYGEN     Path to doxygen command
2499      DOT         Path to the dot graphics utility
2500    
2501    The dot tool checking has been added to util-macros in version 1.18.
2502    
2503    Refer to the table of existing docs in xorg.
2504    XCB will be added for the doxygen generated API manual.
2505    
2506    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2507    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2508
2509commit 9ba6aa759e937e60b231b209b1293a40ad6c7a8a
2510Author: Gaetan Nadon <memsize@videotron.ca>
2511Date:   Tue Jan 7 14:00:43 2014 -0500
2512
2513    autoconf: fix warning by replacing deprecated AC_HELP_STRING
2514    
2515    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2516    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2517
2518commit 942eabaae3600e7277aa09a179ef10c9a06de62f
2519Author: Gaetan Nadon <memsize@videotron.ca>
2520Date:   Tue Jan 7 14:00:42 2014 -0500
2521
2522    autoconf: require libtool minimum level 2.2
2523    
2524    This is the updated minimum level as referenced in:
2525    http://www.x.org/wiki/Building_the_X_Window_System/#index2h3
2526    
2527    Libtool version 2 has been used for several years now. There should be
2528    no surprises.
2529    
2530    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2531    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2532
2533commit 517cb0e888eca4030b4bd00c316619db925032ea
2534Author: Gaetan Nadon <memsize@videotron.ca>
2535Date:   Tue Jan 7 14:00:41 2014 -0500
2536
2537    autoconf: comment and layout the initialization section
2538    
2539    No functional changes. Trying to make it clearer.
2540    
2541    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2542    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2543    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2544
2545commit 0a17b61a8971d3c4600a5c19fa7d753e7196abfb
2546Author: Gaetan Nadon <memsize@videotron.ca>
2547Date:   Tue Jan 7 14:00:40 2014 -0500
2548
2549    autoconf: AC_INIT: add bug url
2550    
2551    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2552    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2553
2554commit 414b1037c5a9770270022e62aa66fe7a4021a96b
2555Author: Gaetan Nadon <memsize@videotron.ca>
2556Date:   Tue Jan 7 14:00:39 2014 -0500
2557
2558    autoconf: use the warning variables from xorg
2559    
2560    The BASE_CFLAGS variable contains only warnings, just like the XCB
2561    version of CWARNFLAGS. This will result in no changes in the binaries
2562    produced. Xorg was missing -fd for SUNCC so it has been added to util-macros
2563    v 1.18.
2564    
2565    Do not get confused with the xorg deprecated CWARNFLAGS variable which
2566    contains an option that is not a warning, -fno-strict-aliasing. This
2567    option, should it be needed, can be added using the XORG_TESTSET_CFLAG
2568    macro.
2569    
2570    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2571    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2572
2573commit c4f2c70bc37a592406b7693562c7513f2f99b34d
2574Author: Gaetan Nadon <memsize@videotron.ca>
2575Date:   Tue Jan 7 14:00:38 2014 -0500
2576
2577    autoconf: use XORG_DEFAULT_OPTIONS
2578    
2579    XCB has been part of X.Org for a while now. This patch will harmonize the XCB
2580    configuration, using xorg-macros series of macros. It is already used in the
2581    XCB utils packages and is needed to build xcb-proto.
2582    
2583    The XORG_DEFAULT_OPTIONS already includes the statement for the silent
2584    rules.
2585    
2586    The AC_PROG_CC statement is removed so as not to override AC_PROG_CC_C99
2587    in XORG_DEFAULT_OPTIONS. The effective change is that xcb now uses c99 as
2588    requested.
2589    
2590    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2591    Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
2592
2593commit e7263931aff3e3450dc938ad465a7577f943549f
2594Author: Peter Harris <pharris@opentext.com>
2595Date:   Tue Jan 14 14:50:55 2014 -0500
2596
2597    Support <pad align="n" />
2598    
2599    Reviewed-By: Ran Benita <ran234@gmail.com>
2600    Signed-off-by: Peter Harris <pharris@opentext.com>
2601
2602commit 3b72a2c9d1d656c74c691a45689e1d637f669e3a
2603Author: Kenneth Graunke <kenneth@whitecape.org>
2604Date:   Fri Jan 3 15:08:33 2014 -0800
2605
2606    Force XCB event structures with 64-bit extended fields to be packed.
2607    
2608    With the advent of the Present extension, some events (such as
2609    PresentCompleteNotify) now use native 64-bit types on the wire.
2610    
2611    For XGE events, we insert an extra "uint32_t full_sequence" field
2612    immediately after the first 32 bytes of data.  Normally, this causes
2613    the subsequent fields to be shifted over by 4 bytes, and the structure
2614    to grow in size by 4 bytes.  Everything works fine.
2615    
2616    However, if event contains 64-bit extended fields, this may result in
2617    the compiler adding an extra 4 bytes of padding so that those fields
2618    remain aligned on 64-bit boundaries.  This causes the structure to grow
2619    by 8 bytes, not 4.  Unfortunately, XCB doesn't realize this, and
2620    always believes that the length only increased by 4.  read_packet()
2621    then fails to malloc enough memory to hold the event, and the event
2622    processing code uses the wrong offsets.
2623    
2624    To fix this, mark any event structures containing 64-bit extended
2625    fields with __attribute__((__packed__)).
2626    
2627    v2: Use any(...) instead of True in (...), as suggested by
2628        Daniel Martin.
2629    
2630    v3 (Alan Coopersmith): Fix build with Solaris Studio 12.3 by moving the
2631    attribute to after the structure definition.
2632    
2633    Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2634    Reviewed-by: Keith Packard <keithp@keithp.com> [v1]
2635    Reviewed-by: Josh Triplett <josh@joshtriplett.org> [v1]
2636    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2637    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2638
2639commit 010872f611a044ced4e96b18a7514796b2a443df
2640Author: Uli Schlachter <psychon@znc.in>
2641Date:   Sun Dec 22 15:59:24 2013 +0100
2642
2643    Release libxcb 1.10
2644    
2645    Signed-off-by: Uli Schlachter <psychon@znc.in>
2646
2647commit f653464554469b5767f1c99abced25a76bace047
2648Author: Julien Cristau <jcristau@debian.org>
2649Date:   Sat Dec 14 06:16:37 2013 +0100
2650
2651    Add NEWS for 1.10
2652    
2653    Signed-off-by: Julien Cristau <jcristau@debian.org>
2654
2655commit 9c2a6dc20c64ce93e0acd2fceec6d3cab8fc9134
2656Author: Uli Schlachter <psychon@znc.in>
2657Date:   Fri Nov 15 22:33:12 2013 +0100
2658
2659    Add NEWS entries for releases 1.9.1 to 1.9.3
2660    
2661    libxcb 1.9.1 was released from a branch and thus its NEWS entries never made it
2662    into the master branch. The other releases didn't update NEWS.
2663    
2664    Signed-off-by: Uli Schlachter <psychon@znc.in>
2665    Signed-off-by: Julien Cristau <jcristau@debian.org>
2666
2667commit ea4406a8e0193cad8effe569e7835f8b67894e54
2668Author: Julien Cristau <jcristau@debian.org>
2669Date:   Sat Dec 14 05:54:20 2013 +0100
2670
2671    Bump xcb-proto requirement to 1.10
2672    
2673    Makes sure we generate the new generic event struct.
2674    
2675    Signed-off-by: Julien Cristau <jcristau@debian.org>
2676
2677commit a1299eb2a210b5788a2b827b82a3d825caa1f201
2678Author: PHO <pho@cielonegro.org>
2679Date:   Tue Dec 3 12:43:04 2013 +0900
2680
2681    Test the value of msg_controllen for platforms whose CMSG_FIRSTHDR() does not test it for us
2682    
2683    As RFC 2292 points out, some platforms (e.g. Darwin 9.8.0) provide
2684    CMSG_FIRSTHDR(msg) which just returns msg.msg_control without first
2685    checking if msg.msg_controllen is non-zero. We need a workaround for
2686    such platforms not to let _xcb_in_read() segfault.
2687    
2688    https://bugs.freedesktop.org/show_bug.cgi?id=72253
2689    
2690    Signed-off-by: Julien Cristau <jcristau@debian.org>
2691
2692commit b30b11ac49d934541312b03c41d1ab83047a59f4
2693Author: Uli Schlachter <psychon@znc.in>
2694Date:   Mon Nov 18 20:28:08 2013 +0100
2695
2696    Increment the "current" version info for sync, xinput and xkb
2697    
2698    Sync: Due to commit e6a246e50e62cbcba3 "sync: Change value list param of
2699    CreateAlarm and ChangeAlarm into switch", various symbols disappeared,
2700    for example xcb_sync_{change,create}_alarm_sizeof.
2701    
2702    xinput: This extension was updated from version 1.4 to 2.3. This means
2703    that lots of new things are generated. However, this change is
2704    backwards-compatible and thus age gets set to 1.
2705    
2706    xkb: In commit 37d0f55392d6 "xkb: Work around alignment problems in
2707    GetNames and GetMap replies", some padding fields were introduced into
2708    structures for which an _unpack() function is generated. This changed
2709    the size of the struct and caused offsets into this struct to change.
2710    
2711    https://bugs.freedesktop.org/show_bug.cgi?id=71507
2712    
2713    Signed-off-by: Uli Schlachter <psychon@znc.in>
2714    Signed-off-by: Julien Cristau <jcristau@debian.org>
2715
2716commit ce5395eb4611341ba7c243ed524d023a616f73bb
2717Author: Uli Schlachter <psychon@znc.in>
2718Date:   Mon Nov 18 20:30:18 2013 +0100
2719
2720    Revert "Remove xcb_ge_event_t from xcb.h"
2721    
2722    This reverts commit f4d5b84800f960831e4fbb3ad9848bbb701020be.
2723    
2724    The version of this struct that the code generator produces breaks the API,
2725    because it gives the fields different (albeit better) names. Thus, we need to
2726    restore the old version of this struct.
2727    
2728    Additionally to the revert, this struct is documented as being deprecated. The
2729    replacement was added to xcb-proto.
2730    
2731    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71502
2732    Signed-off-by: Uli Schlachter <psychon@znc.in>
2733    Signed-off-by: Julien Cristau <jcristau@debian.org>
2734
2735commit 18f0afab3f0de68114fe185e89d8b25a8c072a2c
2736Author: Daniel Martin <consume.noise@gmail.com>
2737Date:   Fri Nov 22 23:27:28 2013 +0100
2738
2739    c_client.py: Fix _sizeof() functions
2740    
2741    Currently, it is not possible to correctly iterate over the replies of
2742    some requests. For example, the list of XIDeviceInfo returned by
2743    the XIQueryDevice request from xinput2 is read as garbage starting from
2744    the second entry.
2745    
2746    The culprits are the _sizeof() used by the iterators. In the above case:
2747    
2748        int
2749        xcb_input_xi_device_info_sizeof (const void  *_buffer  /**< */)
2750        {
2751            char *xcb_tmp = (char *)_buffer;
2752            [...]
2753            unsigned int xcb_block_len = 0;
2754            [...]
2755    
2756            xcb_block_len += sizeof(xcb_input_xi_device_info_t);
2757            xcb_tmp += xcb_block_len;
2758            /* name */
2759            xcb_block_len += (((_aux->name_len + 3) / 4) * 4) * sizeof(char);
2760            xcb_tmp += xcb_block_len;
2761            [...]
2762        }
2763    
2764    The problem here is that `xcb_block_len` is not zero'd right above the
2765    `/* name */` comment, causing `xcb_tmp` to be incremented by
2766    `sizeof(xcb_input_xi_device_info_t)` twice. The returned size is too
2767    large.
2768    
2769    https://bugs.freedesktop.org/show_bug.cgi?id=68387
2770    
2771    Tested-by: Ran Benita <ran234@gmail.com>
2772    Reviewed-by: Ran Benita <ran234@gmail.com>
2773    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
2774    Signed-off-by: Ran Benita <ran234@gmail.com>
2775    Signed-off-by: Julien Cristau <jcristau@debian.org>
2776
2777commit 5d1dbb468bb0f834eaa8adea6daf6729808ca429
2778Author: Uli Schlachter <psychon@znc.in>
2779Date:   Mon Nov 18 19:49:41 2013 +0100
2780
2781    Revert "fix deadlock with xcb_take_socket/return_socket v3"
2782    
2783    This reverts commit 9ae84ad187e2ba440c40f44b8eb21c82c2fdbf12.
2784    
2785    After this patch was merged, there were complaints about it not being a good
2786    idea. Revert this for now until we can agree on this.
2787    
2788    References: http://lists.freedesktop.org/archives/xcb/2013-June/008340.html
2789    Signed-off-by: Uli Schlachter <psychon@znc.in>
2790    
2791    Conflicts:
2792            src/xcbint.h
2793
2794commit c7c5b710f2cc0782412c9e159986c96b52aa0d02
2795Author: Mark Kettenis <kettenis@openbsd.org>
2796Date:   Mon Nov 11 23:11:56 2013 +0100
2797
2798    Fix alignment issues in FD passing code
2799    
2800    A char array on the stack is not guaranteed to have more than byte alignment.
2801    This means that casting it to a 'struct cmsghdr' and accessing its members
2802    may result in unaligned access.  This will generate SIGBUS on struct
2803    alignment architectures like OpenBSD/sparc64.  The canonical solution is to
2804    use a union to force proper alignment.
2805    
2806    Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
2807    Reviewed-by: Matthieu Herrb <matthieu@herrb.eu>
2808    Signed-off-by: Uli Schlachter <psychon@znc.in>
2809
2810commit e8663a935890ff366f49e356211049dfd0d9756a
2811Author: Alan Coopersmith <alan.coopersmith@oracle.com>
2812Date:   Thu Nov 7 20:23:27 2013 -0800
2813
2814    Check if we need to define _XOPEN_SOURCE for struct msghdr.msg_control
2815    
2816    Required to expose the structure members in Solaris headers, since it
2817    was an XPG4/UNIX95 addition to the Solaris ABI.
2818    
2819    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2820
2821commit af8067cbf48561f1e2d43e153292e68e0376a8f9
2822Author: Keith Packard <keithp@keithp.com>
2823Date:   Thu Nov 7 17:36:01 2013 -0800
2824
2825    Add configure option to enable or disable fd passing with sendmsg
2826    
2827    --disable-sendfds or --enable-sendfds
2828    
2829    By default, configure auto-detects based on whether your system
2830    supports sendmsg at all.
2831    
2832    Signed-off-by: Keith Packard <keithp@keithp.com>
2833    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2834    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2835
2836commit 83f28ef8655acff746eab64eabe2e31f8cf0c892
2837Author: Keith Packard <keithp@keithp.com>
2838Date:   Thu Nov 7 17:28:45 2013 -0800
2839
2840    Switch to using the CMSG_* macros for FD passing
2841    
2842    Use these instead of computing the values directly so that it might
2843    work on BSD or other non-Linux systems
2844    
2845    Signed-off-by: Keith Packard <keithp@keithp.com>
2846    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2847    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
2848
2849commit cca607409068ad0948e7283fb8d0465cabc51686
2850Author: Keith Packard <keithp@keithp.com>
2851Date:   Thu Jul 11 16:01:02 2013 -0700
2852
2853    Add Present extension
2854    
2855    Signed-off-by: Keith Packard <keithp@keithp.com>
2856    Reviewed-By: Uli Schlachter <psychon@znc.in>
2857
2858commit 7a9373078e69b2cb2753570f91e5c31062ba25f8
2859Author: Keith Packard <keithp@keithp.com>
2860Date:   Tue Apr 9 21:35:52 2013 -0700
2861
2862    Add DRI3 library
2863    
2864    Signed-off-by: Keith Packard <keithp@keithp.com>
2865    Reviewed-By: Uli Schlachter <psychon@znc.in>
2866
2867commit 93d733e85ded5e92292d36fc7025f0c8ff7b1167
2868Author: Keith Packard <keithp@keithp.com>
2869Date:   Thu Nov 7 05:20:06 2013 -0800
2870
2871    Require xcb proto version 1.9
2872    
2873    Signed-off-by: Keith Packard <keithp@keithp.com>
2874
2875commit 79019541e7c56ddfc3828b7bf96e6e5d3cf81c56
2876Author: Keith Packard <keithp@keithp.com>
2877Date:   Fri Apr 12 20:15:41 2013 -0700
2878
2879    Add event queue splitting
2880    
2881    This allows apps to peel off certain XGE events into separate queues
2882    for custom handling. Designed to support the Present extension
2883    
2884    Signed-off-by: Keith Packard <keithp@keithp.com>
2885    Reviewed-By: Uli Schlachter <psychon@znc.in>
2886
2887commit 7983bf0fbdc2725403f9db6154d0f5bc944040e5
2888Author: Keith Packard <keithp@keithp.com>
2889Date:   Fri Jan 18 01:29:40 2013 -0800
2890
2891    Add support for receiving fds in replies
2892    
2893    Requests signal which replies will have fds, and the replies report
2894    how many fds they expect in byte 1.
2895    
2896    Signed-off-by: Keith Packard <keithp@keithp.com>
2897    Reviewed-By: Uli Schlachter <psychon@znc.in>
2898
2899commit 7b53fb0f9bddae77b3ab8823743db57faee4e99b
2900Author: Keith Packard <keithp@keithp.com>
2901Date:   Mon Jan 14 11:23:00 2013 -0800
2902
2903    Add xcb_send_fd API
2904    
2905    This uses sendmsg to transmit file descriptors from the application to
2906    the X server
2907    
2908    Signed-off-by: Keith Packard <keithp@keithp.com>
2909    Reviewed-By: Uli Schlachter <psychon@znc.in>
2910
2911commit 98c227a2222fb5c7ca7e8101b1ed2bc096a33048
2912Author: Keith Packard <keithp@keithp.com>
2913Date:   Fri Jan 18 01:28:56 2013 -0800
2914
2915    -pendantic is too pendantic
2916    
2917    Many system headers have warnings when compiled with this flag.
2918    
2919    Signed-off-by: Keith Packard <keithp@keithp.com>
2920    Reviewed-By: Uli Schlachter <psychon@znc.in>
2921
2922commit d04a4a03a90f2721d507287938c90f1755d9da0e
2923Author: Keith Packard <keithp@keithp.com>
2924Date:   Fri Jul 12 10:32:03 2013 -0700
2925
2926    Make protocol C files depend on protocol XML files
2927    
2928    When new XML files get installed, make sure the C files are regenerated
2929    
2930    Signed-off-by: Keith Packard <keithp@keithp.com>
2931    Reviewed-By: Uli Schlachter <psychon@znc.in>
2932
2933commit f4d5b84800f960831e4fbb3ad9848bbb701020be
2934Author: Keith Packard <keithp@keithp.com>
2935Date:   Wed Nov 6 19:33:53 2013 -0800
2936
2937    Remove xcb_ge_event_t from xcb.h
2938    
2939    xcb proto now publishes this structure from an XML description
2940    
2941    Signed-off-by: Keith Packard <keithp@keithp.com>
2942
2943commit e4e0c6eec861f4c69da12060dc8dbe7a63fa5eb6
2944Author: Daphne Pfister <daphnediane@mac.com>
2945Date:   Sat Sep 14 17:36:22 2013 -0400
2946
2947    Use /usr/spool/sockets/X11/ on HP-UX for UNIX sockets (#69118).
2948
2949commit f1405d9fe4a6ddcae24585ba254389a4c4f4c8c9
2950Author: Daphne Pfister <daphnediane@mac.com>
2951Date:   Sun Sep 8 16:25:11 2013 -0400
2952
2953    Fix poll() if POLLIN == ROLLRDNORM|POLLRDBAND
2954    
2955    It seems like POLLIN is specified as equivalent to POLLRDNORM | POLLRDBAND. Some
2956    systems (e.g. QNX and HP-UX) take this literaly and have POLLIN defined as the
2957    above bit combination. Other systems (e.g. Linux) have POLLIN as just a single
2958    bit.
2959    
2960    This means that if no out-of-band data is available (which should never be the
2961    case), the result of poll() will not fulfil (fd.revents & POLLIN) == POLLIN on
2962    QNX, because the POLLRDBAND bit is not set.
2963    
2964    In other words, even though poll() signaled that the fd is readable, xcb would
2965    not read from the file descriptor.
2966    
2967    Fix this by checking if any bits from POLLIN are set in the result of poll(),
2968    instead of all of them.
2969    
2970    (This change was independently done by seanb@qnx.com as well)
2971    
2972    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=38001
2973    Acked-by: Julien Cristau <jcristau@debian.org>
2974    Signed-off-by: Uli Schlachter <psychon@znc.in>
2975
2976commit ac47e0ecdb46aa91b191a59364437a8f65947467
2977Author: Uli Schlachter <psychon@znc.in>
2978Date:   Sun Sep 8 22:16:39 2013 +0200
2979
2980    Fix documentation of xcb_poll_for_event()
2981    
2982    In commit 8eba8690adac2, the API documentation for xcb_poll_for_event() was
2983    fixed to remove an argument that was previously removed in commit 34168ab549.
2984    
2985    However, that commit only removed the first line of the documentation, leaving
2986    behind a spurious half-sentence. That commit happened seven years ago and now
2987    finally someone noticed...
2988    
2989    Thanks to Benjamin Herr for reporting this on IRC.
2990    
2991    v2: Thanks again to Benjamin Herr for noticing that my commit message blamed the
2992    wrong commit.
2993    
2994    Signed-off-by: Uli Schlachter <psychon@znc.in>
2995
2996commit c52f2891b43ae77008f63700982f800371458f4d
2997Author: Daniel Martin <consume.noise@gmail.com>
2998Date:   Sun Aug 11 13:25:18 2013 +0200
2999
3000    tests: Add files to .gitignore
3001    
3002    Add check_all.log, check_all.trs and test-suite.log to tests/.gitignore.
3003    
3004    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3005
3006commit a8d11c36edf5c49b718664dd7206f36be150f694
3007Author: Daniel Martin <consume.noise@gmail.com>
3008Date:   Thu Jul 25 11:09:26 2013 +0200
3009
3010    Sort gitignore, adjust pattern for config.h
3011    
3012    Don't ignore the files config.h and config.h.in, adjust the pattern to
3013    ignore config.h*. This matches an additional config.h.in~ too.
3014    
3015    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3016    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3017    Reviewed-by: Uli Schlachter <psychon@znc.in>
3018    Tested-By: Ran Benita <ran234@gmail.com>
3019
3020commit cbe54c97b3f0e4d40e0ee18796f8077cb4a6c16a
3021Author: Daniel Martin <consume.noise@gmail.com>
3022Date:   Thu Jul 25 10:56:30 2013 +0200
3023
3024    Use m4 directory
3025    
3026    - Follow the suggestion by libtoolize:
3027       "Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
3028        rerunning libtoolize, to keep the correct libtool macros in-tree.
3029        Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am."
3030      and add the macro and define.
3031    
3032    - Create the m4 directory and move acinclude.m4 as xcb.m4 there.
3033    
3034    - Ignore the m4 files libtoolize copies into the m4 directory
3035      (m4/l*.m4).
3036    
3037    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3038    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3039    Reviewed-by: Uli Schlachter <psychon@znc.in>
3040    Tested-By: Ran Benita <ran234@gmail.com>
3041
3042commit 6746ab1549d34a146c8383ed5acdabf0d48c1889
3043Author: Daniel Martin <consume.noise@gmail.com>
3044Date:   Wed Jul 24 14:13:41 2013 +0200
3045
3046    Use build-aux as autom4te cache directory
3047    
3048    Remove the generated directory ./autom4te.cache by reusing ./build-aux
3049    as cache directory.
3050    
3051    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3052    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3053    Reviewed-by: Uli Schlachter <psychon@znc.in>
3054    Tested-By: Ran Benita <ran234@gmail.com>
3055
3056commit b6d8c8fe61f25e4eb7b43d3d9b1de81a1f0638a3
3057Author: Daniel Martin <consume.noise@gmail.com>
3058Date:   Wed Jul 24 14:08:38 2013 +0200
3059
3060    Set AC_CONFIG_AUX_DIR to build-aux
3061    
3062    Do not clutter the project directory with generated/copied auxiliary
3063    files, save them in ./build-aux.
3064    
3065    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3066    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3067    Reviewed-by: Uli Schlachter <psychon@znc.in>
3068    Tested-By: Ran Benita <ran234@gmail.com>
3069
3070commit e3b34ad346efcaf0fd28b95e68d99388276802bc
3071Author: Daniel Martin <consume.noise@gmail.com>
3072Date:   Wed Jul 24 19:22:44 2013 +0200
3073
3074    Remove second AC_PREREQ, require version 2.60
3075    
3076    Remove a second AC_PREREQ and bump the required autoconf version to
3077    2.60.
3078    
3079    Version 2.59c was a testing release, published in April 2006. Version
3080    2.60 was the stable release afterwards, released in June 2006.
3081    
3082    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3083    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3084    Reviewed-by: Uli Schlachter <psychon@znc.in>
3085    Tested-By: Ran Benita <ran234@gmail.com>
3086
3087commit 50fb3a6312dd0b6b613fc886ffd6827952d1e286
3088Author: Daniel Martin <consume.noise@gmail.com>
3089Date:   Wed Jul 24 12:51:04 2013 +0200
3090
3091    Initialize automake earlier (bugfix for #66413)
3092    
3093    This fixes:
3094        https://bugs.freedesktop.org/show_bug.cgi?id=66413
3095        Bug 66413 - libxcb 1.9.1: Fails to build on Arch Linux: \
3096            /home/<user>/install-sh: No such file or directory
3097    
3098    Without that patch the search path for `install-sh` will become $HOME
3099    and the `install` target will fail, when DESTDIR doesn't exist in
3100    advance. (occured with automake 1.14 and autoconf 2.69)
3101    
3102    Initial patch by: Alain Kalker <a.c.kalker@gmail.com>
3103    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3104    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3105    Reviewed-by: Uli Schlachter <psychon@znc.in>
3106    Tested-By: Ran Benita <ran234@gmail.com>
3107
3108commit dd01db570c34dd3e2e0f5d07b8d40c837fa51057
3109Author: Daniel Martin <consume.noise@gmail.com>
3110Date:   Mon Dec 31 11:57:49 2012 +0100
3111
3112    Make xsltproc optional
3113    
3114    Fix Bug 23863 - xcb still checks for xsltproc:
3115        https://bugs.freedesktop.org/show_bug.cgi?id=23863
3116    
3117    xsltproc is used to generate the optional html page for `check` results,
3118    only. So, it's not a hard build dependency.
3119    
3120    Additionally, use yes/no instead of true/false in the HTML_CHECK_RESULT
3121    variable for consistent output after a configure run.
3122    
3123    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3124    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3125
3126commit 0289348f2c4ed3b1b286c51df19d82c6787c2b36
3127Author: Daniel Martin <consume.noise@gmail.com>
3128Date:   Fri Dec 28 23:25:16 2012 +0100
3129
3130    c_client.py: Do not create pointers in unions
3131    
3132    Do not create pointers in unions for fields of variadic length.
3133    
3134    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3135    Reviewed-by: Ran Benita <ran234@gmail.com>
3136
3137commit b9efd2a09a45616f6238e4da49b8f2127b6ec6d9
3138Author: Daniel Martin <consume.noise@gmail.com>
3139Date:   Wed Jan 9 12:52:15 2013 +0100
3140
3141    c_client.py: Always initialize xcb_align_to
3142    
3143    to get rid of:
3144        warning: 'xcb_align_to' may be used uninitialized in this function
3145    
3146    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3147    Reviewed-by: Peter Harris <pharris@opentext.com>
3148
3149commit 5648ddd2b97068f549268284129a438a6845e14c
3150Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3151Date:   Sat Aug 3 20:25:23 2013 -0700
3152
3153    Define _xcb_map_new with explicit void arg list instead of empty one
3154    
3155    Fixes Solaris Studio compiler warning:
3156    "xcb_list.c", line 50: warning: old style function definition
3157    
3158    and gcc warning:
3159    xcb_list.c: In function '_xcb_map_new':
3160    xcb_list.c:50:11: warning: old-style function definition [-Wold-style-definition]
3161    
3162    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3163    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3164
3165commit cb51f271b26c6abc76d415553f202bc5139273ca
3166Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3167Date:   Sat Aug 3 20:22:25 2013 -0700
3168
3169    Enable warnings for pre-C89 style definitions for gcc & Solaris Studio
3170    
3171    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3172    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3173
3174commit bc6a4f557ff4e497acdafdcebb006e5a7b4c5b11
3175Author: Michael Stapelberg <michael@stapelberg.de>
3176Date:   Mon Aug 5 22:14:18 2013 +0200
3177
3178    Build xcb-xkb by default
3179    
3180    There have not been any big issues with xcb-xkb recently.
3181    
3182    Also, Wayland is using xcb-xkb actively, making distributions compile
3183    libxcb with xkb support anyway, so let’s reflect reality :).
3184    
3185    Reviewed-by: Jamey Sharp <jamey@minilop.net>
3186    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3187
3188commit 45619dc71e9411a526d7c69595cf615b1b1206cf
3189Author: Daniel Martin <consume.noise@gmail.com>
3190Date:   Sat Jun 8 11:20:39 2013 +0200
3191
3192    c_client.py: Inject full_sequence into GE events
3193    
3194    The generic event structure xcb_ge_event_t has the full_sequence field
3195    at the 32byte boundary. That's why we've to inject this field into GE
3196    events while generating the structure for them. Otherwise we would read
3197    garbage (the internal full_sequence) when accessing normal event fields
3198    there.
3199    
3200    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3201    Reviewed-by: Keith Packard <keithp@keithp.com>
3202    Signed-off-by: Peter Harris <pharris@opentext.com>
3203
3204commit a1e67b141a57d39cbcaff1a703d6fc0da1fbb56d
3205Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3206Date:   Mon Jul 8 17:54:35 2013 -0700
3207
3208    Fix "indention" typos in xcb-examples.3 man page
3209    
3210    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3211
3212commit 8b6bb1a71977116d382f45eef803aedd3e313d37
3213Author: Marc Deslauriers <marc.deslauriers@canonical.com>
3214Date:   Wed Jun 5 16:38:00 2013 -0400
3215
3216    Update Makefile.am for newer automake
3217    
3218    Debian Bug #710344
3219    
3220    Reviewed-by: Daniel Martin <consume.noise@gmail.com>
3221
3222commit 9ae84ad187e2ba440c40f44b8eb21c82c2fdbf12
3223Author: Christian König <christian.koenig@amd.com>
3224Date:   Wed May 15 11:21:36 2013 +0200
3225
3226    fix deadlock with xcb_take_socket/return_socket v3
3227    
3228    To prevent different threads from stealing the socket from each other the
3229    caller of "xcb_take_socket" must hold a lock that is also acquired in
3230    "return_socket". Unfortunately xcb tries to prevent calling return_socket
3231    from multiple threads and this can lead to a deadlock situation.
3232    
3233    A simple example:
3234    - X11 has taken the socket
3235    - Thread A has locked the display.
3236    - Thread B does xcb_no_operation() and thus ends up in libX11's return_socket(),
3237      waiting for the display lock.
3238    - Thread A calls e.g. xcb_no_operation(), too, ends up in return_socket() and
3239      because socket_moving == 1, ends up waiting for thread B
3240    => Deadlock
3241    
3242    This patch allows calling return_socket from different threads at the same time
3243    an so resolves the deadlock situation.
3244    
3245    Partially fixes: https://bugs.freedesktop.org/show_bug.cgi?id=20708
3246    
3247    v2: fixes additional pthread_cond_wait dependencies,
3248        rework comments and patch description
3249    
3250    v3: separate pthread_cond_wait dependencies and unrelated whitespace
3251        change into their own patch, use unsigned for socket_seq
3252    
3253    Signed-off-by: Christian König <christian.koenig@amd.com>
3254    Signed-off-by: Uli Schlachter <psychon@znc.in>
3255
3256commit 1b33867fa996034deb50819ae54640be501f8d20
3257Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3258Date:   Wed May 1 17:59:31 2013 -0700
3259
3260    integer overflow in read_packet() [CVE-2013-2064]
3261    
3262    Ensure that when calculating the size of the incoming response from the
3263    Xserver, we don't overflow the integer used in the calculations when we
3264    multiply the int32_t length by 4 and add it to the default response size.
3265    
3266    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3267
3268commit e602b653c191e18cbb63db6526aac77c368ed70b
3269Author: Daniel Martin <consume.noise@gmail.com>
3270Date:   Mon May 13 23:33:04 2013 +0200
3271
3272    c_client.py: Handle multiple expr. in a bitcase
3273    
3274    Adopt a change from xcbgen. With that modification the expression in a
3275    bitcase became a list of expressions to support multiple <enumref> in a
3276    <bitcase>.
3277    
3278    Signed-off-by: Daniel Martin <consume.noise@gmail.com>
3279    Signed-off-by: Peter Harris <pharris@opentext.com>
3280
3281commit 6b6044cb8aacdf1b637da7b5dda392f9ff41ed39
3282Author: Christian König <christian.koenig@amd.com>
3283Date:   Wed May 15 11:21:35 2013 +0200
3284
3285    whitespace fix in xcb_take_socket
3286    
3287    Signed-off-by: Christian König <christian.koenig@amd.com>
3288    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3289
3290commit 0dd8f8d26a758bc385e79d9239bf6ef2e3d7bf13
3291Author: Colin Walters <walters@verbum.org>
3292Date:   Wed Jan 4 17:37:06 2012 -0500
3293
3294    autogen.sh: Implement GNOME Build API
3295    
3296    http://people.gnome.org/~walters/docs/build-api.txt
3297    
3298    Signed-off-by: Adam Jackson <ajax@redhat.com>
3299
3300commit 76a2166de9c80b35f987fdc3f3a228bafa0de94e
3301Author: Chí-Thanh Christopher Nguyễn <chithanh@gentoo.org>
3302Date:   Thu Oct 11 01:14:12 2012 +0200
3303
3304    c_client.py: Fix python-3 invalid except statement
3305    
3306    Replace except statement with a PEP-3110 compliant one. This fixes a regression
3307    introduced by c3deeaf714630531d693a6a902b8dabf791858b1
3308    https://bugs.freedesktop.org/show_bug.cgi?id=55690
3309    
3310    Reviewed-by: Peter Harris <pharris@opentext.com>
3311    Signed-off-by: Uli Schlachter <psychon@znc.in>
3312
3313commit 9db4517c87f56bb0ac82b647a08db30850ee2b04
3314Author: Chí-Thanh Christopher Nguyễn <chithanh@gentoo.org>
3315Date:   Thu Oct 11 01:14:11 2012 +0200
3316
3317    c-client.py: Fix python-3 AttributeError: 'dict' object has no attribute 'iteritems'
3318    
3319    This fixes a regression introduced by ea71d7d7e3f5d8189b80747678e9ca9a417b1d37
3320    https://bugs.freedesktop.org/show_bug.cgi?id=55690
3321    
3322    Reviewed-by: Peter Harris <pharris@opentext.com>
3323    Signed-off-by: Uli Schlachter <psychon@znc.in>
3324
3325commit 4ffe54f69049e6792a35a287fd9ff83abbd4fd8d
3326Author: Uli Schlachter <psychon@znc.in>
3327Date:   Fri Oct 5 14:53:37 2012 +0200
3328
3329    Release libxcb 1.9
3330    
3331    Signed-off-by: Uli Schlachter <psychon@znc.in>
3332
3333commit 4f52f884f42b72087f3323f2bab204223664a488
3334Author: Uli Schlachter <psychon@znc.in>
3335Date:   Fri Oct 5 11:23:26 2012 +0200
3336
3337    Include static man pages in "make dist"
3338    
3339    This was found by distcheck. It tried to install src/man/xcb-examples.3 and
3340    src/man/xcb-requests.3, but those files weren't in the distribution.
3341    
3342    Fix this by explicitly telling automake to distribute those files.
3343    
3344    Signed-off-by: Uli Schlachter <psychon@znc.in>
3345
3346commit 23911a707b8845bff52cd7853fc5d59fb0823cef
3347Author: Uli Schlachter <psychon@znc.in>
3348Date:   Mon Sep 24 22:07:51 2012 +0200
3349
3350    Fix a multi-thread deadlock
3351    
3352    This fixes a deadlock which was seen in-the-wild with wine.
3353    
3354    It could happen that two threads tried to read from the socket at the same time
3355    and one of the thread got stuck inside of poll()/select().
3356    
3357    The fix works by making sure that the writing thread doesn't steal the reading
3358    thread's reply.
3359    
3360    Debugged-by: Erich Hoover <ehoover@mines.edu>
3361    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54671
3362    Signed-off-by: Uli Schlachter <psychon@znc.in>
3363
3364commit c16cc5467eb0af7c5cdee16e6efaee54eb53bba6
3365Author: Uli Schlachter <psychon@znc.in>
3366Date:   Mon Sep 24 22:07:30 2012 +0200
3367
3368    Add a .gitignore for src/man/
3369    
3370    Signed-off-by: Uli Schlachter <psychon@znc.in>
3371
3372commit 08cc068ead7b8e678cdb119b38ada5261d5cc3ea
3373Author: Peter Harris <pharris@opentext.com>
3374Date:   Thu Aug 16 11:59:14 2012 -0400
3375
3376    Allow xcb_send_request with >MAX_IOV iovecs
3377    
3378    This allows an application to do a scatter/gather operation on a large
3379    image buffer to avoid the extra memcpy.
3380    
3381    Use autoconf to use UIO_MAXIOV where IOV_MAX is not available (and the
3382    POSIX minimum of 16 where neither are available).
3383    
3384    Reviewed-by: Uli Schlachter <psychon@znc.in>
3385    Signed-off-by: Peter Harris <pharris@opentext.com>
3386
3387commit ff53285ae3f604e9f2cc5f4837255220459b5e44
3388Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3389Date:   Sat Aug 25 13:53:37 2012 -0700
3390
3391    Return connection failure if display string specifies non-existent screen
3392    
3393    Matches the behaviour of Xlib - if you set DISPLAY to :0.1 but only have
3394    one screen, closes connection and returns error.
3395    
3396    This introduces a new connection error code:
3397    XCB_CONN_CLOSED_INVALID_SCREEN
3398    
3399    Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
3400    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3401    Reviewed-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
3402    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3403
3404commit 90889794ad882a6847bcffe52c4cc5dfd168f1f4
3405Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3406Date:   Fri Aug 24 23:35:41 2012 -0700
3407
3408    Add AC_USE_SYSTEM_EXTENSIONS to allow use of more system functionality
3409    
3410    Copied from libX11 configure.ac
3411    
3412    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3413
3414commit b52790e8ed4bb077eabdeca803935d2910558acc
3415Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3416Date:   Fri Aug 24 23:32:32 2012 -0700
3417
3418    Always include "config.h" at the start of all C source files.
3419    
3420    Allows configure to set defines such as _POSIX_SOURCE in config.h
3421    that affect functions exposed by system headers and get consistent
3422    results across all the source files.
3423    
3424    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3425
3426commit ed93a6a2a8e23f12380709de6e04b2d833df7e71
3427Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3428Date:   Sat Aug 25 12:33:35 2012 -0700
3429
3430    Fix "sppported" typo in doxygen comment for XCB_CONN_CLOSED_EXT_NOTSUPPORTED
3431    
3432    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3433
3434commit c3deeaf714630531d693a6a902b8dabf791858b1
3435Author: Colin Walters <walters@verbum.org>
3436Date:   Mon Aug 13 15:32:31 2012 -0400
3437
3438    c_client: Fix parallel-make issue creating 'man' directory
3439    
3440    With make -j, it was possible to hit a race condition in the code to
3441    make the 'man' directory.
3442    
3443    Signed-off-by: Julien Danjou <julien@danjou.info>
3444
3445commit 5f8f2ba1c4f9ac74c8f301dcca8566e296e37995
3446Author: Jeremy Huddleston <jeremyhu@apple.com>
3447Date:   Sat Apr 21 22:42:51 2012 -0700
3448
3449    xcb_connect: launchd: Don't fall back on tcp if $DISPLAY is a path to a launchd socket
3450    
3451    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
3452
3453commit 71a295082e07ff20d4c4cc97feed03b94cceb251
3454Author: Julien Danjou <julien@danjou.info>
3455Date:   Tue Mar 27 12:10:59 2012 +0200
3456
3457    Move static man to man
3458    
3459    Signed-off-by: Julien Danjou <julien@danjou.info>
3460
3461commit df217bf7c930d4433c991d86f857ecf63cc2d25a
3462Author: Julien Danjou <julien@danjou.info>
3463Date:   Tue Mar 27 12:10:15 2012 +0200
3464
3465    Do not list manpages, use a wildcard
3466    
3467    Signed-off-by: Julien Danjou <julien@danjou.info>
3468
3469commit a4417b1611f0bf02b81b54be90ef3353010da10a
3470Author: Julien Danjou <julien@danjou.info>
3471Date:   Tue Mar 27 12:06:54 2012 +0200
3472
3473    Split manpage list in two (static/built)
3474    
3475    Signed-off-by: Julien Danjou <julien@danjou.info>
3476
3477commit cc7fab2d5e912629d4a2a6adfb7666dc2ba45db2
3478Author: Julien Danjou <julien@danjou.info>
3479Date:   Mon Mar 26 18:29:35 2012 +0200
3480
3481    Allow undocumented code to be built
3482    
3483    Signed-off-by: Julien Danjou <julien@danjou.info>
3484
3485commit ea71d7d7e3f5d8189b80747678e9ca9a417b1d37
3486Author: Michael Stapelberg <michael@stapelberg.de>
3487Date:   Sun Nov 27 10:38:26 2011 +0000
3488
3489    c_client.py: generate manpages
3490    
3491    Signed-off-by: Julien Danjou <julien@danjou.info>
3492
3493commit 57a62e99b1241d5a0e6cf0b72f52090862a9c07d
3494Author: Julien Danjou <julien@danjou.info>
3495Date:   Fri Mar 9 15:38:56 2012 +0100
3496
3497    Release libxcb 1.8.1
3498    
3499    Signed-off-by: Julien Danjou <julien@danjou.info>
3500
3501commit 236f914ea7205f5f74e87fcc1b06d87bd0789a7a
3502Author: Uli Schlachter <psychon@znc.in>
3503Date:   Thu Mar 1 20:26:39 2012 +0100
3504
3505    Fix a busy loop on BSD and Mac OS
3506    
3507    On FreeBSD MSG_WAITALL on a non-blocking socket fails immediately if less bytes
3508    than were asked for are available. This is different than the behavior on linux
3509    where as many bytes as are available are returned in this case. Other OS
3510    apparently follow the FreeBSD behavior.
3511    
3512    _xcb_in_read() is used to fill xcb's read buffer, thus this function will call
3513    recv() with a big length argument (xcb's read buffer is by default 16 KiB
3514    large). That many bytes are highly unlikely to be available in the kernel
3515    buffer.
3516    
3517    This means that _xcb_in_read() always failed on FreeBSD. Since the socket was
3518    still signaled as readable by poll(), this bug even resulted in a busy loop.
3519    
3520    The same issue is present in read_block(), but here it is slightly different.
3521    read_block() is called when we read the first few bytes of an event or a reply,
3522    so that we already know its length. This means that we should be able to use
3523    MSG_WAITALL here, because we know how many bytes there have to be.
3524    
3525    However, that function could busy loop, too, when only the first few bytes of
3526    the packet were sent while the rest is stuck somewhere on the way to us. Thus,
3527    MSG_WAITALL should be removed here, too.
3528    
3529    Thanks to Christoph Egger from Debian for noticing the problem, doing all the
3530    necessary debugging and figuring out what the problem was! This patch is 99%
3531    from debian. Thanks for all the work.
3532    
3533    This bug was introduced in commit 2dcf8b025be88a25d4333abdc28d425b88238d96.
3534    
3535    This commit also reverts commit 9061ee45b8dbe5431c23e3f628089d703ccad0b1.
3536    
3537    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=45776
3538    
3539    Signed-off-by: Uli Schlachter <psychon@znc.in>
3540    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3541
3542commit 9061ee45b8dbe5431c23e3f628089d703ccad0b1
3543Author: Jeremy Huddleston <jeremyhu@apple.com>
3544Date:   Thu Mar 8 00:50:13 2012 -0800
3545
3546    darwin: Use read(2) rather than recv(2)
3547    
3548    2dcf8b025be88a25d4333abdc28d425b88238d96 was causing some regressions on
3549    darwin, so go back to using read(2) there until I have time to investigate
3550    further.
3551    
3552    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
3553
3554commit 7d235c62f0d5bd0df1236cc52141c10c5d272a18
3555Author: Julien Cristau <jcristau@debian.org>
3556Date:   Sun Feb 19 13:43:54 2012 +0100
3557
3558    Fallback to TCP if no protocol is specified and the UNIX connection fails
3559    
3560    Signed-off-by: Julien Cristau <jcristau@debian.org>
3561    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3562
3563commit f7bd139616d228b20eeb2c96b85e4093606c93fc
3564Author: Arnaud Fontaine <arnau@debian.org>
3565Date:   Wed Jan 25 17:18:20 2012 +0900
3566
3567    Add xkb_internals and xkb_issues to EXTRA_DIST.
3568
3569commit 68d30adde982f1be33a934707fa105c0db6f7f8f
3570Author: Jon TURNEY <jon.turney@dronecode.org.uk>
3571Date:   Tue Jan 17 20:59:56 2012 +0000
3572
3573    Update use of error_connection under WIN32 to _xcb_conn_ret_error()
3574    
3575    Unfortunately, commit 31b57676 adding WSACleanup/WSAShutdown on Win32 adds a new use
3576    of error_connection, which was removed in commit 769acff0, applied 5 minutes earlier.
3577    
3578    src/xcb_util.c: In function 'xcb_connect_to_display_with_auth_info':
3579    src/xcb_util.c:433:39: error: 'error_connection' undeclared (first use in this function)
3580    
3581    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
3582    Reviewed-by: Arvind Umrao <arvind.umrao@oracle.com>
3583    Signed-off-by: Uli Schlachter <psychon@znc.in>
3584
3585commit 87b7bf875e0105924ae306e90ca79512d6c0cf47
3586Author: Jon TURNEY <jon.turney@dronecode.org.uk>
3587Date:   Wed Jan 18 14:11:07 2012 +0000
3588
3589    Fix build of xcb_auth.c with XDMCP on WIN32
3590    
3591    Fix a redefinition problem which shows up when building for _WIN32 and
3592    libXdmcp is installed, so HASXDMAUTH is enabled
3593    
3594    It seems this is a special place in xcb as it uses other X11 library headers here
3595    
3596    If HASXDMAUTH is defined, include the wrapped windows.h before any header which
3597    includes it unwrapped, to avoid conflicts with types defined in X headers
3598    
3599    We need to include config.h and check HASXDMAUTH to avoid an unconditional dependency
3600    on x11proto headers
3601    
3602    In file included from install/include/X11/Xdmcp.h:19:0,
3603                     from git/xcb/libxcb/src/xcb_auth.c:52:
3604    install/include/X11/Xmd.h:120:14: error: conflicting types for 'INT32'
3605    /usr/i686-pc-mingw32/sys-root/mingw/include/basetsd.h:54:13: note: previous declaration of 'INT32' was here
3606    install/include/X11/Xmd.h:143:15: error: conflicting types for 'BOOL'
3607    /usr/i686-pc-mingw32/sys-root/mingw/include/windef.h:234:17: note: previous declaration of 'BOOL' was here
3608    
3609    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
3610    Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
3611
3612commit 6db1a2686f0f073438d36f5fa0f97b787842b0f2
3613Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
3614Date:   Tue Jan 17 23:55:23 2012 -0800
3615
3616    Revert "Fix include order with Xdmcp on WIN32"
3617    
3618    This reverts commit 0e9246def562be97cc816f824f702233a826ec56.
3619    
3620    This change caused build failures because <X11/Xdmcp.h> was never
3621    included under any circumstance.  This is because the check for
3622    HASXDMAUTH was moved before the inclusion of config.h (via xcbint.h)
3623    which defined it.
3624    
3625    Found-by: Tinderbox
3626    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
3627    Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
3628
3629commit da1d15082baab844a3b9b2d5cc48620af0b806ec
3630Author: Julien Danjou <julien@danjou.info>
3631Date:   Thu Jan 12 10:25:07 2012 +0100
3632
3633    Bump xcb-proto requirement
3634    
3635    We are now unable to build xcb-proto before 1.7.
3636    
3637    Signed-off-by: Julien Danjou <julien@danjou.info>
3638
3639commit b95b33e8c04c90c2240df19acea0c841d6e3450b
3640Author: Julien Danjou <julien@danjou.info>
3641Date:   Wed Jan 11 18:27:38 2012 +0100
3642
3643    Release libxcb 1.8
3644    
3645    Signed-off-by: Julien Danjou <julien@danjou.info>
3646
3647commit 31b57676e8d7ab6048dbfb145187833fac5e478c
3648Author: Ryan Pavlik <rpavlik@iastate.edu>
3649Date:   Wed Jan 11 18:06:50 2012 +0100
3650
3651    Use WSAStartup()/WSACleanup() on WIN32
3652    
3653    The alternative is to use these in every WIN32 application which uses xcb. Doing
3654    it this way should be safe, as, according to MSDN, "There must be a call to
3655    WSACleanup for each successful call to WSAStartup. Only the final WSACleanup
3656    function call performs the actual cleanup. The preceding calls simply decrement
3657    an internal reference count"
3658    
3659    (We should probably also include ws2_32 in Libs.private for libxcb, as anything
3660    which links with libxcb will also need that, but there seems to be some pkg-config
3661    issues to resolve first...)
3662    
3663    v2: Check for errors so WSAStartup()/WSACleanup() uses are balanced
3664    v3: Use same indentation style as surrounding code
3665    
3666    Reviewed-by: Peter Harris <pharris@opentext.com>
3667    Signed-off-by: Julien Danjou <julien@danjou.info>
3668
3669commit 0e9246def562be97cc816f824f702233a826ec56
3670Author: Ryan Pavlik <rpavlik@iastate.edu>
3671Date:   Thu Jan 5 20:57:53 2012 +0000
3672
3673    Fix include order with Xdmcp on WIN32
3674    
3675    Fix a redefinition problem due to include order which shows up when
3676    building for _WIN32 and libXdmcp is installed, so HASXDMAUTH is enabled
3677    
3678    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
3679    Reviewed-by: Peter Harris <pharris@opentext.com>
3680    Signed-off-by: Julien Danjou <julien@danjou.info>
3681
3682commit 4aa7a2c849a9536febb2dc7773e06c12a69c5213
3683Author: Jon TURNEY <jon.turney@dronecode.org.uk>
3684Date:   Thu Jan 5 20:57:52 2012 +0000
3685
3686    Fix WIN32 compilation after commit 163c47bdc0d32785d831e4c93fea9ab7e023446b
3687    
3688    WIN32 does not have arpa/inet.h, so do not try to include it unless _WIN32 is
3689    not defined
3690    
3691    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
3692    Reviewed-by: Peter Harris <pharris@opentext.com>
3693    Signed-off-by: Julien Danjou <julien@danjou.info>
3694
3695commit 769acff0da8b9859ebdf052dce80045465c7598c
3696Author: Arvind Umrao <arvind.umrao@oracle.com>
3697Date:   Fri Nov 4 15:42:05 2011 +0530
3698
3699    Added more error states and removed global error_connection
3700    
3701    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=41443
3702    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=42304
3703    
3704    I have added more xcb connection error states at xcb.h header.
3705    Also I have removed global error_connection variable, and added
3706    an interface that returns connection error state.
3707    
3708    TBD:
3709    I will segregate errors states in a separate header file and try to
3710    provide more precise error states, in future. Also I will give patch
3711    for libX11, in that patch xcb_connection_t::has_error will be passed
3712    to default io handler of libX11. This value can then be used for
3713    displaying error messages.
3714    
3715    Reviewed-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
3716    Reviewed-by: Uli Schlachter <psychon@znc.in>
3717    Signed-off-by: Arvind Umrao <arvind.umrao@oracle.com>
3718
3719commit 9b4d6f30a339e2d18ebaea98149da81aba453216
3720Author: Keith Packard <keithp@keithp.com>
3721Date:   Thu Dec 1 10:28:51 2011 +0000
3722
3723    Make xcb_take_socket keep flushing until idle
3724    
3725    _xcb_out_flush_to will drop the iolock in pthread_cond_wait allowing
3726    other threads to queue new requests. When this happened,
3727    there would be requests queued for the socket after _xcb_out_flush_to
3728    returned, and xcb_take_socket would throw an assert.
3729    
3730    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29875
3731    Signed-off-by: Keith Packard <keithp@keithp.com>
3732    Signed-off-by: Julien Danjou <julien@danjou.info>
3733
3734commit 5ceeaaa4294201b3f613c07f9ec610c0e5f673c7
3735Author: Uli Schlachter <psychon@znc.in>
3736Date:   Thu Aug 25 14:18:16 2011 +0200
3737
3738    Fix a dead-lock due to xcb_poll_for_reply
3739    
3740    Imagine two threads:
3741    
3742    Thread#1: for(;;) { xcb_get_input_focus_reply(c, xcb_get_input_focus(c), 0); }
3743    
3744    Thread#2: for(;;) { xcb_poll_for_event(c); }
3745    
3746    Since xcb_poll_for_event() calls _xcb_in_read() directly without synchronizing
3747    with any other readers, this causes two threads to end up calling recv() at the
3748    same time. We now have a race because any of these two threads could get read
3749    the GetInputFocus reply.
3750    
3751    If thread#2 reads this reply, it will be put in the appropriate queue and
3752    thread#1 will still be stuck in recv(), although its reply was already received.
3753    If no other reply or event causes this thread to wake up, the process deadlocks.
3754    
3755    To fix this, we have to make sure that there is only ever one thread reading
3756    from the connection. The obvious solution is to check in poll_for_next_event()
3757    if another thread is already reading (in which case c->in.reading != 0) and not
3758    to read from the wire in this case.
3759    
3760    This solution is actually correct if we assume that the other thread is blocked
3761    in poll() which means there isn't any data which can be read. Since we already
3762    checked that there is no event in the queue this means that
3763    poll_for_next_event() didn't find any event to return.
3764    
3765    There might be a small race here where the other thread already determined that
3766    there is data to read, but it still has to wait for c->iolock. However, this
3767    means that the next poll_for_next_event() will be able to read the event, so
3768    this shouldn't cause any problems.
3769    
3770    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=40372
3771    
3772    Signed-off-by: Uli Schlachter <psychon@znc.in>
3773    Signed-off-by: Peter Harris <pharris@opentext.com>
3774
3775commit b12038e9ae5343c4176f11d68c963c752bc35c03
3776Author: Jamey Sharp <jamey@minilop.net>
3777Date:   Wed Aug 24 08:52:02 2011 -0700
3778
3779    Keep ALIGNOF definition out of the public namespace.
3780    
3781    Uli's patch is an excellent solution; I just want to keep the new
3782    ALIGNOF macro hidden from XCB's users, as they don't need it to call
3783    XCB.
3784    
3785    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3786
3787commit 163c47bdc0d32785d831e4c93fea9ab7e023446b
3788Author: Markus Duft <mduft@gentoo.org>
3789Date:   Wed Aug 24 10:49:06 2011 -0400
3790
3791    Support pre-IPv6 systems (without getaddrinfo)
3792    
3793    Some of these systems (eg. Interix on XP) are still in use.
3794    
3795    Reviewed-by: Josh Triplett <josh@joshtriplett.org>
3796    Signed-off-by: Peter Harris <pharris@opentext.com>
3797
3798commit aa02096b8e7f94ad3c998a8d5af54963ee860b13
3799Author: Uli Schlachter <psychon@znc.in>
3800Date:   Wed Aug 24 12:47:16 2011 +0200
3801
3802    Compute alignment correctly
3803    
3804    The code previously assumed that everything has to be aligned to a 4 byte
3805    boundary. This assumption is wrong as e.g. the STR struct from xproto shows.
3806    
3807    Instead, each type has to be aligned to its natural alignment. So a char doesn't
3808    need any alignment, a INT16 gets aligned to a 2-byte-boundary and a INT32 gets
3809    the old 4 byte alignment.
3810    
3811    I'm not 100% sure that this commit is correct, but some quick tests with awesome
3812    and cairo-xcb went well.
3813    
3814    This commit causes lots of dead assignments to xcb_align_to since only the last
3815    field's alignment is actually used, but this simplified this patch a lot.
3816    
3817    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=34037
3818    
3819    Signed-off-by: Uli Schlachter <psychon@znc.in>
3820    Signed-off-by: Peter Harris <pharris@opentext.com>
3821
3822commit 4f25ee16443b29e1a25bd26a724e1e0a577e21ff
3823Author: Michael Stapelberg <michael@stapelberg.de>
3824Date:   Thu Aug 18 21:38:28 2011 +0200
3825
3826    Drop AI_ADDRCONFIG when resolving TCP addresses
3827    
3828    When a system is completely offline (no interface has an IP address but 'lo'),
3829    xcb could not connect to localhost via TCP, e.g. connections with
3830    DISPLAY=127.0.0.1:0 fail.
3831    
3832    AI_ADDRCONFIG will only return IPv4 addresses if the system has an IPv4
3833    address configured (likewise for IPv6). This also takes place when
3834    resolving localhost (or 127.0.0.0/8 or ::1). Also, as per RFC 3493,
3835    loopback addresses are not considered as valid addresses when
3836    determining whether to return IPv4 or IPv6 addresses.
3837    
3838    As per mailing-list discussion on the xcb list started with message
3839    20110813215405.5818a0c1@x200, the AI_ADDRCONFIG flag is there for historical
3840    reasons:
3841    
3842        In the old days, the "default on-link" assumption in IPv6 made the flag vey
3843        much indispensable for dual-stack hosts on IPv4-only networks. Without it,
3844        there would be long timeouts trying non-existent IPv6 connectivity. Nowadays,
3845        this assumption has been flagged as historic bad practice by IETF, and hosts
3846        should have been updated to not make it anymore.
3847    
3848        Then AI_ADDRCONFIG became mostly cosmetic: it avoids phony "Protocol family
3849        not supported" or "Host unreachable" errors while trying to connect to a dual-
3850        stack mode from a host with no support for source address selection.
3851    
3852        Nowadays, on up-to-date systems, this flag is completely useless. Then again,
3853        I understood only the very latest MacOS release is "up-to-date" with this
3854        definition.
3855
3856commit 662ad589c5d6f03757ae57a926d3800bfb528b30
3857Author: James Jones <jajones@nvidia.com>
3858Date:   Wed May 11 23:22:22 2011 -0700
3859
3860    Insert, not append explicit xcbgen dir python path
3861    
3862    If a the path to the xcb python generate libs is
3863    explicitly specified to c_client.py, insert it in
3864    the python path list just after the local dir entry,
3865    rather than appending it to the existing paths.
3866    This keeps a global/distro install of xcb from
3867    overriding a local build of the xcb proto files.
3868    
3869    Signed-off-by: James Jones <jajones@nvidia.com>
3870    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3871
3872commit 294c9f455c3534d836b010dacd2e7aa62a7dde9d
3873Author: David Coles <dcoles@gaikai.com>
3874Date:   Fri Apr 8 17:47:05 2011 -0700
3875
3876    Add support for building with Python 3
3877    
3878    Python 3 introduces some language changes that cause issues when running
3879    c_client.py. This also breaks compatibility with Python 2.5 since it does not
3880    support the "as" statement in try/except blocks and does not have reduce() in
3881    the functools package.
3882    
3883    The main changes are:
3884    * try/except blocks require `except ... as ...:` to resolve syntactical ambiguity
3885    * map() and filter() return iterators rather than lists in Python 3
3886    * reduce() is now in functools package (and not built-in in Python 3)
3887    * Dictionaries don't have a has_key() method in Python 3
3888    * None and int types can't be directly compared in Python 3
3889    * print() is a statement in Python 3
3890    
3891    See http://diveintopython3.org/porting-code-to-python-3-with-2to3.html and
3892    PEP-3110 for details.
3893    
3894    Verified on Python 2.6.5 and 3.1.3.
3895    
3896    Signed-off-by: David Coles <dcoles@gaikai.com>
3897    Signed-off-by: Julien Danjou <julien@danjou.info>
3898
3899commit e300ee4920bf4618f58618f3063b362f811154c1
3900Author: Jamey Sharp <jamey@minilop.net>
3901Date:   Tue Apr 12 13:09:23 2011 -0700
3902
3903    Revert "Introduce xcb_wait_for_event_until, for consuming responses in wire-order."
3904    
3905    This function was intended to allow libX11 to fix a multi-threaded hang,
3906    but the corresponding libX11 patch caused single-threaded apps to spin
3907    sometimes. Since I've retracted that patch, this patch has no users and
3908    shouldn't go into a release unless/until that changes.
3909    
3910    This reverts commit 2415c11dec5e5adb0c17f98aa52fbb371a4f8f23.
3911    
3912    Conflicts:
3913    
3914            src/xcb.h
3915            src/xcb_in.c
3916    
3917    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3918
3919commit 527df3c84bd71113cedc7f55089c02d1c099ecad
3920Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
3921Date:   Tue Mar 22 14:33:23 2011 +0200
3922
3923    Introduce a variant of xcb_poll_for_event for examining event queue.
3924    
3925    In some circumstances using xcb_poll_for_event is suboptimal because
3926    it checks the connection for new events. This may lead to a lot of
3927    failed nonblocking read system calls.
3928    
3929    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
3930    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3931
3932commit b64cd0df884e7901ff13def0272df74962035920
3933Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3934Date:   Mon Apr 4 21:36:47 2011 -0700
3935
3936    If protocol is "unix", use a Unix domain socket, not TCP
3937    
3938    Fixes fallback to local connections from Xlib's XOpenDisplay(), which
3939    will try with protocol "unix" if a hostname is specified and tcp fails
3940    (as it usually will now that most OS'es ship with -nolisten tcp enabled).
3941    
3942    Also fixes explicitly specifying DISPLAY as "unix/foo:0", which Xlib
3943    previously accepted for Unix domain sockets.
3944    
3945    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3946
3947commit b027922ebf1931885e00629c20e26f14f184998d
3948Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3949Date:   Mon Apr 4 21:28:00 2011 -0700
3950
3951    Make launchd code in xcb_util.c match surrounding code indent levels
3952    
3953    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3954
3955commit 82b1f3919a82a730f6b2f952d4090fe15702694e
3956Author: Carlos Garnacho <carlosg@gnome.org>
3957Date:   Tue Oct 5 18:03:22 2010 +0200
3958
3959    Handle XGE events with the "send event" flag
3960    
3961    This patch is necessary so xcb reads the payload after the message
3962    for GenericEvents with the 0x80 flag turned on.
3963    
3964    Signed-off-by: Carlos Garnacho <carlosg@gnome.org>
3965    Signed-off-by: Jamey Sharp <jamey@minilop.net>
3966
3967commit 42c4adeff4a6aedfba30e22f71800c1b73942923
3968Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3969Date:   Mon Apr 4 20:20:16 2011 -0700
3970
3971    Add #include <sys/socket.h> to xcb_conn.c
3972    
3973    Solves compiler warning on Solaris:
3974    "xcb_conn.c", line 304: warning: implicit function declaration: shutdown
3975    
3976    Also provides system definition of SHUT_RDWR on Solaris 11.
3977    
3978    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3979    Reviewed-by: Jamey Sharp <jamey@minilop.net>
3980
3981commit 4b502dd696cf7f59a961bcf71c9255ae28f0765a
3982Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3983Date:   Mon Apr 4 20:12:56 2011 -0700
3984
3985    Remove unused DECnet code
3986    
3987    "unifdef -UDNETCONN src/xcb_util.c" plus re-indenting code that was
3988    formerly in the else clause after a DECnet check.
3989    
3990    DECnet support has been removed from most of the X.Org code base for
3991    several years, and it appears DNETCONN was never defined in XCB.
3992    
3993    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
3994    Reviewed-by: Jamey Sharp <jamey@minilop.net>
3995
3996commit 7131d5d0706f2b63caad13c815e893627872114c
3997Author: Alan Coopersmith <alan.coopersmith@oracle.com>
3998Date:   Mon Apr 4 16:32:45 2011 -0700
3999
4000    Use special path to sockets when running under Solaris Trusted Extensions
4001    
4002    Solaris Trusted Extensions puts the endpoints for the X server's Unix
4003    domain sockets in a special directory shared from the global zone to
4004    each of the labeled zones, since each labeled zone has a separate /tmp.
4005    
4006    Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
4007    Reviewed-by: Peter Harris <pharris@opentext.com>
4008
4009commit 70976d87f18d15c2ccc28eb7728e4822d3849e0d
4010Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
4011Date:   Wed Mar 23 17:47:50 2011 +0200
4012
4013    Prevent theoretical double free and leak on get_peer_sock_name.
4014    
4015    Variable new_sockname will leak and sockname will be double freed if
4016    both of the cases shown below are true.
4017    
4018    1. realloc succeeds and doesn't return the original pointer
4019    2. calling socket_func fails
4020    
4021    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4022    Signed-off-by: Erkki Seppälä <erkki.seppala@vincit.fi>
4023    Reviewed-by: Arnaud Fontaine <arnau@debian.org>
4024    Signed-off-by: Peter Harris <pharris@opentext.com>
4025
4026commit 3678159e4ed64502f9ce218a63c8d069649f2215
4027Author: Jamey Sharp <jamey@minilop.net>
4028Date:   Sat Mar 19 20:04:55 2011 -0700
4029
4030    Delete the old c-client.xsl.
4031    
4032    It hasn't been used since libxcb 1.1.90.1, released in 2008.
4033    
4034    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4035
4036commit 2415c11dec5e5adb0c17f98aa52fbb371a4f8f23
4037Author: Jamey Sharp <jamey@minilop.net>
4038Date:   Fri Mar 18 20:56:07 2011 -0700
4039
4040    Introduce xcb_wait_for_event_until, for consuming responses in wire-order.
4041    
4042    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4043    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4044
4045commit 29a974f212aae9eeff4fde99f110cee08f0312f3
4046Author: Jamey Sharp <jamey@minilop.net>
4047Date:   Fri Mar 18 17:36:32 2011 -0700
4048
4049    Dequeue readers that can't receive any new responses.
4050    
4051    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4052    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4053
4054commit 131e867fca5cda94e634af69214ad54e066ac871
4055Author: Jamey Sharp <jamey@minilop.net>
4056Date:   Fri Mar 18 15:37:34 2011 -0700
4057
4058    Factor reader_list management out of wait_for_reply.
4059    
4060    Later patches will insert reader_list entries from other entry points.
4061    
4062    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4063    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4064
4065commit 1469e879655b20351530059538a7b89612028ae2
4066Author: Jamey Sharp <jamey@minilop.net>
4067Date:   Fri Mar 18 18:18:41 2011 -0700
4068
4069    Enable AM_SILENT_RULES on automake 1.11 or newer.
4070    
4071    This incantation is supposed to be a no-op on earlier automake versions.
4072    
4073    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4074    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4075
4076commit 6310475e23eac6917db54f1425e20d8434bee679
4077Author: Rami Ylimäki <rami.ylimaki@vincit.fi>
4078Date:   Wed Oct 13 17:48:13 2010 +0300
4079
4080    Prevent reply waiters from being blocked.
4081    
4082    It's possible to call xcb_wait_for_reply more than once for a single
4083    request. In this case we are nice and let reply waiters continue so
4084    that they can notice that the reply is not available
4085    anymore. Otherwise an event waiter could just signal the reply waiter
4086    that got its reply to continue but leave a waiter for an earlier reply
4087    blocked.
4088    
4089    Below is an example sequence for reproducing this problem.
4090    
4091    thread #1 (XNextEvent)
4092      - waits for events
4093    thread #2 (XSync)
4094      - executes request #2
4095      - waits for reply #2
4096    thread #1
4097      - reads reply #2
4098      - signals waiter of reply #2 to continue
4099      - waits for events
4100    thread #2
4101      - handles reply #2
4102    thread #3 (XCloseDisplay)
4103      - executes request #3
4104      - waits for reply #2
4105    thread #1
4106      - reads reply #3
4107      - nobody is waiting for reply #3 so don't signal
4108      - wait for events
4109    
4110    Of course it may be questionable to wait for a reply twice, but XCB
4111    should be smart enough to let clients continue if they choose to do
4112    so.
4113    
4114    Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
4115    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4116
4117commit 29ab5aeb9b1b1daf7f0659b134a4cfe9f42ca71a
4118Author: Jamey Sharp <jamey@minilop.net>
4119Date:   Sun Mar 13 09:41:10 2011 -0700
4120
4121    Include XKB in ./configure's summary output.
4122    
4123    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4124
4125commit 2edfd5c375bf980b484b7cfbfc1f4fb751621859
4126Merge: ed37b08 8c3325f
4127Author: Jamey Sharp <jamey@minilop.net>
4128Date:   Sun Mar 13 09:18:24 2011 -0700
4129
4130    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
4131    
4132    Apparently I forgot to push these months ago.
4133
4134commit 8c3325f8bbdb1e8975bdb01525a52d6b0f80cfa3
4135Author: Jeremy Huddleston <jeremyhu@apple.com>
4136Date:   Fri Mar 4 12:41:55 2011 -0800
4137
4138    darwin: Don't use poll() when expected to run on darwin10 and prior
4139    
4140    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4141
4142commit c2e023646298da05e212056fbc6b329e4dd9a100
4143Author: Peter Harris <pharris@opentext.com>
4144Date:   Thu Dec 23 13:32:17 2010 -0500
4145
4146    Don't try to sizeof(void)
4147    
4148    sizeof(void) is a gcc extension, and not portable.
4149    
4150    Xorg Bugzilla 31959
4151    http://bugs.freedesktop.org/show_bug.cgi?id=31959
4152    http://lists.freedesktop.org/archives/xcb/2010-May/006039.html
4153    
4154    Signed-off-by: Peter Harris <pharris@opentext.com>
4155    Tested-by: Cyril Brulebois <kibi@debian.org>
4156
4157commit 9efced72a3cb8072fa60fbed4f04d61cde412494
4158Author: Vincent Torri <vincent dot torri at gmail dot com>
4159Date:   Sun Nov 28 14:02:40 2010 +0100
4160
4161    fix Windows build and installation
4162
4163commit 3c5813697169a33ecfd6ac0ab5641dec654f6612
4164Merge: 69b78ce b672d15
4165Author: Peter Harris <pharris@opentext.com>
4166Date:   Thu Dec 23 13:04:40 2010 -0500
4167
4168    Merge branch 'master' of git://anongit.freedesktop.org/~peterh/libxcb
4169
4170commit 69b78ced1a7bcdca538c0720fde9cf3e6f70d040
4171Author: Jeetu Golani <jeetu.golani@gmail.com>
4172Date:   Sun Dec 12 16:48:41 2010 -0500
4173
4174    Don't validate FD_SETSIZE on Win32
4175    
4176    Windows' file handles have never been small or consecutive, so Windows'
4177    select has always been implemented the same way as everyone else's poll.
4178    
4179    On Windows, FD_SETSIZE is the size of the poll array, not the maximum
4180    SOCKET number.
4181    
4182    Signed-off-by: Peter Harris <git@peter.is-a-geek.org>
4183
4184commit 8ecd754b168a0352783bf1ba0f0887f7ff479ee8
4185Author: Uli Schlachter <psychon@znc.in>
4186Date:   Wed Nov 17 20:26:11 2010 +0100
4187
4188    xcb_take_socket: Document sequence wrap requirements
4189    
4190    If lots of requests are send without one causing a reply, xcb can get confused
4191    about the current sequence number of a reply. Document the requirements of an
4192    external socket owner to avoid this problem.
4193    
4194    The return_socket callback for xcb_take_socket() originally was supposed to
4195    return the last sequence number used, but the version committed to libxcb never
4196    actually had this signature. This fixes the function's documentation not to
4197    mention this non-existent return value.
4198    
4199    Signed-off-by: Uli Schlachter <psychon@znc.in>
4200    Signed-off-by: Julien Danjou <julien@danjou.info>
4201
4202commit 5755582444ad0ba79e661ab3173cc38e9e588d83
4203Author: Nick Bowler <nbowler@draconx.ca>
4204Date:   Wed Nov 10 20:49:41 2010 -0500
4205
4206    xcb_auth: Fix memory leak in _xcb_get_auth_info.
4207    
4208    If the initial get_peer_sock_name(getpeername ...) succeeds, the
4209    pointer to allocated memory is overwritten by the later call to
4210    get_peer_sock_name(getsockname ...).  Fix that up by freeing
4211    the allocated memory before overwriting the pointer.
4212    
4213    Signed-off-by: Nick Bowler <nbowler@draconx.ca>
4214    Signed-off-by: Julien Danjou <julien@danjou.info>
4215
4216commit ed37b087519ecb9e74412e4df8f8a217ab6d12a9
4217Author: Jamey Sharp <jamey@minilop.net>
4218Date:   Sat Oct 9 17:13:45 2010 -0700
4219
4220    xcb_in: Use 64-bit sequence numbers internally everywhere.
4221    
4222    Widen sequence numbers on entry to those public APIs that still take
4223    32-bit sequence numbers.
4224    
4225    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4226
4227commit 6c8b539c2a2e53bf3deb0e749a941ab52b7e8834
4228Author: Jamey Sharp <jamey@minilop.net>
4229Date:   Sat Oct 9 13:19:05 2010 -0700
4230
4231    xcb_discard_reply: Simplify by re-using poll_for_reply helper.
4232    
4233    If you discard a sequence number that has multiple responses already
4234    read, this will do more allocations than necessary. But nobody cares
4235    about ListFontsWithInfo.
4236    
4237    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4238
4239commit 3a74b5e7a1aab0619b7e34d90d2b8b2b1e386129
4240Author: Jamey Sharp <jamey@minilop.net>
4241Date:   Sat Oct 9 12:32:05 2010 -0700
4242
4243    xcb_request_check: Hold the I/O lock while deciding to sync.
4244    
4245    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4246
4247commit ee1bc1d28a1bda0526db90139edc1304d2ef3d7c
4248Author: Jamey Sharp <jamey@minilop.net>
4249Date:   Sat Oct 9 04:08:18 2010 -0700
4250
4251    xcb_send_request: Send all requests using a common internal send_request.
4252    
4253    This simplifies the critical section of xcb_send_request and fixes a
4254    couple of subtle bugs:
4255    
4256    - It's possible for xcb_send_request to need to issue two sync requests
4257      before it can issue the real request. Previously, we counted sequence
4258      numbers as if both were issued, but only one went out on the wire.
4259    
4260    - The test for whether to sync at 32-bit sequence number wrap has been
4261      incorrect since we switched to 64-bit sequence numbers internally.
4262    
4263    This change means that if the output queue was already full and the
4264    current request is bigger than the output queue, XCB will do one more
4265    write syscall than it did before. But syncs are rare and small requests
4266    are the norm, so this shouldn't be a measurable difference.
4267    
4268    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4269
4270commit b672d1514c88e119f3aaeded8c8a488cad36db52
4271Author: Peter Harris <pharris@opentext.com>
4272Date:   Wed Sep 22 23:15:38 2010 -0400
4273
4274    Fix _unserialize of reply headers
4275    
4276    This cleans up a number of warnings, and passes the sequence number
4277    through correctly.
4278    
4279    Signed-off-by: Peter Harris <pharris@opentext.com>
4280
4281commit 29cca33b9001961fa2c33bb9d9fe4a9983913fce
4282Author: Peter Harris <pharris@opentext.com>
4283Date:   Wed Sep 22 22:32:34 2010 -0400
4284
4285    Clean up a couple of warnings in xprint
4286    
4287    Signed-off-by: Peter Harris <pharris@opentext.com>
4288
4289commit 8c1d2021ca611c1452a8d2ff2a705569e4ebd056
4290Author: Peter Harris <pharris@opentext.com>
4291Date:   Wed Sep 22 21:16:51 2010 -0400
4292
4293    Make *_unserialize safe to use on buffers in-place
4294    
4295    By calling memmove instead of memcpy, and walking the buffer backward
4296    from the end, *_unserialize is safe to use in-place.
4297    
4298    Signed-off-by: Peter Harris <pharris@opentext.com>
4299
4300commit 28a71c6567d08272dc9c1c2b32f0529f11f62b9e
4301Author: Peter Harris <pharris@opentext.com>
4302Date:   Fri Sep 10 15:51:56 2010 -0400
4303
4304    Fix memory leak in _sizeof implemented with _unserialize
4305    
4306    Signed-off-by: Peter Harris <pharris@opentext.com>
4307
4308commit a22909c0f513fe28347c56be65a345831c3ce744
4309Author: Peter Harris <pharris@opentext.com>
4310Date:   Wed Sep 8 15:57:00 2010 -0400
4311
4312    Don't emit out-of-module sizeof definitions
4313    
4314    Signed-off-by: Peter Harris <pharris@opentext.com>
4315
4316commit 1c4717de367fe3bf1cf56bd8ef2bd30586bed023
4317Author: Josh Triplett <josh@joshtriplett.org>
4318Date:   Sun Sep 19 20:38:06 2010 +0200
4319
4320    Allow disconnecting connections that are in error state.
4321    
4322    In support of this, consolidate the two static error_connection
4323    definitions into one so we don't try to free the static out-of-memory
4324    error_connection.
4325    
4326    Commit by Josh Triplett and Jamey Sharp.
4327    
4328    Signed-off-by: Josh Triplett <josh@joshtriplett.org>
4329    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4330
4331commit 03bcccb132249142ba908a075e0bd5075fc20d97
4332Author: Peter Harris <pharris@opentext.com>
4333Date:   Fri Sep 10 15:53:13 2010 -0400
4334
4335    Add xkb.* to gitignore
4336    
4337    Signed-off-by: Peter Harris <pharris@opentext.com>
4338
4339commit 28d39258008fcc8ced84dc6c1dd2644e2c908c87
4340Merge: f0565e8 5e8a7ad
4341Author: Peter Harris <pharris@opentext.com>
4342Date:   Wed Sep 8 14:41:52 2010 -0400
4343
4344    Merge branch 'gsoc2010' of git://anongit.freedesktop.org/~chr/libxcb
4345
4346commit f0565e8f06aadf760a9065a97b8cf5ab9cbd18de
4347Author: Jamey Sharp <jamey@minilop.net>
4348Date:   Sat Sep 4 10:17:21 2010 -0700
4349
4350    _xcb_conn_wait: Shut down the connection on unexpected poll() events.
4351    
4352    If a client calls close(2) on the connection's file descriptor and then
4353    flushes writes, libxcb causes a hang in the client.
4354    
4355    Any flush eventually calls _xcb_out_send() with has the following loop:
4356       while(ret && *count)
4357           ret = _xcb_conn_wait(c, &c->out.cond, vector, count);
4358    
4359    _xcb_conn_wait(), if built with USE_POLL, gets the POLLNVAL error. It only
4360    checks for POLLIN and POLLOUT though, ignoring the error. Return value is 1,
4361    count is unmodified, leaving us with an endless loop and a client hang.
4362    
4363    XTS testcase Xlib3/XConnectionNumber triggers this bug. It creates a display
4364    connection, closes its file descriptor, tries to send a no-op, and then expects
4365    an error.
4366    http://cgit.freedesktop.org/xorg/test/xts/tree/xts5/Xlib3/XConnectionNumber.m
4367    
4368    If poll returned POLLHUP or POLLERR, we might see the same result.
4369    
4370    If poll returns any event we didn't ask for, this patch causes
4371    _xcb_conn_shutdown() to be invoked and an error returned. This matches the
4372    behaviour if select(2) is used instead of poll(2): select(2) returns -1 and
4373    EBADF for an already closed file descriptor.
4374    
4375    I believe this fix both is safe and will handle any similar error. POSIX says
4376    that the only bits poll is permitted to set in revents are those bits that were
4377    set in events, plus POLLHUP, POLLERR, and POLLNVAL. So if we see any flags we
4378    didn't ask for then something has gone wrong.
4379    
4380    Patch inspired by earlier proposals from Peter Hutterer and Aaron
4381    Plattner--thanks!
4382    
4383    Reported-by: Peter Hutterer <peter.hutterer@who-t.net>
4384    Reported-by: Aaron Plattner <aplattner@nvidia.com>
4385    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4386    Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
4387    Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
4388    Tested-by: Aaron Plattner <aplattner@nvidia.com>
4389    Cc: Peter Hutterer <peter.hutterer@who-t.net>
4390    Cc: Dan Nicholson <dbn.lists@gmail.com>
4391    Signed-off-by: Peter Harris <pharris@opentext.com>
4392
4393commit 20da10490f8dac75ec9fe1df28cb9e862e171be5
4394Merge: 7f5cfcc 2dcf8b0
4395Author: Peter Harris <pharris@opentext.com>
4396Date:   Tue Aug 31 18:33:36 2010 -0400
4397
4398    Merge branch 'master' of git://github.com/topcat/xcb-win32
4399    
4400    Conflicts:
4401            src/xcb_conn.c
4402            src/xcb_util.c
4403    
4404    Signed-off-by: Peter Harris <pharris@opentext.com>
4405
4406commit 7f5cfcc2fd0168d505504cc088bfdcba5c71f0ea
4407Author: Aaron Plattner <aplattner@nvidia.com>
4408Date:   Tue Aug 17 08:04:41 2010 -0700
4409
4410    xcb_disconnect: call shutdown() to force a disconnect
4411    
4412    Fixes the X Test Suite's XCloseDisplay-6 test, which has this (admittedly
4413    ridiculous) behavior:
4414    
4415     1. Create a window w.
4416     2. Open two display connections, dpy1, and dpy2.
4417     3. Grab the server using dpy1.
4418     4. Fork.
4419     5 (child). XSetProperty on w using dpy2.
4420     5 (parent). Verify that no event was recieved on dpy1.
4421     6 (parent). XCloseDisplay(dpy1).
4422     6 (child). Verify that an event was received on dpy2.
4423    
4424    It was failing because at step 6 (child), the server had not actually ungrabbed
4425    yet because the file descriptor for dpy1 was still open in the child process.
4426    
4427    Shutting down the socket during XCloseDisplay matches the behavior of non-XCB
4428    Xlib, which calls shutdown() from _X11TransSocketDisconnect.
4429    
4430    Thanks to Julien Cristau for noticing this.
4431    
4432    Signed-off-by: Aaron Plattner <aplattner at nvidia.com>
4433    Reviewed-by: Julien Cristau <jcristau@debian.org>
4434    Signed-off-by: Peter Harris <pharris@opentext.com>
4435
4436commit 2040f10a4efa95092bc9409c5b20347989b5f0a2
4437Author: Jamey Sharp <jamey@minilop.net>
4438Date:   Tue Aug 24 09:17:38 2010 -0700
4439
4440    xcb_request_check: Sync even if an event was read for this sequence.
4441    
4442    This fixes the test case I have so far for Havoc's report that
4443    xcb_request_check hangs.
4444    
4445    Rationale: Since we have a void cookie, request_expected can't have been
4446    set equal to this sequence number when the request was sent; it can only
4447    have become equal due to the arrival of an event or error. If it became
4448    equal due to an event then we still need to sync. If it became equal due
4449    to an error, then request_completed will have been updated, which means
4450    we correctly won't sync.
4451    
4452    Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=29599
4453    
4454    However, Havoc reports that he can still reproduce the problem, so we
4455    may be revisiting this later.
4456    
4457    Reported-by: Havoc Pennington <hp@pobox.com>
4458    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4459
4460commit 5e8a7ade2dc8aeeeb8013785ca3f24c6057ae443
4461Author: Christoph Reimann <oss@arcor.de>
4462Date:   Mon Aug 16 20:24:40 2010 +0200
4463
4464    small fix to get rid of some compiler warnings
4465    also added very basic documentation for xkb
4466
4467commit b89f634ff9b321f21874cd45e398d661a6ff726e
4468Author: Christoph Reimann <oss@arcor.de>
4469Date:   Mon Aug 16 18:22:42 2010 +0200
4470
4471    small fix in the xkb pkg file
4472
4473commit 22e1013131984a217e9bddeac3a6a4183e35f0c1
4474Author: Christoph Reimann <oss@arcor.de>
4475Date:   Mon Aug 16 18:19:16 2010 +0200
4476
4477    added accessors for special cases
4478    major bugfixes include: rewrite of prefix related functions, merge of serialize/unserialize/... generators, extended field name resolution
4479
4480commit 35f901a0f2bc3f5bb30dc6ff9d791679c9e84c05
4481Author: Julien Danjou <julien@danjou.info>
4482Date:   Fri Aug 13 13:46:37 2010 +0200
4483
4484    Release libxcb 1.7
4485    
4486    Signed-off-by: Julien Danjou <julien@danjou.info>
4487
4488commit fe0e32b5fa3923fae97210e974c0f96a085116cb
4489Author: Christoph Reimann <oss@arcor.de>
4490Date:   Sun Aug 8 21:25:13 2010 +0200
4491
4492    special case 'intermixed variable and fixed size fields': fixed reply side, needs testing
4493
4494commit 77b594f9583ea0247ff27130316d8e045da7f921
4495Author: Christoph Reimann <oss@arcor.de>
4496Date:   Thu Aug 5 15:55:28 2010 +0200
4497
4498    renamed most _unserialize() functions to _sizeof() and fixed _unserialize() for the special case of intermixed variable and fixed size fields
4499
4500commit dd1a4dbe20d6b5fd33aeb65e662bb2ca18665518
4501Author: Eamon Walsh <efw@eamonwalsh.com>
4502Date:   Thu Aug 5 00:48:08 2010 -0400
4503
4504    Tutorial uses wrong function.
4505    
4506    https://bugs.freedesktop.org/show_bug.cgi?id=29392
4507    
4508    Signed-off-by: Eamon Walsh <efw@eamonwalsh.com>
4509
4510commit b187f029d6bb693f0294bad5261ec486b140f185
4511Author: Christoph Reimann <oss@arcor.de>
4512Date:   Mon Aug 2 23:30:42 2010 +0200
4513
4514    attempt to fix special case: variable fields followed by fixed size fields
4515
4516commit a700eeb502b5fe902e4c93cc707100ec868ce946
4517Author: Christoph Reimann <oss@arcor.de>
4518Date:   Sun Aug 1 23:40:20 2010 +0200
4519
4520    bug fixes for all kinds of 'special cases'
4521
4522commit 1c590d5a86ae854e53f388e40c952e92f11d59e6
4523Author: Christoph Reimann <oss@arcor.de>
4524Date:   Thu Jul 22 16:41:15 2010 +0200
4525
4526    partial rewrite of serialize helper functions completed;
4527    _serialize() & _unserialize() have been tested for switch derived from valueparam
4528
4529commit 566ae9baee20fb6147b94b89a26796087461bae8
4530Author: Christoph Reimann <oss@arcor.de>
4531Date:   Tue Jul 20 22:46:37 2010 +0200
4532
4533    preliminary handling of further special cases in unserialize
4534    first attempts to unify serialize and unserialize
4535
4536commit 4e665e1580ece7bc9505f3a2f657959669ffcd05
4537Author: Christoph Reimann <oss@arcor.de>
4538Date:   Thu Jul 15 18:43:43 2010 +0200
4539
4540    added generating code for _serialize() in case of variable sized structs (largely untested)
4541
4542commit d0031456097f009bdb82fb979924e71ca38c539b
4543Author: Christoph Reimann <oss@arcor.de>
4544Date:   Tue Jul 13 20:08:51 2010 +0200
4545
4546    xkb: added pkg config file
4547
4548commit 86704124b1fd62c30441ace1f3f8e2c316801c53
4549Author: Christoph Reimann <oss@arcor.de>
4550Date:   Tue Jul 13 20:06:08 2010 +0200
4551
4552    new and still preliminary functions for switch; feautures include
4553    - API compatibility with valueparam
4554    - request _aux() auxiliary functions
4555    - _serialize() and _unserialize() auxiliary functions
4556    - new data type that allows mixing of fixed and variable size members
4557
4558commit 80322d11636dd638902660d80481080d2fad40fe
4559Author: Christoph Reimann <oss@arcor.de>
4560Date:   Tue Jul 13 19:59:23 2010 +0200
4561
4562    xkb: updated configure.ac/Makefile.am
4563
4564commit 8c2707773b3621fb8bbda9021d23944f5be34aab
4565Author: Christoph Reimann <oss@arcor.de>
4566Date:   Tue Jul 13 19:56:44 2010 +0200
4567
4568    added xcb_sumof() with restriction to uint8_t
4569
4570commit 496efb7624d14b4ca0391f44926edc448cbd605e
4571Author: Jamey Sharp <jamey@minilop.net>
4572Date:   Tue Jul 13 07:01:06 2010 -0700
4573
4574    _xcb_conn_wait: Fix whitespace.
4575    
4576    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4577
4578commit 74057c7eb6836353960ce3849703ce20e45089bc
4579Author: Jeremy Huddleston <jeremyhu@apple.com>
4580Date:   Mon Jul 12 16:53:53 2010 -0700
4581
4582    AC_CHECK_PROG(LAUNCHD, [launchd], [yes], [no], [$PATH$PATH_SEPARATOR/sbin])
4583    
4584    launchd: Explicitly search /sbin
4585    
4586    Previously, launchd wasn't found if /sbin wasn't in the user's PATH.
4587    https://bugs.freedesktop.org/show_bug.cgi?id=29028
4588    
4589    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4590
4591commit 75ff427d41fc10d00b780c965c289fc02c81aaac
4592Author: Vincent Torri <vtorri@univ-evry.fr>
4593Date:   Sun Jun 20 20:50:06 2010 -0700
4594
4595    configure.ac: Report which extensions are being built.
4596    
4597    I was surprised to see that xinput was not installed. Looking at
4598    configure.ac, it seems that it is disabled by default. Maybe configure
4599    should output the status of the different extensions.
4600
4601commit de3cdad87a341c238c544425f4dd574b19c58ae3
4602Author: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
4603Date:   Fri Jun 11 16:30:47 2010 +0300
4604
4605    xcb_connect_to_display_with_auth_info: Fix memory leak
4606    
4607    protocol and host are allocated in _xcb_parse_display but ownership of
4608    them is passed to the caller. They have to be freed in
4609    xcb_connect_to_display_with_auth_info.
4610    
4611    Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
4612    Signed-off-by: Peter Harris <pharris@opentext.com>
4613
4614commit 18718d483e0982c779a61c71176fb0e64f850015
4615Author: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
4616Date:   Fri Jun 11 16:30:46 2010 +0300
4617
4618    _xcb_parse_display: Fix error path
4619    
4620    xcb_parse_display claims that there is no side effects when failing.
4621    That requires _xcb_parse_display to free the memory in failure case.
4622    
4623    Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
4624    Signed-off-by: Peter Harris <pharris@opentext.com>
4625
4626commit 3f79628becbd3b0eff1aef804902eb739fac4403
4627Author: Jeremy Huddleston <jeremyhu@apple.com>
4628Date:   Wed May 12 19:53:45 2010 -0700
4629
4630    xcb_open: Improve protocol/host parsing
4631    
4632    Support scenarios where host is not set and protocol is.  eg:
4633    
4634    DISPLAY=tcp/:0
4635    
4636    as well as the "inet" and "inet6" alias for "tcp" for compatability
4637    with Xlib
4638    
4639    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4640    Reviewed-by: Jamey Sharp <jamey@minilop.net>
4641
4642commit e4b746ac13e89b99abd80b3d1fa2a16796da3b6d
4643Author: Marcin Kościelnicki <koriakin@0x04.net>
4644Date:   Thu May 13 21:05:57 2010 +0000
4645
4646    Add ~ operator support in code generator
4647    
4648    Reviewed-by: Julien Cristau <jcristau@debian.org>
4649    Signed-off-by: Julien Danjou <julien@danjou.info>
4650
4651commit 5e86cb05666c448de2f61c23ae94e94ef4b38d64
4652Author: Arnaud Fontaine <arnau@debian.org>
4653Date:   Fri Apr 30 18:49:18 2010 +0100
4654
4655    Fix GCC error on undeclared variable when not using abstract socket
4656    
4657    This is a regression found by tinderbox in previous commit:
4658    
4659    xcb_util.c: In function '_xcb_open':
4660    xcb_util.c:213: error: 'fd' undeclared (first use in this function)
4661
4662commit a546d00091de0ab16374dec55e8e2fa87d6bbebf
4663Author: Arnaud Fontaine <arnau@debian.org>
4664Date:   Fri Apr 30 14:47:16 2010 +0200
4665
4666    Get rid of PATH_MAX and MAXPATHLEN
4667    
4668    There could be no upper limit on the length of a path according
4669    to POSIX, therefore these macros may not be defined at all on
4670    some systems (such as GNU Hurd).
4671    
4672    Signed-off-by: Arnaud Fontaine <arnau@debian.org>
4673    Reviewed-by: Peter Harris <pharris@opentext.com>
4674
4675commit d06857217328c2283a8956788d72646fc67216fb
4676Author: Jeremy Huddleston <jeremyhu@apple.com>
4677Date:   Fri Apr 23 21:57:26 2010 -0700
4678
4679    Use limits.h instead of syslimits.h
4680    
4681    Regression found by tinderbox in 89b3485dadef47a30264a5bf150b96522183376b
4682    
4683    xcb_util.c:31:27: error: sys/syslimits.h: No such file or directory
4684    xcb_util.c: In function '_xcb_open':
4685    xcb_util.c:148: error: 'PATH_MAX' undeclared (first use in this function)
4686    
4687    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4688
4689commit 89b3485dadef47a30264a5bf150b96522183376b
4690Author: Jeremy Huddleston <jeremyhu@apple.com>
4691Date:   Fri Apr 23 17:29:25 2010 -0700
4692
4693    Reworked launchd support to work better with _xcb_parse_display
4694    
4695    Fixes: http://xquartz.macosforge.org/trac/ticket/390
4696    
4697    Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
4698
4699commit 2dcf8b025be88a25d4333abdc28d425b88238d96
4700Author: Jeetu Golani <jeetu.golani@gmail.com>
4701Date:   Fri Apr 23 00:47:16 2010 +0530
4702
4703    Replaced read() in read_block and _xcb_in_read() with recv for all
4704    platforms. MSG_WAITALL is undefined in MinGW so it's been explicitly
4705    defined in xcb_in.c
4706
4707commit 56962e42a509dc4d0d9541e46b93689dac61c4fd
4708Author: Jeetu Golani <jeetu.golani@gmail.com>
4709Date:   Thu Apr 22 23:23:27 2010 +0530
4710
4711    Set errno=0 in read_block. On Win32 there is no errno and this makes the
4712    do..while loop execute only once. Also set the return value to -1 in
4713    _xcb_open if control reaches the end - if all goes well it shouldn't
4714    reach there.
4715
4716commit b0525e242368fffbc77ebb45293f34e80847e65a
4717Author: Jamey Sharp <jamey@minilop.net>
4718Date:   Sat Apr 17 17:59:11 2010 -0700
4719
4720    Always wake up readers after writing.
4721    
4722    Since writers must make sure they read as well, threads may have gone to
4723    sleep waiting for the opportunity to read. The writer must wake up one
4724    of those readers or the application can hang.
4725    
4726    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4727    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4728
4729commit eff3851ba80c42b5b3ba240f7e9049d7b0fac6f0
4730Author: Jamey Sharp <jamey@minilop.net>
4731Date:   Sun Mar 28 10:31:55 2010 -0700
4732
4733    Fix strict-aliasing warning when getting generic event length.
4734    
4735    xcb_ge_event_t has its length field in the same place that
4736    xcb_generic_reply_t does, so there's no need to cast the generic reply.
4737    
4738    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4739    Cc: Peter Hutterer <peter.hutterer@who-t.net>
4740    Reviewed-by: Julien Danjou <julien@danjou.info>
4741
4742commit 6dd8228a137d280ce24cec604a419129d8ed0e8e
4743Author: Jamey Sharp <jamey@minilop.net>
4744Date:   Fri Feb 12 12:25:05 2010 -0800
4745
4746    Delete a useless level of indirection from _xcb_out_send's parameters.
4747    
4748    _xcb_out_send needs _xcb_conn_wait to store back its progress so it can
4749    be reinvoked to pick up where it left off---but then _xcb_out_send
4750    guarantees that it leaves either an empty output vector or a shut-down
4751    connection, so *its* callers never care how much progress was made.
4752    
4753    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4754    Reviewed-by: Josh Triplett <josh@freedesktop.org>
4755
4756commit a63fbc9d6c484e5ad7a5e9d56f81b8e1a2f38a82
4757Author: Julien Danjou <julien@danjou.info>
4758Date:   Fri Apr 9 16:57:51 2010 +0200
4759
4760    Release libxcb 1.6
4761
4762commit a1d9aa6e07a297f4108b4ad787336f74c86a3312
4763Author: Rémi Denis-Courmont <remi.denis-courmont@nokia.com>
4764Date:   Fri Mar 26 23:12:47 2010 +0200
4765
4766    Fail if fd is too big and poll() is not available
4767    
4768    Depending on the process file limit, a file descriptor can be larger
4769    than the capacity of fd_set. There is no portable way to create a
4770    large enough fd_set at run-time. So we just fail if the file descriptor
4771    number is too high and poll() is not available.
4772    
4773    Reviewed-by: Jamey Sharp <jamey@minilop.net>
4774    Signed-off-by: Julien Danjou <julien@danjou.info>
4775
4776commit 0e0c80e749eccf121e55c1e855c48d03b54f33ef
4777Author: Jeetu Golani <jeetu.golani@gmail.com>
4778Date:   Wed Mar 31 22:00:04 2010 +0530
4779
4780    xcb_in.c #ifndef _WIN32 inside of #if USE_POLL redundant and removed
4781
4782commit e8009194c9f5a6995c4a9b03a7a49d5bc09e96fc
4783Author: Jeetu Golani <jeetu.golani@gmail.com>
4784Date:   Wed Mar 31 09:50:51 2010 +0530
4785
4786    restablished inclusion of fcntl.h and netinet/tcp.h in xcb_util.c -- without these the code no longer compiled on *ix
4787
4788commit d302f1e9b158d0a51936c28e5dc66251d90d1d56
4789Author: Jeetu Golani <jeetu.golani@gmail.com>
4790Date:   Mon Mar 29 22:37:33 2010 +0530
4791
4792    changes in xcb_windefs.h - the flag  _XCB_WINDEFS_H replaces WINDEFS_H
4793
4794commit 36c9a985aaee655c118c9f7b8425d3ac9ce0f840
4795Author: Jeetu Golani <jeetu.golani@gmail.com>
4796Date:   Mon Mar 29 22:31:49 2010 +0530
4797
4798    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
4799
4800commit bce72f63d2dfb61661f81e305ad3a7db0334403c
4801Author: Jeetu Golani <jeetu.golani@gmail.com>
4802Date:   Fri Mar 26 09:40:09 2010 +0530
4803
4804    Win32 code for xcb-1.5
4805
4806commit d18d03d6f37ec220805855d840950716e22354e4
4807Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
4808Date:   Fri Mar 12 23:51:32 2010 +0100
4809
4810    Fix authentication on hpux and Hurd
4811    
4812    libxcb's 010e5661 (Fix XDM-AUTHORIZATION-1 (bug #14202)) mistakenly
4813    inverted a few lines of code, making local socket authentication fail on
4814    hpux and Hurd: when getpeername fails, sockname needs to be initialized
4815    by getsockname before its address family can be checked.
4816    
4817    Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
4818    Signed-off-by: Julien Danjou <julien@danjou.info>
4819
4820commit 53a9834e4cdd11aba8c1cc49347f09a958107de8
4821Author: Rémi Denis-Courmont <remi@remlab.net>
4822Date:   Sat Feb 13 12:23:51 2010 +0200
4823
4824    Open the X11 socket with close-on-exec flag
4825    
4826    This saves the X11 connection from leaking into children processes.
4827    On Linux, this is fully thread-safe using SOCK_CLOEXEC. On other
4828    systems, there is a small race condition.
4829    
4830    Signed-off-by: Julien Danjou <julien@danjou.info>
4831
4832commit 367882fa32489ebafcd9afc04fbf732b02ceb33a
4833Author: Peter Harris <pharris@opentext.com>
4834Date:   Tue Feb 9 16:27:22 2010 -0500
4835
4836    Support xcb_discard_reply
4837    
4838    This function is useful for dynamic language garbage collectors. Frequently
4839    a GC cycle may run before you want to block wainting for a reply.
4840    
4841    This function is also marginally useful for libxcb apps that issue
4842    speculative requests (eg. xlsclients).
4843    
4844    Reviewed-by: Jamey Sharp <jamey@minilop.net>
4845    Tested-by: Eamon Walsh <efw@eamonwalsh.com>
4846    Signed-off-by: Peter Harris <pharris@opentext.com>
4847
4848commit be7e528eae62ddee14fa50f2c0e9036bafbc9f81
4849Author: Rémi Denis-Courmont <remi@remlab.net>
4850Date:   Thu Jan 7 18:08:35 2010 +0200
4851
4852    xcb_connect_to_fd: fix descriptor leak on memory error path
4853    
4854    Signed-off-by: Julien Danjou <julien@danjou.info>
4855
4856commit c7a57043da1717c18703a38772555fea6aa811bf
4857Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
4858Date:   Fri Dec 11 17:24:43 2009 -0800
4859
4860    configure.ac: Fix a typo on the last commit.
4861    
4862    Signed-off-by: Jeremy Huddleston <jeremyhu@freedesktop.org>
4863
4864commit 58c96da9283d3c2f65c818bd952f611ec4e4dad0
4865Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
4866Date:   Fri Dec 11 17:15:16 2009 -0800
4867
4868    darwin: xnu doesn't support poll on ttys on the master side.
4869    
4870    <rdar://problem/7360546>
4871    
4872    Signed-off-by: Jeremy Huddleston <jeremyhu@freedesktop.org>
4873
4874commit 1e746fd89ae93965183c759b969ff9f4d5dbc9d8
4875Author: Jim Ingram <ingramj@gmail.com>
4876Date:   Mon Dec 7 14:41:18 2009 -0600
4877
4878    Tutorial code fix
4879    
4880    Accessed elements of names[] after freeing them in the first example.
4881    
4882    Signed-off-by: Julien Danjou <julien@danjou.info>
4883
4884commit a470579ba29c8f39e77668558a08bb173297711f
4885Author: Rémi Cardona <remi@gentoo.org>
4886Date:   Fri Dec 4 22:06:20 2009 +0100
4887
4888    DRI2 extension requires xcb-proto 1.6 or newer
4889    
4890    Signed-off-by: Rémi Cardona <remi@gentoo.org>
4891    Signed-off-by: Julien Danjou <julien@danjou.info>
4892
4893commit 6e875a82a8a1a324067970a5f9d7442585ee0eef
4894Author: Julien Danjou <julien@danjou.info>
4895Date:   Thu Dec 3 11:34:18 2009 +0100
4896
4897    build: simplify extension building
4898    
4899    Signed-off-by: Julien Danjou <julien@danjou.info>
4900
4901commit 48217ac986d77eba40f3af4df597a734f4fd3690
4902Author: Julien Danjou <julien@danjou.info>
4903Date:   Thu Dec 3 10:08:59 2009 +0100
4904
4905    Release libxcb 1.5
4906    
4907    Signed-off-by: Julien Danjou <julien@danjou.info>
4908
4909commit 1cf2a87def76f4646fe05e282b45605d572f2513
4910Author: Adam Jackson <ajax@redhat.com>
4911Date:   Wed Dec 2 14:31:56 2009 -0500
4912
4913    setsockopt(SO_KEEPALIVE) on TCP display connections.
4914    
4915    This matches xtrans behaviour in SocketINETConnect, and makes it so apps
4916    don't hang forever if their display dies.
4917    
4918    Signed-off-by: Adam Jackson <ajax@redhat.com>
4919
4920commit e82c34c2f7ac3fbb23ab14cbee8df2dd9178f5a4
4921Author: Eric Anholt <eric@anholt.net>
4922Date:   Thu Oct 15 13:13:47 2009 -0700
4923
4924    Add DRI2 support. (v2)
4925    
4926    v2: Build fix from jcristau.
4927
4928commit a27c77ccae059fb64e0506648b81677858d3b05b
4929Author: Eric Anholt <eric@anholt.net>
4930Date:   Thu Oct 15 13:14:49 2009 -0700
4931
4932    Fix typo in the tutorial.
4933
4934commit 17af34b4a438a80d3dcad93bb5254366791a488f
4935Author: Paulo R. Zanoni <pzanoni@mandriva.com>
4936Date:   Fri Sep 18 15:05:46 2009 -0300
4937
4938    Don't build docs if 'dot' is not found Signed-off-by: Paulo R. Zanoni <pzanoni@mandriva.com>
4939
4940commit 29207e2943ad56fb8b4d2b7f07b1530cb1d7d9be
4941Author: Julien Cristau <jcristau@debian.org>
4942Date:   Mon Aug 31 17:51:36 2009 +0200
4943
4944    Fix check dependency
4945    
4946    Bugzilla #21992
4947    
4948    make -j check fails because the check-local rule gets executed before
4949    the tests actually ran, so CheckLog*.xml doesn't exist.
4950    
4951    Signed-off-by: Julien Danjou <julien@danjou.info>
4952
4953commit ff2e2e35bcac1d835c9eecf9ac8b3072005989a5
4954Author: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
4955Date:   Sat Jun 6 15:52:20 2009 +0100
4956
4957    Cygwin build fix: Add -no-undefined to libtool flags
4958    
4959    -no-undefined is needed to tell libtool a shared library can be built
4960    on platforms which require all references to be statisfied at link time.
4961    
4962    Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
4963    Signed-off-by: Julien Danjou <julien@danjou.info>
4964
4965commit 96ff5fb635bc333e80ebce6cc03008a1b9992ff6
4966Author: Arnaud Fontaine <arnau@debian.org>
4967Date:   Wed Jul 15 16:15:02 2009 +0100
4968
4969    Release libxcb 1.4
4970
4971commit f4c2794bf5990a0b2f6168f2b22b60b15e08ac44
4972Author: Arnaud Fontaine <arnau@debian.org>
4973Date:   Wed Jul 15 16:03:56 2009 +0100
4974
4975    Add majorCode, minorCode and resourceID fields to X generic error
4976
4977commit e06955ed66cb499ada52b8add6709edd38d70054
4978Author: Jamey Sharp <jamey@minilop.net>
4979Date:   Mon Jul 6 13:14:35 2009 -0700
4980
4981    Fix precedence bug: wrong length for big-requests preceded by sync.
4982    
4983    Also replace excessively clever use of bitwise OR with equivalent
4984    addition.
4985    
4986    Reported-by: Geoffrey Li <geoffrey@seitopos.com>
4987    Signed-off-by: Jamey Sharp <jamey@minilop.net>
4988    Signed-off-by: Josh Triplett <josh@joshtriplett.org>
4989
4990commit 9e191c722a0be4b5f3d993165055a71f85c21882
4991Author: Julien Cristau <jcristau@debian.org>
4992Date:   Fri May 29 14:41:59 2009 +0200
4993
4994    Fix libxcb-randr version info
4995    
4996    The SONAME shouldn't have been bumped in 1.3, only new symbols were
4997    added.
4998    
4999    Signed-off-by: Julien Danjou <julien@danjou.info>
5000
5001commit f44dc519c5b324335d05f6e7fb31b78062a8c5f5
5002Author: Julien Danjou <julien@danjou.info>
5003Date:   Fri May 29 08:26:05 2009 +0200
5004
5005    Release libxcb 1.3
5006
5007commit ee89850e68205a7f8961ace0839b5be86040dade
5008Author: elupus <elupus@ecce.se>
5009Date:   Tue May 26 16:14:48 2009 +0200
5010
5011    Disable Nagle on TCP socket
5012    
5013    Signed-off-by: Julien Danjou <julien@danjou.info>
5014
5015commit 62fe187e2d617eb0feb1ca03d8b4a64db9dd952b
5016Author: Bob Ham <rah@bash.sh>
5017Date:   Mon May 25 12:20:23 2009 +0200
5018
5019    Store xcbproto version libxcb was compiled with
5020    
5021    Signed-off-by: Julien Danjou <julien@danjou.info>
5022
5023commit efbe96ee0ab0c5511035eee99f8fe7b38d5f65fa
5024Author: Julien Danjou <julien@danjou.info>
5025Date:   Sat May 9 17:39:34 2009 +0200
5026
5027    depends on recent xcb-proto and bump version of randr
5028    
5029    Signed-off-by: Julien Danjou <julien@danjou.info>
5030
5031commit 010e566126625c56bdf9989085bacf731520ff87
5032Author: Bart Massey <bart@cs.pdx.edu>
5033Date:   Tue Apr 21 08:39:52 2009 +0200
5034
5035    Fix XDM-AUTHORIZATION-1 (bug #14202)
5036    
5037    With this patch, we know use correctly the socket address or peer
5038    address for authentication purpose.
5039    
5040    Signed-off-by: Julien Danjou <julien@danjou.info>
5041
5042commit ca978a9dae621126075712f9e2c29591208570bc
5043Author: Julien Danjou <julien@danjou.info>
5044Date:   Tue Apr 7 14:22:57 2009 +0200
5045
5046    util: remove useless strlen calls from decnet opening
5047    
5048    Signed-off-by: Julien Danjou <julien@danjou.info>
5049
5050commit cc191431412d8764c645a51b0f106c0dfe652213
5051Author: Julien Danjou <julien@danjou.info>
5052Date:   Tue Apr 7 14:18:40 2009 +0200
5053
5054    util: merge common code for xcb_connect
5055    
5056    Many code was duplicated between xcb_connect_to_display_with_auth_info
5057    and xcb_connect(). We merge both, since the difference is just about the
5058    xcb_auth_info_t pointer being supplied, or not.
5059    
5060    Signed-off-by: Julien Danjou <julien@danjou.info>
5061
5062commit 8797e053b2b2ee989f47490c7687b9a2fbdb0021
5063Author: Julien Danjou <julien@danjou.info>
5064Date:   Tue Apr 7 13:37:40 2009 +0200
5065
5066    util: open_abstract gets filelen as parameters
5067    
5068    That saves us from a couple of strlen() calls.
5069    
5070    Signed-off-by: Julien Danjou <julien@danjou.info>
5071
5072commit f0b29819749b769e5a8d313bf1bab80d6513208b
5073Author: Julien Danjou <julien@danjou.info>
5074Date:   Tue Apr 7 11:55:30 2009 +0200
5075
5076    auth: use snprintf() return value
5077    
5078    That save us from a strlen().
5079    
5080    Signed-off-by: Julien Danjou <julien@danjou.info>
5081
5082commit 9f24c91f91dd68a52e46191b686283b0df38d2f5
5083Author: Julien Danjou <julien@danjou.info>
5084Date:   Tue Apr 7 11:49:13 2009 +0200
5085
5086    auth: precompute authnameslen
5087    
5088    Signed-off-by: Julien Danjou <julien@danjou.info>
5089
5090commit 2a4f1cf3801080276694f5026d35220b65201038
5091Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5092Date:   Thu Apr 9 05:12:02 2009 -0700
5093
5094    darwin: Don't use poll() on versions of darwin before darwin10
5095
5096commit 6e2e87d0bbdff87f127986a0666445160d52e6a5
5097Author: Samuel Thibault <samuel.thibault@ens-lyon.org>
5098Date:   Mon Apr 6 03:31:23 2009 +0200
5099
5100    Local socket connections do not work on hurd-i386
5101    
5102    Local socket connections currently do not work on hurd-i386 because
5103    xcb_auth calls getpeername() on the client socket, but hurd-i386 does
5104    not implement anything in that case (I actually wonder what reasonable
5105    value could be returned).  In such case the xcb code does not actually
5106    need the peer name anyway.
5107    
5108    Signed-off-by: Julien Danjou <julien@danjou.info>
5109
5110commit f916062edf0e04cd4e0a78f6975892f59fae3b60
5111Author: Michael Ost <most@museresearch.com>
5112Date:   Mon Mar 30 11:09:32 2009 +0200
5113
5114    use poll() instead of select() when available
5115    
5116    Signed-off-by: Julien Danjou <julien@danjou.info>
5117
5118commit beccb0be15f5699c942a0af33307d9e4bf797e2a
5119Author: Bart Massey <bart@cs.pdx.edu>
5120Date:   Tue Mar 24 16:24:04 2009 -0700
5121
5122    kludgily hand-merged xid fixes
5123    
5124    Signed-off-by: Bart Massey <bart@cs.pdx.edu>
5125    Signed-off-by: Julien Danjou <julien@danjou.info>
5126
5127commit 1e9c0f1012b6d349f92ea7246194b8667d27f849
5128Author: Julien Danjou <julien@danjou.info>
5129Date:   Mon Mar 16 10:26:02 2009 +0100
5130
5131    Fix do_append() arguments
5132    
5133    Signed-off-by: Julien Danjou <julien@danjou.info>
5134
5135commit fcb433db80315a44154248a9229c9cfcbab63f04
5136Author: Julien Danjou <julien@danjou.info>
5137Date:   Sun Mar 15 10:18:50 2009 +0100
5138
5139    Copy full IPv4 mapping (Bug #20665)
5140    
5141    Signed-off-by: Julien Danjou <julien@danjou.info>
5142
5143commit eaa71eac02c6a862ab23e8afcce12d9f38590338
5144Author: Peter Harris <pharris@opentext.com>
5145Date:   Fri Mar 13 15:24:55 2009 -0400
5146
5147    Avoid name collisions between xidtype and enum.
5148    
5149    These changes are necessary to build with latest xcb/proto.
5150    
5151    Signed-off-by: Peter Harris <pharris@opentext.com>
5152
5153commit e986d1ee5a126dc38113125075a1e986235ba7c7
5154Author: Peter Harris <pharris@opentext.com>
5155Date:   Fri Mar 13 15:25:30 2009 -0400
5156
5157    Revert "Don't use enums in generated C code"
5158    
5159    This commit broke xcb/util.
5160    
5161    This reverts commit 9984b72888108a038d6b3f7dee374d17e26ef9e2.
5162    
5163    Signed-off-by: Peter Harris <pharris@opentext.com>
5164
5165commit 9984b72888108a038d6b3f7dee374d17e26ef9e2
5166Author: Peter Harris <pharris@opentext.com>
5167Date:   Wed Feb 25 18:48:50 2009 -0500
5168
5169    Don't use enums in generated C code - use integer constants instead.
5170    
5171    Signed-off-by: Peter Harris <pharris@opentext.com>
5172
5173commit b08e1535cf0716fc917eaa1f5d5f6d1960bf1e3c
5174Author: Julien Danjou <julien@danjou.info>
5175Date:   Tue Feb 17 13:37:29 2009 +0100
5176
5177    Release libxcb 1.2
5178    
5179    Signed-off-by: Julien Danjou <julien@danjou.info>
5180
5181commit 2e65188f91800d7e6a8d74fa077b46f8f67b6893
5182Author: Julien Danjou <julien@danjou.info>
5183Date:   Mon Feb 16 11:44:20 2009 +0100
5184
5185    Stop providing autogenerated files in tarball
5186    
5187    Signed-off-by: Julien Danjou <julien@danjou.info>
5188
5189commit 37f5ce3ef46c8af38ec0103b9d960615a0a1e058
5190Author: Julien Danjou <julien@danjou.info>
5191Date:   Thu Dec 11 11:17:13 2008 +0100
5192
5193    Release libxcb 1.1.93
5194    
5195    Signed-off-by: Julien Danjou <julien@danjou.info>
5196
5197commit f896ae8c53e99f39b347f9f0ac2b4a8cc12cad6f
5198Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5199Date:   Sun Nov 23 17:25:21 2008 -0800
5200
5201    Shutup compiler warning about unused variable...
5202
5203commit d79621b25ba6784135b1a3aa51e9561fcf72fe7f
5204Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5205Date:   Sun Nov 23 17:23:17 2008 -0800
5206
5207    Apple: Apple launchd cleanup
5208    Added --with-launchd option instead of just using __APPLE__
5209    Fixed opening launchd fd when displayname=NULL
5210
5211commit 9b79ae49f709707e99b8487b01f9d3f102754bd3
5212Author: Jeremy Huddleston <jeremyhu@freedesktop.org>
5213Date:   Sat Nov 8 14:41:23 2008 -0800
5214
5215    Apple: Enable support for launchd DISPLAY socket
5216
5217commit 43b612a5e99ccdfb787a11b2b7c3179fc242edf3
5218Author: Peter Harris <peter.harris@hummingbird.com>
5219Date:   Wed Nov 12 14:45:04 2008 -0500
5220
5221    Treat XIDs the same as other cardinal values.
5222    
5223    This fixes a bug where c_client.py wasn't generating *_end functions,
5224    but expected them to exist in order to find the subsequent list's start.
5225    
5226    Signed-off-by: Peter Harris <peter.harris@hummingbird.com>
5227
5228commit 902cade8ec9e7eb5d29f6f6d61cac0470c300aec
5229Author: Julien Danjou <julien@danjou.info>
5230Date:   Tue Nov 4 10:19:14 2008 +0100
5231
5232    Release libxcb 1.1.92
5233    
5234    Signed-off-by: Julien Danjou <julien@danjou.info>
5235
5236commit fa452cc9b2bb69fa0603dfd97e00e540b6b52840
5237Author: Josh Triplett <josh@freedesktop.org>
5238Date:   Sun Mar 16 23:16:31 2008 -0700
5239
5240    Support handing off socket write permission to external code.
5241    
5242    Libraries like Xlib, some XCB language bindings, and potentially others
5243    have a common problem: they want to share the X connection with XCB. This
5244    requires coordination of request sequence numbers.  Previously, XCB had an
5245    Xlib-specific lock, and allowed Xlib to block XCB from making requests.
5246    Now we've replaced that lock with a handoff mechanism, xcb_take_socket,
5247    allowing external code to ask XCB for permission to take over the write
5248    side of the socket and send raw data with xcb_writev.  The caller of
5249    xcb_take_socket must supply a callback which XCB can call when it wants
5250    the write side of the socket back to make a request.  This callback
5251    synchronizes with the external socket owner, flushes any output queues if
5252    appropriate, and then returns the sequence number of the last request sent
5253    over the socket.
5254    
5255    Commit by Josh Triplett and Jamey Sharp.
5256    Handoff mechanism inspired by Keith Packard.
5257
5258commit baff35a04b0e8d21821850a405a550d86a8aeb6f
5259Author: Jamey Sharp <jamey@minilop.net>
5260Date:   Wed May 21 14:44:16 2008 -0700
5261
5262    Track 64-bit sequence numbers internally.
5263    
5264    External APIs that used 32-bit sequence numbers continue to do so.
5265    
5266    Commit by Josh Triplett and Jamey Sharp.
5267
5268commit 96e55444b9b9500420f9132a1ace720100a26398
5269Author: Jamey Sharp <jamey@minilop.net>
5270Date:   Fri Mar 14 20:18:52 2008 -0700
5271
5272    Use sequence number ranges in pending replies
5273    
5274    This allows optimizing adjacent pending replies with the same flags, and
5275    will help support default flags for a range of future requests.
5276    
5277    Commit by Josh Triplett and Jamey Sharp.
5278
5279commit 059ca642c76639fee958dc6054070de85e257e98
5280Author: Jamey Sharp <jamey@minilop.net>
5281Date:   Fri Mar 14 12:08:58 2008 -0700
5282
5283    Inline _xcb_lock_io, _xcb_unlock_io, and _xcb_wait_io.
5284    
5285    These functions are once again a single pthread call, so just make that
5286    call directly.
5287
5288commit d989656cde2ee7a4a66b2065209ef389495f3452
5289Author: Jamey Sharp <jamey@minilop.net>
5290Date:   Fri Mar 14 12:08:32 2008 -0700
5291
5292    Remove libxcb-xlib and xcbxlib.h.
5293
5294commit dcbef23d730b95ef7dc9ef524a4c3fc0017b63d2
5295Author: Julien Danjou <julien@danjou.info>
5296Date:   Tue Oct 14 23:39:07 2008 +0200
5297
5298    build: fix configure.ac AC_DEFINE
5299    
5300    Rather use AC_DEFINE_UNQUOTED and only once.
5301    
5302    Signed-off-by: Julien Danjou <julien@danjou.info>
5303
5304commit cebd482a20fcc2b2dae0683c38e917a3740638a6
5305Author: Julien Danjou <julien@danjou.info>
5306Date:   Fri Sep 26 15:26:42 2008 +0200
5307
5308    allow compile-time setting for XCB queue buffer size
5309    
5310    Signed-off-by: Julien Danjou <julien@danjou.info>
5311
5312commit 625ed596cae6dd8175aeb6cb6f26784928042f22
5313Author: Josh Triplett <josh@freedesktop.org>
5314Date:   Wed Oct 8 16:04:25 2008 -0700
5315
5316    Remove duplicate XCB_EXTENSION calls for Composite extension
5317
5318commit db332dcda989b5b021dc220c102666f695c772cf
5319Author: Henning Sten <henning.sten@yahoo.com>
5320Date:   Sat Sep 20 13:08:58 2008 +0200
5321
5322    fix tiny memory leak in read_packet (leak only happens when malloc returns NULL so it's very rare)
5323    
5324    Signed-off-by: Julien Danjou <julien@danjou.info>
5325
5326commit 9afadd2aef8af89a4f4ab70baeae0b848904c367
5327Author: Carsten Meier <cm@trexity.de>
5328Date:   Tue Sep 9 12:11:37 2008 +0200
5329
5330    Added generation of extern "C" for compatibility with C++
5331    
5332    The auto-generated header files now include an extern "C"
5333    declaration for compatibility with C++.
5334    
5335    Signed-off-by: Julien Danjou <julien@danjou.info>
5336
5337commit 25e59ccc0dc8baf344145d6d739229e8120330db
5338Author: Julien Cristau <jcristau@debian.org>
5339Date:   Tue Sep 9 04:42:36 2008 +0100
5340
5341    Add support for the abstract socket namespace under Linux
5342    
5343    Based on same in Xtrans.
5344    
5345    Signed-off-by: Julien Danjou <julien@danjou.info>
5346
5347commit f3f8738436d09f7e590b76e22a7a2cc4b16abd1d
5348Author: Julien Cristau <jcristau@debian.org>
5349Date:   Tue Sep 9 04:42:35 2008 +0100
5350
5351    Fix some fd leaks in _xcb_open_*()
5352    
5353    Signed-off-by: Julien Danjou <julien@danjou.info>
5354
5355commit d6d44e1bf09cca8aefbf4ce9875d7f794bf19fb1
5356Author: Bart Massey <bart@cs.pdx.edu>
5357Date:   Wed Sep 3 13:52:58 2008 -0700
5358
5359    fixed overly aggressive warning about fixed field following variable
5360
5361commit 7e0674e76186ee4491a089350511fc0d22fb3af3
5362Author: Bart Massey <bart@cs.pdx.edu>
5363Date:   Sun Aug 31 00:42:23 2008 -0700
5364
5365    added small fix to support trailing fixed fields; also warning for non-pad fixed fields
5366
5367commit 2d04a1e6cedcdc832e2db3c65ababc3aff904ec4
5368Author: Vincent Torri <vtorri at univ-evry dot fr>
5369Date:   Sun Aug 31 10:33:31 2008 +0200
5370
5371    factorize m4 macros and add one to set X extensions
5372    
5373    Signed-off-by: Julien Danjou <julien@danjou.info>
5374
5375commit bcf662c1b433b296060c66ae1656fcb5c6e697ef
5376Author: Julien Danjou <julien@danjou.info>
5377Date:   Thu Aug 28 14:35:54 2008 +0200
5378
5379    Initialize all fields of addrinfo
5380    
5381    Signed-off-by: Julien Danjou <julien@danjou.info>
5382
5383commit baf31b1bf20b49ec00d0f64bb7cc9c788a28c088
5384Author: Julien Danjou <julien@danjou.info>
5385Date:   Thu Aug 28 13:51:38 2008 +0200
5386
5387    Use ifdef instead of if for defined value
5388    
5389    Signed-off-by: Julien Danjou <julien@danjou.info>
5390
5391commit 38d5de3a5573b2e89e97d04a809a3dd38a0fe8a7
5392Author: Julien Danjou <julien@danjou.info>
5393Date:   Wed Aug 27 13:56:28 2008 +0200
5394
5395    Set namelen unsigned
5396    
5397    Signed-off-by: Julien Danjou <julien@danjou.info>
5398
5399commit 9c9c09b376fe1ddcedd03c52cfc0b06867d998c9
5400Author: Julien Danjou <julien@danjou.info>
5401Date:   Wed Aug 27 13:56:26 2008 +0200
5402
5403    Rename index to idx to avoid shadowing
5404    
5405    Signed-off-by: Julien Danjou <julien@danjou.info>
5406
5407commit c5b2e53abf0b113d4cc4105127cf848ee450aa98
5408Author: Julien Danjou <julien@danjou.info>
5409Date:   Wed Aug 27 13:56:25 2008 +0200
5410
5411    Use a signed size in read_block()
5412    
5413    Signed-off-by: Julien Danjou <julien@danjou.info>
5414
5415commit 1bbdba52116f127bed3ce812a00240b4009bbf22
5416Author: Julien Danjou <julien@danjou.info>
5417Date:   Wed Aug 27 13:56:24 2008 +0200
5418
5419    Use unsigned to compare and rename sync
5420    
5421    - i must be unsigned to be compare in the loop
5422    - sync shadow global sync() function
5423    
5424    Signed-off-by: Julien Danjou <julien@danjou.info>
5425
5426commit 6438584285de72858f97be891e16a125d13471d8
5427Author: Julien Danjou <julien@danjou.info>
5428Date:   Wed Aug 27 13:56:23 2008 +0200
5429
5430    Fix htonl() arg & convert sizeof() to signed
5431    
5432    Signed-off-by: Julien Danjou <julien@danjou.info>
5433
5434commit 1ead02e88eb9f37757eeb1cc8c762fc48d6d08ee
5435Author: Julien Danjou <julien@danjou.info>
5436Date:   Wed Aug 27 13:56:22 2008 +0200
5437
5438    initialize global_id to 0
5439    
5440    Signed-off-by: Julien Danjou <julien@danjou.info>
5441
5442commit f209d0ef7ad57a395c01ca09ecf48117a648e39c
5443Author: Peter Hutterer <peter.hutterer@who-t.net>
5444Date:   Thu Jul 17 13:57:41 2008 +0930
5445
5446    Bump to 1.1.91.
5447
5448commit cdc347938702dddbacb5af5c24988e9152b5447c
5449Author: Julien Cristau <jcristau@debian.org>
5450Date:   Wed Jul 16 23:25:04 2008 +0930
5451
5452    Make EXTHEADERS, EXTSOURCES, EXTENSION_XML unconditional of configure flags.
5453    
5454    yay, make distcheck works now even when some extensions are disabled.
5455    
5456    Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
5457
5458commit a9d15a08451c76a9250642c9f662f296196f60a0
5459Author: Petr Salinger <Petr.Salinger@seznam.cz>
5460Date:   Mon Jul 7 17:57:37 2008 +0200
5461
5462    fix FreeBSD support
5463    
5464    The GNU/kFreeBSD (and BSDs in general) have a different
5465    layout of struct sockaddr, sockaddr_in, sockaddr_un ...
5466    
5467    The first member do not have to be "sa_family",
5468    they also have "sa_len" field.
5469    
5470    Signed-off-by: Julien Danjou <julien@danjou.info>
5471
5472commit ee78071902e93ce22a3170f0937c158fd16894d8
5473Author: Peter Hutterer <peter@cs.unisa.edu.au>
5474Date:   Wed May 28 17:41:35 2008 +0930
5475
5476    Bump to 1.1.90.
5477
5478commit 424ad131b7d97d6196181c31677655e6d52b41df
5479Author: Josh Triplett <josh@freedesktop.org>
5480Date:   Wed May 28 12:26:13 2008 -0700
5481
5482    Fix variable declaration formatting
5483
5484commit 6532c715c3805128b9976ab208f1426f691056a2
5485Author: Peter Hutterer <peter@cs.unisa.edu.au>
5486Date:   Tue May 15 16:28:19 2007 +0930
5487
5488    Add xcb_ge_event_t and handling for long events.
5489    
5490    GenericEvent can be more than 32 bytes long. Ensure that the required data is
5491    pulled off the wire and tack it onto the event.
5492    
5493    Due to the structure of the xcb_generic_event_t, the data is appended AFTER
5494    the full_sequence field.
5495
5496commit b08a5909daf589d5e06c17c55d044f39c1d3479a
5497Author: Oswald Buddenhagen <ossi@kde.org>
5498Date:   Thu May 1 16:17:55 2008 -0400
5499
5500    Fix libxcb/src compile with srcdir != builddir.
5501
5502commit 4a405feba8cde8490d847a57b7e833176e18b90f
5503Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5504Date:   Wed Apr 23 20:26:28 2008 -0400
5505
5506    Replace a stray c-client.xsl in the libxcb SOURCES.  Fixes make distcheck.
5507
5508commit 40566c36d543edc2118cbb358e0303d9e8862892
5509Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5510Date:   Wed Apr 23 20:25:57 2008 -0400
5511
5512    Use the python install path from xcb-xproto.pc to locate the xcbgen package.
5513
5514commit b3832bcc46d85110fdb2b553df6249a831cfe0fa
5515Author: Jeremy Kolb <jkolb@glorfindel.hsd1.ma.comcast.net>
5516Date:   Sun Apr 20 16:26:51 2008 -0400
5517
5518    Add mention of PYTHONPATH if xcbgen cannot be found.
5519
5520commit 5ee915e12a102e86e141981bbce60ed81037dfdc
5521Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5522Date:   Fri Apr 18 16:30:08 2008 -0400
5523
5524    Add Python parser C language-dependent part.
5525
5526commit 947a2e26e4217531e612a5110e6f95296c94614f
5527Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5528Date:   Thu Jan 24 16:02:34 2008 -0500
5529
5530    Add SELinux extension support, disabled by default.
5531
5532commit 0b5f684eb0f8a8ad1887514568532037bb46020c
5533Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5534Date:   Thu Jan 24 15:57:35 2008 -0500
5535
5536    Add XInput extension support, disabled by default.
5537
5538commit c72581c844efbaaa7e632377b57678f6668327fe
5539Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5540Date:   Tue Jan 15 17:06:14 2008 -0500
5541
5542    Inputs to AC_CONFIG_FILES are automatically distributed.
5543    Hence, it is not necessary to explicitly add them to EXTRA_DIST.
5544
5545commit 7a74ba3d0212f9bfe021d6da9070f71cbc53f85b
5546Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5547Date:   Fri Dec 7 16:22:04 2007 -0500
5548
5549    Generated the configure.ac and Makefile.am's in libxcb with the idea of
5550    making each extension library individually selectable for build.
5551    
5552    Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
5553
5554commit bcd1dcec9b242d7e2185c1ae83d3884844a2647f
5555Merge: baae582 9bf8329
5556Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5557Date:   Fri Dec 7 16:18:00 2007 -0500
5558
5559    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
5560
5561commit 9bf8329b50a8edacf00efb074f73c3bb759f7c8d
5562Author: Bart Massey <bart@cs.pdx.edu>
5563Date:   Sat Nov 24 14:53:54 2007 -0800
5564
5565    make IPv6 optional
5566
5567commit 0593989103c48cc10165066d985c9f2d3355926b
5568Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5569Date:   Fri Nov 16 19:38:40 2007 -0500
5570
5571    Remove file that wasn't meant to be committed.
5572
5573commit baae5826a6f51490e842be931c8b9f76086c4d98
5574Merge: c3136d1 46413cd
5575Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5576Date:   Fri Nov 16 19:36:08 2007 -0500
5577
5578    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
5579
5580commit 46413cd85ee4f3d51a3a3e1d8ee13bc5fa6c2d5d
5581Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5582Date:   Fri Nov 16 19:34:42 2007 -0500
5583
5584    Add comment noting the requirement to free replies when finished.
5585
5586commit c3136d18321df31caa7f582d475132c2e02155de
5587Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5588Date:   Fri Nov 16 19:33:20 2007 -0500
5589
5590    Add generated comment noting requirement to free replies after use.
5591
5592commit 3c6c8f127c2bce4f45bface7dd45cc719af9de0d
5593Author: Jamey Sharp <jamey@minilop.net>
5594Date:   Sun Nov 4 17:29:13 2007 -0800
5595
5596    Release libxcb 1.1
5597
5598commit af50de26c10c93ccc4cd3bc61e92aff47651b961
5599Author: Jamey Sharp <jamey@minilop.net>
5600Date:   Sun Nov 4 17:26:21 2007 -0800
5601
5602    Revert "Generate error constants as XCB_BAD_*, similar to Xlib."
5603    
5604    Since several extensions named their errors like "BadFoo", this patch
5605    results in names like XCB_EXT_BAD_BAD_FOO, which is really awful. Those
5606    extensions are already kind of awful, as they produce structure names
5607    like xcb_ext_bad_foo_error_t, which is redundant.
5608    
5609    A patch that removes "Bad" from the XML extension descriptions, while
5610    maintaining API and ABI compatibility in XCB, is needed before this
5611    patch can be released.
5612    
5613    This reverts commit 158c9b6ba18b39f424bd524fceb66f3fec0d1616.
5614
5615commit a29fbc2645fabb96d02c382ffef499b48fb1514a
5616Author: Jamey Sharp <jamey@minilop.net>
5617Date:   Sun Oct 28 13:28:18 2007 -0700
5618
5619    Don't hold the xlib-xcb lock while sleeping: that allows deadlock.
5620    
5621    With this patch, `ico -threads 2` runs without deadlock.
5622    
5623    Many thanks to Christoph Pfister <christophpfister@gmail.com> for
5624    pointing out the problem, providing detailed analyses, explaining it to
5625    me repeatedly until I understood what was going on, and proposing and
5626    reviewing possible solutions.
5627    
5628    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5629    Acked-by: Christoph Pfister <christophpfister@gmail.com>
5630
5631commit f6b75d6090dc40918196d2b902e9616d0199af42
5632Author: Jamey Sharp <jamey@minilop.net>
5633Date:   Sun Oct 28 11:56:08 2007 -0700
5634
5635    Factor pthread_cond_wait(iolock) to _xcb_wait_io.
5636    
5637    This parallels the _xcb_lock_io and _xcb_unlock_io factoring.
5638
5639commit 4d828c5eba9fc7161c5f18650f2dbe218e1db06f
5640Author: Jamey Sharp <jamey@minilop.net>
5641Date:   Tue Oct 23 11:03:33 2007 -0700
5642
5643    Don't abort() on locking assertions if LIBXCB_ALLOW_SLOPPY_LOCK is set.
5644    
5645    But do still print a full backtrace, on platforms where that's
5646    supported.
5647    
5648    This commit follows the spirit of Novell's libxcb-sloppy-lock.diff.
5649    
5650    I strongly opposed proposals like this one for a long time. Originally I
5651    had a very good reason: libX11, when compiled to use XCB, would crash
5652    soon after a locking correctness violation, so it was better to have an
5653    informative assert failure than a mystifying crash soon after.
5654    
5655    It took some time for me to realize that I'd changed the libX11
5656    implementation (for unrelated reasons) so that it could survive most
5657    invalid locking situations, as long as it wasn't actually being used
5658    from multiple threads concurrently.
5659    
5660    The other thing that has changed is that most of the code with incorrect
5661    locking has now been fixed. The value of the assert is accordingly
5662    lower.
5663    
5664    However, remaining broken callers do need to be fixed. That's why libXCB
5665    will still noisily print a stacktrace (if possible) on each assertion
5666    failure, even when assert isn't actually invoked to abort() the program;
5667    and that's why aborting is still default. This environment variable is
5668    provided only for use as a temporary workaround for broken applications.
5669    
5670    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5671    Acked-by: Josh Triplett <josh@freedesktop.org>
5672
5673commit 09045eaac34973662aaa820a94ca8ed66d9dcb4e
5674Author: Egbert Eich <eich@freedesktop.org>
5675Date:   Thu Jul 19 17:00:18 2007 +0200
5676
5677    Allow unix:<screen> style display names again.
5678    
5679    https://bugzilla.novell.com/show_bug.cgi?id=289007
5680    This notion is used in a lot of scripts.
5681
5682commit 65ffbc6cfdb97b14689d3baef183cd50fbd31a7f
5683Author: Josh Triplett <josh@freedesktop.org>
5684Date:   Wed Jun 13 23:46:37 2007 -0700
5685
5686    Send locking assertion backtraces to stderr.  Improve the heading on the backtrace.
5687
5688commit 605c778e695a4535c35c5324325f310b5faf80e2
5689Author: Christoph Pfister <christophpfister@gmail.com>
5690Date:   Wed Jun 6 17:17:49 2007 +0200
5691
5692    Print backtraces in case an assert fails inside xlib/xcb.
5693    
5694    As you know there are some nasty libs / apps doing locking
5695    incorrectly. In order to improve the information given to the user
5696    when he encounters such a situation (people don't run apps in gdb
5697    normally) I created the patch attached.
5698    It's very non-intrusive (and affects only xlib/xcb, Josh told me on
5699    irc that it could be useful for other areas too, personally I don't
5700    think that it's really needed at other places ...).
5701    
5702    Some same outputs and the discussion of them:
5703    
5704        lxuser@pdln:/tmp$ ./main
5705        Got a backtrace:
5706        #0 /tmp/usr/lib/libxcb-xlib.so.0 [0xb7f9d728]
5707        #1 /tmp/usr/lib/libxcb-xlib.so.0(xcb_xlib_unlock+0x31) [0xb7f9d861]
5708        #2 ./test.so(function_a+0x11) [0xb7f9f3fd]
5709        #3 ./test.so(function_b+0x11) [0xb7f9f410]
5710        #4 ./main [0x80484a7]
5711        #5 /lib/libc.so.6(__libc_start_main+0xdc) [0xb7e60ebc]
5712        #6 ./main [0x80483f1]
5713        main: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
5714        Aborted
5715    
5716    That's kinda the normal situation.
5717    
5718        lxuser@pdln:/tmp$ ./main
5719        Got a backtrace:
5720        #0 /tmp/usr/lib/libxcb-xlib.so.0 [0xb7f90728]
5721        #1 /tmp/usr/lib/libxcb-xlib.so.0(xcb_xlib_unlock+0x31) [0xb7f90861]
5722        #2 /tmp/test.so [0xb7f923cd]
5723        #3 /tmp/test.so(function_b+0x11) [0xb7f923e0]
5724        #4 ./main [0x80484ab]
5725        #5 /lib/libc.so.6(__libc_start_main+0xdc) [0xb7e53ebc]
5726        #6 ./main [0x80483f1]
5727        main: xcb_xlib.c:82: xcb_xlib_unlock: Assertion `c->xlib.lock' failed.
5728        Aborted
5729    
5730    There are two possible reasons that the name doesn't appear in #2:
5731    a) a hidden symbol or a symbol with statical linkage in a library
5732    b) a symbol in an app not compiled with -rdynamic.
5733    But in both cases you still know _where_ the caller is.
5734    
5735    Note that in this example test.so was compiled with
5736    -fomit-frame-pointer; this isn't an issue as _one_ (= the caller)
5737    stack trace is still valid (as long as you don't have the insane idea
5738    to compile xcb with -fo-f-p).
5739    
5740    Another issue that may appear is "tail call elimination" (some entries
5741    are mysteriously missing; this is quite ugly, but you still get enough
5742    information so that you can do something useful with the issue e.g. by
5743    disassembling the relevant parts with gdb).
5744    
5745    Signed-off-by: Jamey Sharp <jamey@minilop.net>
5746
5747commit e20a31d72b8838cdf31b568431b5ad78492c1481
5748Author: Jamey Sharp <jamey@minilop.net>
5749Date:   Sat Jun 2 18:29:37 2007 -0700
5750
5751    xcb_poll_for_event: Return already-read events before read(2)ing again.
5752
5753commit 2ec1383a68bf9f4baf7125a7d6544167f38d8d62
5754Merge: 3abd416 158c9b6
5755Author: Matthias Hopf <mhopf@suse.de>
5756Date:   Wed Apr 18 12:46:08 2007 +0200
5757
5758    Merge branch 'master' of git://anongit.freedesktop.org/git/xcb/libxcb
5759
5760commit 158c9b6ba18b39f424bd524fceb66f3fec0d1616
5761Author: Ian Osgood <iano@quirkster.com>
5762Date:   Fri Apr 13 15:14:12 2007 -0700
5763
5764    Generate error constants as XCB_BAD_*, similar to Xlib.
5765    
5766    The previous constants remain for compatibility, but should be deprecated.
5767
5768commit f7279d8c8ae33d6e3029e85b14ed8495d3b00e30
5769Merge: 0925e47 eaa380e
5770Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5771Date:   Thu Apr 12 15:48:44 2007 +0200
5772
5773    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
5774
5775commit 0925e470b2abf8837c62e81428660a01bbb990fa
5776Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5777Date:   Thu Apr 12 15:46:05 2007 +0200
5778
5779    remove the 7th bit of the response_type for the event loops
5780
5781commit 3abd41625c7d6db6d01f3167d6bac2b7481965cf
5782Author: Matthias Hopf <mhopf@suse.de>
5783Date:   Wed Apr 11 17:31:31 2007 +0200
5784
5785    Update autogen.sh to one that does objdir != srcdir
5786
5787commit eaa380efefd347abcd11a6c24c008686beaf8257
5788Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5789Date:   Tue Apr 10 11:56:06 2007 -0400
5790
5791    Modify new attribute from previous patch so that it is necessary only on
5792    extensions with split names.  Tested with diff and found no difference with
5793    previous stylesheet header-file output.
5794
5795commit 91be36f845352dea7838853d0f98ae144fe571a4
5796Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
5797Date:   Thu Mar 29 12:28:07 2007 -0700
5798
5799    Replaces special-casing in c-client stylesheet with support
5800    for new attribute.  Tested with diff and found no difference with
5801    previous stylesheet header-file output.
5802
5803commit 4a60950b7433eb41e08cb0c74dc8ced8f92fd78a
5804Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5805Date:   Tue Feb 27 13:35:02 2007 +0100
5806
5807    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 ?
5808
5809commit bca41cdcdb11e2a610d7b7a3d077cbb77bc81e45
5810Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5811Date:   Wed Feb 7 20:08:27 2007 +0100
5812
5813    and make the html code valid...
5814
5815commit acefe83bea9b2b17a956d68f5be504866d34196f
5816Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5817Date:   Wed Feb 7 20:04:10 2007 +0100
5818
5819    font part
5820
5821commit 8a8c1fa184939ef23f96421990c171b49d16ee33
5822Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5823Date:   Wed Feb 7 18:57:46 2007 +0100
5824
5825    no more xid or id fields
5826
5827commit 6cedaece0e294d39fe090b87e2f60569b40a4554
5828Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5829Date:   Tue Feb 6 09:31:54 2007 +0100
5830
5831    add the complete cursor example. Make the html code valid
5832
5833commit ab22a4d61665bbe637b8c1f349fcfaf04e386e88
5834Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5835Date:   Mon Jan 22 11:40:15 2007 +0100
5836
5837    add doxygen doc for the *_end functions
5838
5839commit 342e9cefcc7d942714488f6b424fee7a5bdc741f
5840Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5841Date:   Sat Jan 13 11:20:55 2007 +0100
5842
5843    fix all the occurence where a drawable where considered as a union
5844
5845commit 70a72f65e438888a5530e9911c36aad68833790f
5846Author: Josh Triplett <josh@freedesktop.org>
5847Date:   Sun Dec 10 21:10:06 2006 -0800
5848
5849    Add autogen.sh to EXTRA_DIST.
5850
5851commit 867ae5eb5c19091365daad621b923f56ff699415
5852Author: Josh Triplett <josh@freedesktop.org>
5853Date:   Sun Dec 10 21:09:30 2006 -0800
5854
5855    Add tools/* to EXTRA_DIST.
5856
5857commit 67af2d24e4c46580479570cf09586a54b84b1b63
5858Author: TORRI Vincent <torri@doursse.(none)>
5859Date:   Tue Nov 28 20:26:54 2006 +0100
5860
5861    add doc tag for the _next functions. It creates doxygen doc in the header files for these functions
5862
5863commit 4c8777f87a28ff5bf45cbdddce509163dbcf8137
5864Merge: e624cca f486075
5865Author: TORRI Vincent <torri@doursse.(none)>
5866Date:   Tue Nov 28 20:15:27 2006 +0100
5867
5868    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
5869
5870commit e624cca7df4c2c5bc5e44af1e851e5a0d17c88bd
5871Author: TORRI Vincent <torri@doursse.(none)>
5872Date:   Tue Nov 28 20:14:16 2006 +0100
5873
5874    add some output informations to configure script
5875
5876commit f486075fa093846c3f32b3e4b9624c039ea26ba4
5877Author: Tilman Sauerbeck <tilman@code-monkey.de>
5878Date:   Fri Nov 24 15:48:08 2006 +0100
5879
5880    If enable_build_docs is "no", we don't even try to look for doxygen.
5881    
5882    Also set BUILD_DOCS in that branch. Now the disable-build-docs works
5883    as expected.
5884
5885commit 2e8e6debac39864450c0a69633086ad92459c25e
5886Author: Tilman Sauerbeck <tilman@code-monkey.de>
5887Date:   Fri Nov 24 15:43:39 2006 +0100
5888
5889    Fixed evaluation of the disable-build-docs argument.
5890    
5891    Now at least enable_build_docs is set correctly.
5892
5893commit 19dfaf93ac1b5e9d3ce09b1f8e2338e53a5d7324
5894Author: Ian Osgood <iano@quirkster.com>
5895Date:   Sun Nov 26 09:26:32 2006 -0800
5896
5897    Fix unit tests for FreeBSD
5898    
5899    putenv() string must contain '='
5900    environment failure test is invalid if argument is NULL
5901
5902commit 98e2a5617ef1c9955b3b5553224c34f55c7c5d29
5903Author: Josh Triplett <josh@freedesktop.org>
5904Date:   Sat Nov 25 22:27:33 2006 -0800
5905
5906    Use substitition variables in xcb-xinerama.pc.in, not instances of their values
5907    
5908    xcb-xinerama.pc.in looked more like a generated .pc file; replace specific
5909    instances of values provided by an invocation of configure with the general
5910    substitution variables configure replaces.
5911
5912commit e74cdcd02e6814222a76c0a237efca16c423bb26
5913Author: Ian Osgood <iano@quirkster.com>
5914Date:   Sat Nov 25 11:00:14 2006 -0800
5915
5916    Bug #9119: test xcb_popcount
5917
5918commit 05d23a724d4dde42b11d6e9dec9ccaf5a516e287
5919Author: Jamey Sharp <jamey@minilop.net>
5920Date:   Fri Nov 24 16:22:13 2006 -0800
5921
5922    libXau didn't have a correct pkg-config file until 0.99.2: fail if an older version is found.
5923
5924commit 3360d0c79e98cf6b7f30b2d84f117aea0a28595d
5925Author: Jamey Sharp <jamey@minilop.net>
5926Date:   Fri Nov 24 13:24:05 2006 -0800
5927
5928    NetBSD doesn't have AI_ADDRCONFIG: use it only if it's available.
5929
5930commit 818811a7ac660e46d0dca1cbf9e53ad0475af330
5931Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
5932Date:   Fri Nov 24 12:33:15 2006 +0100
5933
5934    replace all the _new functions with xcb_generate_id. repalce X11/XCB with xcb. Fix a description of the default background of a window
5935
5936commit 27f98afc2f7039f8546d6cab5a72f609c72299d6
5937Author: Josh Triplett <josh@freedesktop.org>
5938Date:   Thu Nov 23 07:30:51 2006 -0800
5939
5940    Release libxcb 1.0
5941
5942commit 11738b2af0d8bfcf5b2f7c0d3e6ade1a14866b94
5943Author: Diego 'Flameeyes' Pettenò <flameeyes@gentoo.org>
5944Date:   Thu Nov 23 15:15:30 2006 +0100
5945
5946    Avoid race condition when using multiple make jobs
5947    
5948    Avoid race condition when symlinking XML files.
5949    When declaring a rule with many files as target, the rule is called
5950    when any of them is requested, resulting in multiple for loops happening
5951    during a make process using more than one job.
5952    Also, use '$(LN_S) -f' rather than removing and recreating a file,
5953    that one should be as supported as 'rm -f' and requires one less command.
5954
5955commit 30c768b322f613d697a61997e081a19aecd0b07f
5956Author: Josh Triplett <josh@freedesktop.org>
5957Date:   Thu Nov 23 02:11:02 2006 -0800
5958
5959    Rewrite automake's data installation rules, because they suck.
5960    
5961    Specifically, they didn't handle installing data from both srcdir and builddir.
5962    We have the tutorial in the srcdir, and build the manual in the builddir.
5963    
5964    Also, stop rebuilding the manual for each make target in the doc directory, and
5965    every time any of those targets get called.  This change now makes the manual
5966    never rebuild once built; we plan to fix that later, by rewriting the makefiles
5967    to avoid recursive make, and then making the manual depend on the source files.
5968    
5969    Commit by Jamey Sharp and Josh Triplett.
5970
5971commit af3a15838fc7dc5b4e52197854deb54da6baeaf1
5972Author: Josh Triplett <josh@freedesktop.org>
5973Date:   Thu Nov 23 00:08:30 2006 -0800
5974
5975    Rework doxygen build and install to work with srcdir != builddir
5976    
5977    The documentation generation with doxygen now works when built out of tree,
5978    with srcdir != builddir.  xcb.doxygen now gets generated from xcb.doxygen.in,
5979    so that it can use top_builddir and top_srcdir to find source and to output
5980    documentation.  Also fill in PROJECT_NUMBER from @VERSION@, now that we have
5981    it readily available via autoconf.
5982
5983commit 608058ec80edb041ca012d530b42d97474f80320
5984Author: Josh Triplett <josh@freedesktop.org>
5985Date:   Wed Nov 22 22:47:15 2006 -0800
5986
5987    Remove --with-opt and --with-debug options from configure.ac; use CFLAGS instead
5988    
5989    configure supports using custom CFLAGS, so remove the --with-opt and
5990    --with-debug options from configure.ac, and the corresponding usage of
5991    COPTFLAGS and CDEBUGFLAGS in src/Makefile.am.
5992
5993commit 1aade6a15f767c49db5c1b69b97a5b1bf8cdce57
5994Author: Josh Triplett <josh@freedesktop.org>
5995Date:   Wed Nov 22 21:49:52 2006 -0800
5996
5997    Check for getaddrinfo rather than gethostbyname in configure.ac
5998    
5999    Since the addition of IPv6 support, we now use getaddrinfo rather than
6000    gethostbyname; update configure.ac accordingly.
6001
6002commit 183c2ba4cca71f4b12f106f801f0bd30c4c527f1
6003Author: Jamey Sharp <jamey@minilop.net>
6004Date:   Wed Nov 22 21:27:41 2006 -0800
6005
6006    xcb-proto has no libraries or headers, so don't use XCBPROTO_CFLAGS or XCBPROTO_LIBS.
6007
6008commit e54dfd73a9652bbc103f3f874abc303aed4b5d76
6009Author: Jamey Sharp <jamey@minilop.net>
6010Date:   Wed Nov 22 20:29:04 2006 -0800
6011
6012    Use pthread-stubs as needed, and list xdmcp in Requires.private when XCB is built to use it.
6013
6014commit 3de6ab7b786775d9e7df1523c27cdfe3ffd3a25f
6015Author: Josh Triplett <josh@freedesktop.org>
6016Date:   Tue Nov 21 20:29:34 2006 -0800
6017
6018    Replace uses of "long" with uint32_t, and similar; fixes 64-bit bugs
6019
6020commit 7fbfebaa3fb3a5ca2d2a307a0a5c40c015e18115
6021Author: Ian Osgood <iano@quirkster.com>
6022Date:   Tue Nov 21 09:04:37 2006 -0800
6023
6024    Fix IP6 work for FreeBSD/Mac.
6025
6026commit 684b8271a4539527daa15da82ad0cc302fb44727
6027Author: Josh Triplett <josh@freedesktop.org>
6028Date:   Tue Nov 21 00:57:49 2006 -0800
6029
6030    Support XDM-AUTHORIZATION-1 on IPv6
6031    
6032    Commit by Jamey Sharp and Josh Triplett.
6033
6034commit d6abe93b06c421b78e92d76ceb5ca181e3adff31
6035Author: Josh Triplett <josh@freedesktop.org>
6036Date:   Mon Nov 20 23:25:41 2006 -0800
6037
6038    Refactor auth code to get display number from xcb_connect
6039    
6040    Change xcb_connect to pass the display number to _xcb_get_auth_info, which
6041    passes it to get_authptr.  This allows get_authptr to stop hacking the display
6042    number out of the sockaddrs of various address families, such as
6043    port - X_TCP_PORT, or the number after the last X in the UNIX socket path. This
6044    also removes a portability bug introduced during the IPv6 changes: relying on
6045    '\0'-termination of the UNIX socket path in a sockaddr_un.
6046    
6047    Commit by Jamey Sharp and Josh Triplett.
6048
6049commit 4a928de402a6e69886921fe428bbffb909c6405e
6050Author: Josh Triplett <josh@freedesktop.org>
6051Date:   Mon Nov 20 22:26:40 2006 -0800
6052
6053    Only use AI_NUMERICSERV if defined.
6054
6055commit 907f8c8c4906cbd2352f3bbddfe144ad81a2cfa8
6056Author: Josh Triplett <josh@freedesktop.org>
6057Date:   Mon Nov 20 22:16:24 2006 -0800
6058
6059    Support authentication for IPv6 connections
6060    
6061    Support AF_INET6 in get_authptr, and refactor to use common code for IPv4 and
6062    v4-mapped IPv6 addresses.
6063    
6064    Commit by Jamey Sharp and Josh Triplett.
6065
6066commit 48776ce233bf77fbaddbe972d2356bca69094239
6067Author: Josh Triplett <josh@freedesktop.org>
6068Date:   Mon Nov 20 17:53:30 2006 -0800
6069
6070    Support displays with IPv6 addresses or hosts which resolve to IPv6 addresses
6071    
6072    xcb_parse_display already correctly handled IPv6 displays.  Now, _xcb_open_tcp
6073    uses getaddrinfo, and correctly connects to IPv6 displays.  Displays can use
6074    bare IPv6 addresses, square-bracketed IPv6 addresses, or hostnames which
6075    resolve to IPv6 addresses.
6076    
6077    Since IPv6 addresses may include colons, including at the end, revise the
6078    DECnet display parsing code to avoid triggering on IPv6 addresses.
6079    
6080    Authorization may not work with IPv6 connections yet.
6081    
6082    This commit brought to you by the (display) number ::1:1.1, the letter X,
6083    the Gobby collaborative editor, Josh Triplett, and Jamey Sharp.
6084
6085commit 6c3a8db3f69aebf2b5c1330ec6c11b3138f9bdfe
6086Author: Josh Triplett <josh@freedesktop.org>
6087Date:   Mon Nov 20 17:48:49 2006 -0800
6088
6089    Add new xcb_parse_display test cases, most related to IPv6 addresses
6090
6091commit 3d9bb02012b35fd51ed3352d0bd974a0f6e439d8
6092Author: Jamey Sharp <jamey@minilop.net>
6093Date:   Sun Nov 19 21:48:27 2006 -0800
6094
6095    Bug #5958: Also zero out the implicit pad byte in empty requests.
6096
6097commit 3fa50020d2d9b78e22cf3597f4688c4e22df5121
6098Author: Jamey Sharp <jamey@minilop.net>
6099Date:   Sun Nov 19 18:31:48 2006 -0800
6100
6101    Remove support for the <localfield> tag: nothing needs it.
6102    
6103    Only one use of <localfield> remained, for a list length expression in
6104    xv.xml. List length parameters that don't actually appear in the
6105    protocol should be left implicit: if no length expression is given, then
6106    a localfield will be automatically created by c-client.xsl.
6107
6108commit 16e7328b4a366b29d54696aafffe62c92e0e0130
6109Author: Jamey Sharp <jamey@minilop.net>
6110Date:   Sun Nov 19 17:56:32 2006 -0800
6111
6112    Bug #5958: zero out padding bytes in requests.
6113
6114commit 6eee1707ea7bc303620e39911706d5fb71fa16b0
6115Author: Jamey Sharp <jamey@minilop.net>
6116Date:   Sun Nov 19 00:30:37 2006 -0800
6117
6118    htmldir was introduced in autoconf 2.59c: set a reasonable value for earlier versions.
6119
6120commit da4d56ef5a880eb24014a141e6e16668ab51f180
6121Author: Jamey Sharp <jamey@minilop.net>
6122Date:   Sun Nov 12 15:30:10 2006 -0800
6123
6124    Provide xcb_prefetch_maximum_request_length counterpart to xcb_get_maximum_request_length.
6125
6126commit aedfa1fe1d91a10ccfe3ee6ac6b7a25885623dc6
6127Author: Jamey Sharp <jamey@minilop.net>
6128Date:   Sat Nov 18 21:30:12 2006 -0800
6129
6130    Fix Doxygen warnings.
6131    
6132    Specifically, fixes these two warnings which were emitted for every
6133    generated source file:
6134    
6135    * Warning: end of file while inside a group
6136    * Warning: group XCB_BigRequests_API already documented. Skipping documentation.
6137
6138commit 44a2160c95f6c90d461d66c0278348c62d2d77fa
6139Author: Jamey Sharp <jamey@minilop.net>
6140Date:   Sat Nov 18 19:36:48 2006 -0800
6141
6142    Make doc installation work and simplify doc/Makefile.am.
6143    
6144    Don't override the default htmldir with an unquoted copy.
6145    
6146    Don't suppress `make` echoing on Doxygen commands.
6147    
6148    Ensure the tutorial is always installed even if Doxygen isn't
6149    available.
6150    
6151    Take better advantage of the automake installation infrastructure.
6152
6153commit 142fe9a6f2b13d42b0e4baaa25155067f45d065b
6154Author: Jamey Sharp <jamey@minilop.net>
6155Date:   Sat Nov 18 16:51:01 2006 -0800
6156
6157    Turn off Doxygen warnings for undocumented API, for now.
6158    
6159    But warn about absent parameter documentation for functions that are
6160    otherwise documented.
6161
6162commit 3634299ecc5223e8505d19468cf04770ac08ec14
6163Author: Jamey Sharp <jamey@minilop.net>
6164Date:   Thu Nov 16 17:02:17 2006 -0800
6165
6166    Switch from the old AM_PATH_CHECK macro to pkg-config.
6167    
6168    check 0.9.4 is now required to build XCB's unit tests.
6169    
6170    The version that we were requiring was not actually new enough to let
6171    our unit tests compile, and the AM_PATH_CHECK macro is now considered
6172    deprecated. We know that versions of check using pkg-config are new
6173    enough to work, and the check dependency was optional anyway, so we've
6174    dropped support for older versions.
6175
6176commit 82762555a1af0b170d01e2042eefe34086e6c132
6177Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6178Date:   Sun Nov 5 18:24:30 2006 +0100
6179
6180    add Makefile.am for building/installing doxygen doc
6181
6182commit f32c7131cc89c64e09e9a88c5f859824c0dfa54f
6183Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6184Date:   Sun Nov 5 18:22:55 2006 +0100
6185
6186    add xcb.doxygen to EXTRA_DIST
6187
6188commit e3b7009dc0be6515d23fbda9c744082506df4cb0
6189Merge: 0262cd6 32cd2d2
6190Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6191Date:   Sun Nov 5 18:21:34 2006 +0100
6192
6193    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb/libxcb
6194
6195commit 0262cd63a2c10e52daa3bd71cb21a6a37017f234
6196Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6197Date:   Sun Nov 5 18:20:45 2006 +0100
6198
6199    add AM_CHECK_DOXYGEN macro
6200
6201commit 32cd2d20477ab17704f0a2696c2b3bc2ff8f42de
6202Author: Jamey Sharp <jamey@minilop.net>
6203Date:   Thu Nov 2 18:04:16 2006 -0800
6204
6205    Release libxcb 1.0 RC3 (0.9.93).
6206
6207commit 9d155160546798052533c277c877879e83f85979
6208Author: Jeremy Kolb jkolb@brandeis.edu <jkolb@glorfindel.(none)>
6209Date:   Sun Oct 29 19:31:09 2006 -0500
6210
6211    Added initial doxygen generation stuff.  This should probably be cleaned up later.
6212    
6213    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.
6214
6215commit 8eba8690adac2c4425f4cac490c77f9f3b1da6f7
6216Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6217Date:   Fri Oct 20 07:00:15 2006 +0200
6218
6219    fix doxy doc
6220
6221commit a5529a4f2f1a9ed117e3465f0e422b707db21c68
6222Author: Jamey Sharp <jamey@minilop.net>
6223Date:   Wed Oct 18 23:28:27 2006 -0700
6224
6225    Add doxygen config file.
6226    
6227    Thanks to Jeremy Kolb <jkolb@brandeis.edu> for the initial version.
6228
6229commit 094248405fd9c0957b92f13acb92461ab61b2c3f
6230Author: Jamey Sharp <jamey@minilop.net>
6231Date:   Wed Oct 18 23:23:00 2006 -0700
6232
6233    Quit treating xproto specially: handle it like all the extensions.
6234
6235commit 6bc0b37303f26faf6fbdcbbee444c227e83a329c
6236Author: Jamey Sharp <jamey@minilop.net>
6237Date:   Wed Oct 18 23:21:39 2006 -0700
6238
6239    Ignore generated xinerama files.
6240
6241commit 5abb10d63ffebbe34054ae17bbdfd142e0b36396
6242Author: Josh Triplett <josh@freedesktop.org>
6243Date:   Mon Oct 16 05:11:57 2006 -0700
6244
6245    Also test xcb_parse_display with NULL argument and display in $DISPLAY
6246
6247commit 9d619d14e347b9dd1d544627060023a4bca2452b
6248Author: Jamey Sharp <jamey@minilop.net>
6249Date:   Sun Oct 15 12:34:30 2006 -0700
6250
6251    Match only XCB-namespaced XID generators when converting to xcb_generate_id.
6252    
6253    Thanks to Vincent Torri for pointing out that I screwed up. :-)
6254
6255commit 5ec55dc9c0c51137d89dd870ede05cc3268f75ae
6256Author: Zephaniah E. Hull <warp@aehallh.com>
6257Date:   Sat Oct 14 19:52:23 2006 -0700
6258
6259    Define and use constants for opcode numbers.
6260    
6261    Hard coding the opcode numbers in the function just makes it harder to figure
6262    out what's going on, but much more to the point, not defining the opcodes in
6263    the header makes it impossible to use the generated headers instead of the
6264    x11proto headers in the server.
6265    
6266    The name I settled on is very simple, for an extension by the name of xconf,
6267    and a request by the name of list_devices, we get XCB_XCONF_LIST_DEVICES.  If
6268    this somehow causes problems, we can probably add a _OP somewhere in there,
6269    but.
6270    
6271    Acked-by: Jamey Sharp <jamey@minilop.net>
6272    Closes: #8641
6273
6274commit c1504691ec2786594e8ffc33f3962f460f041971
6275Author: Jeremy Kolb <jkolb@brandeis.edu>
6276Date:   Thu Oct 12 23:52:59 2006 -0400
6277
6278    Add library support for xcb-xinerama.
6279    
6280    This is version 1.1 of the Xinerama (PanoramiX) extension.
6281
6282commit 7f74dd6b643f7769701c31587bf3cfa9c5a8e980
6283Author: Josh Triplett <josh@freedesktop.org>
6284Date:   Sun Oct 8 18:16:14 2006 -0700
6285
6286    Add note to xcbxlib.h that nothing except Xlib/XCB should use it.
6287
6288commit 431f210bbb20e39a24ac879af5f2a8a7744e6a6c
6289Author: Jamey Sharp <jamey@minilop.net>
6290Date:   Sat Oct 7 19:50:55 2006 -0700
6291
6292    Release libxcb 1.0 RC2 (0.9.92).
6293
6294commit e0fac22caaf27b3e461807f8c563d0457938baa6
6295Author: Jamey Sharp <jamey@minilop.net>
6296Date:   Sat Oct 7 18:35:53 2006 -0700
6297
6298    Bugfix: make Plan 7 'checked' requests work correctly.
6299    
6300    The initial implementation of Plan 7 dumped all X errors into the event
6301    queue, because the record of a pending reply was pruned too early if an
6302    error occurred in place of the expected reply.
6303
6304commit 7b84d8b650a611d4d76083340e50cb7ea815014d
6305Author: Ian Osgood <iano@quirkster.com>
6306Date:   Sat Oct 7 14:09:29 2006 -0700
6307
6308    Document xcb_generate_id.
6309
6310commit 9e10819a678970928c9f9a1fffb4ba00f5ac7c57
6311Author: Josh Triplett <josh@freedesktop.org>
6312Date:   Sat Oct 7 13:16:43 2006 -0700
6313
6314    Move xcb_generate_id from xcbext.h to xcb.h
6315    
6316    Since extensions no longer provide type-specific XID-generation functions,
6317    xcb_generate_id now forms part of the xcb client API, rather than the
6318    extension API; move it from xcbext.h to xcb.h accordingly.
6319
6320commit b9e49b2a47e7388e20f8a5c009cdaf0b369e117c
6321Author: Josh Triplett <josh@freedesktop.org>
6322Date:   Sat Oct 7 01:27:26 2006 -0700
6323
6324    Stop implicitly importing xproto; goes with proto change to explicitly import it
6325
6326commit dd5ece606587d3e41058df803ab1804654195bbb
6327Author: Jamey Sharp <jamey@minilop.net>
6328Date:   Sat Oct 7 01:27:03 2006 -0700
6329
6330    Declare "struct foo" or "enum foo" as well as the typedef'd name "foo".
6331
6332commit 16ec51397e0711526457709605cf6b5ebf00b5e2
6333Author: Josh Triplett <josh@freedesktop.org>
6334Date:   Sat Oct 7 00:14:50 2006 -0700
6335
6336    Handle "xidunion" instead of "union" for XID unions like DRAWABLE and FONTABLE
6337
6338commit 7b786bd8f34e67617beda4f952a5a636a370396a
6339Author: Jamey Sharp <jamey@minilop.net>
6340Date:   Fri Oct 6 23:32:18 2006 -0700
6341
6342    api_conv.pl: replace xcb_*_new with calls directly to xcb_generate_id, now that we have no type-safety XID wrappers.
6343
6344commit c73ff37b907c8af6d03ab69cca27bc837c1cc189
6345Author: Josh Triplett <josh@freedesktop.org>
6346Date:   Fri Oct 6 20:11:19 2006 -0700
6347
6348    Refer to the "xlib lock" rather than the "IOLock" in xcbxlib.h
6349
6350commit 90eeb461be4975ff35d187ba64fedbb16ff13d15
6351Author: Josh Triplett <josh@freedesktop.org>
6352Date:   Fri Oct 6 17:44:53 2006 -0700
6353
6354    Remove XID wrapper structures and replace them with uint32_t typedefs
6355    
6356    After positive feedback from several people, we have decided to remove the XID
6357    wrapper structures that attempted to provide C type safety, and replace them
6358    with uint32_t typedefs.  Feedback has indicated that these type-safety hacks
6359    generated more trouble than help.
6360    
6361    We will bump the libxcb soname at the next release.
6362
6363commit 9bd2c0c58fbe4c994e76619c2d67e46ed8b43030
6364Author: Jamey Sharp <jamey@minilop.net>
6365Date:   Fri Oct 6 16:47:50 2006 -0700
6366
6367    Remove xcb_get_io_lock from the Xlib-specific API: it is no longer used.
6368
6369commit 34168ab549fdf7c6c8a6fbe39824116698d01b12
6370Author: Jamey Sharp <jamey@minilop.net>
6371Date:   Fri Oct 6 16:12:04 2006 -0700
6372
6373    Remove the 'int *error' out-parameter for xcb_poll_for_event.
6374
6375commit 40589db8124b8c72894deb86a825c6117b0a2cd2
6376Author: Jamey Sharp <jamey@minilop.net>
6377Date:   Wed Oct 4 15:01:00 2006 -0700
6378
6379    Add xcb_xlib_lock and xcb_xlib_unlock, a special-purpose two-level recursive lock just for libX11.
6380
6381commit 57b0cd8fea498a32ff2322583c7278d5e86aa4e8
6382Author: Jamey Sharp <jamey@minilop.net>
6383Date:   Wed Oct 4 12:23:45 2006 -0700
6384
6385    Factor out pthread_mutex_lock and unlock calls for the iolock.
6386
6387commit e7f473afbd02c87cc6b1fc9c7c240d6c5cc26763
6388Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6389Date:   Sat Sep 30 13:06:58 2006 +0200
6390
6391    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.
6392
6393commit 408e74538676230e289d79c6b0195b464bfc6606
6394Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6395Date:   Sat Sep 30 12:45:12 2006 +0200
6396
6397    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)
6398
6399commit 15993fccd0b0cba870b793345fb5c5ec13ece0c8
6400Author: Jamey Sharp <jamey@minilop.net>
6401Date:   Mon Sep 25 05:12:07 2006 -0700
6402
6403    Release libxcb 1.0 RC1 (0.9.91).
6404
6405commit f7cd80142fa83d531850df2c2a3b4a514129170b
6406Author: Josh Triplett <josh@freedesktop.org>
6407Date:   Mon Sep 25 05:10:37 2006 -0700
6408
6409    Stop installing the protocol descriptions for extensions to an extensions/
6410    subdirectory
6411
6412commit 0d7beff580232d4b7adfc840bcd1cae894b57682
6413Author: Josh Triplett <josh@freedesktop.org>
6414Date:   Mon Sep 25 04:43:04 2006 -0700
6415
6416    Actually install xcb-xlib.pc, and ship xcb-xlib.pc.in in distribution tarballs
6417
6418commit d04e7777d3547de2cece0734c0b9be49d3ddb88c
6419Author: Jamey Sharp <jamey@minilop.net>
6420Date:   Mon Sep 25 04:14:26 2006 -0700
6421
6422    Ensure api_conv.pl is idempotent on names of constants: XCB_[A-Z0-9_]+.
6423
6424commit 0bccf7d2eed45ba15a012d090b9b950075e75803
6425Author: Jamey Sharp <jamey@minilop.net>
6426Date:   Mon Sep 25 04:03:01 2006 -0700
6427
6428    Move header files to $includedir/xcb (generally $prefix/include/xcb).
6429
6430commit 9100981d359f6366ab92b456e6069bac89afeb15
6431Author: Josh Triplett <josh@freedesktop.org>
6432Date:   Mon Sep 25 02:02:16 2006 -0700
6433
6434    Split the Xlib compatibility functions into a separate library libxcb-xlib
6435    
6436    We don't want to have to change the libxcb soname if we later manage to remove
6437    the Xlib compatibility functions, and nothing except Xlib should ever use
6438    them, so split them into a separate library.
6439
6440commit 810dfb886247dc4484886aff2e2e7d80ea945de4
6441Author: Jamey Sharp <jamey@minilop.net>
6442Date:   Mon Sep 25 02:10:44 2006 -0700
6443
6444    Move debian/ directory to new branch "debian".
6445
6446commit 46b7dbdd9306235ae04da9d0e3431a1b58f89048
6447Author: Jamey Sharp <jamey@minilop.net>
6448Date:   Mon Sep 25 02:00:23 2006 -0700
6449
6450    We no longer need xproto: do not list it as a dependency.
6451
6452commit cf80bab8e6410915ad27a61cef3bb5f456c7d80f
6453Author: Jamey Sharp <jamey@minilop.net>
6454Date:   Mon Sep 25 01:21:42 2006 -0700
6455
6456    Lowercase library names to correspond with the new API.
6457
6458commit a1a7646d5a791c2e918bc1a9f46d64334d5ab575
6459Author: Ian Osgood <iano@quirkster.com>
6460Date:   Sun Sep 24 14:10:20 2006 -0700
6461
6462    Add XCBAllPlanes from xcb-image util library.
6463
6464commit dfbde9a4e972ed9bbd701fd6f89c3a6e6f641740
6465Author: Josh Triplett <josh@freedesktop.org>
6466Date:   Sun Sep 24 03:15:39 2006 -0700
6467
6468    Integrate top-level .gitignore into .gitignore for each subdirectory
6469    
6470    In preparation for the repository split, move the relevant contents of the
6471    top-level .gitignore into the .gitignore for each immediate subdirectory.
6472
6473commit 905379e4474c0137e5dd535798ae0afb07070df6
6474Author: TORRI Vincent <torri@alf94-3-82-66-248-160.fbx.proxad.net>
6475Date:   Sun Sep 24 08:51:20 2006 +0200
6476
6477    now, I know how to use api_conv.pl :)
6478
6479commit 16516a5e94205edffddde546abc833106c68068e
6480Author: Josh Triplett <josh@freedesktop.org>
6481Date:   Sat Sep 23 14:44:15 2006 -0700
6482
6483    Remove the xcb_[extension]_init functions; use xcb_get_extension_data directly
6484
6485commit 9691890529fddc051c15e191a8a5f06017514c1c
6486Author: Josh Triplett <josh@freedesktop.org>
6487Date:   Sat Sep 23 14:17:52 2006 -0700
6488
6489    More fixups for incorrect API conversions by api_conv.pl
6490
6491commit 0a867d72337afa88f4ab453b6d3fbb519368c308
6492Author: Josh Triplett <josh@freedesktop.org>
6493Date:   Sat Sep 23 13:59:40 2006 -0700
6494
6495    Convert the XCB test suite to the new API
6496
6497commit 63b38f31974b0a61c923a62a2d224e7024a0fc2b
6498Author: Jamey Sharp <jamey@minilop.net>
6499Date:   Sat Sep 23 13:51:10 2006 -0700
6500
6501    Apply const-names.xsl to the un-renamed XML; replace the XSL with the result.
6502
6503commit 86a4c0cc284366bbb01898a77df360278d5a8194
6504Author: Josh Triplett <josh@freedesktop.org>
6505Date:   Sat Sep 23 13:33:27 2006 -0700
6506
6507    We do not conflict with Xmd.h anymore; remove the include hack from xcb_auth.c
6508
6509commit 4ff12a1a94895837dc85ac9e37fd1a5f40819cf3
6510Author: Josh Triplett <josh@freedesktop.org>
6511Date:   Sat Sep 23 13:31:13 2006 -0700
6512
6513    Fix some mis-conversions by api_conv.pl, and remove the now-unused Xmd types
6514
6515commit a3bd6f4760b5b3f5f360a690920839646e2b9d06
6516Author: Josh Triplett <josh@freedesktop.org>
6517Date:   Sat Sep 23 12:22:22 2006 -0700
6518
6519    The Great XCB Renaming
6520    
6521    Rename API to follow a new naming convention:
6522    * XCB_CONSTANTS_UPPERCASE_WITH_UNDERSCORES
6523    * xcb_functions_lowercase_with_underscores
6524    * xcb_types_lowercase_with_underscores_and_suffix_t
6525    * expand all abbreviations like "req", "rep", and "iter"
6526    
6527    Word boundaries for the names in the protocol descriptions fall:
6528    * Wherever the protocol descriptions already have an underscore
6529    * Between a lowercase letter and a subsequent uppercase letter
6530    * Before the last uppercase letter in a string of uppercase letters followed
6531      by a lowercase letter (such as in LSBFirst between LSB and First)
6532    * Before and after a string of digits (with exceptions for sized types like
6533      xcb_char2b_t and xcb_glx_float32_t to match the stdint.h convention)
6534    
6535    Also fix up some particular naming issues:
6536    * Rename shape_op and shape_kind to drop the "shape_" prefix, since otherwise
6537      these types end up as xcb_shape_shape_{op,kind}_t.
6538    * Remove leading underscores from enums in the GLX protocol description,
6539      previously needed to ensure a word separator, but now redundant.
6540    
6541    This renaming breaks code written for the previous API naming convention.  The
6542    scripts in XCB's tools directory will convert code written for the old API to
6543    use the new API; they work well enough that we used them to convert the
6544    non-program-generated code in XCB, and when run on the old program-generated
6545    code, they almost exactly reproduce the new program-generated code (modulo
6546    whitespace and bugs in the old code generator).
6547    
6548    Authors: Vincent Torri, Thomas Hunger, Josh Triplett
6549
6550commit ca72e777740f917753f3c180fdfdb55df91c09c4
6551Author: Jamey Sharp <jamey@minilop.net>
6552Date:   Sat Sep 23 01:33:45 2006 -0700
6553
6554    Special cases to agree with the conventions Josh and I have hashed out.
6555
6556commit 4168ddc13dff7bf2479c2229f42b114d75098112
6557Author: Jamey Sharp <jamey@minilop.net>
6558Date:   Fri Sep 22 22:53:17 2006 -0700
6559
6560    Handle remaining incorrect API translations in api_conv.pl.
6561    
6562    The big change is treating numbers as separate "words" in the
6563    translation, which leads to funny-looking names like xcb_char_2_b_t, but
6564    makes more sense than the alternative in other situations and was
6565    suggested on the mailing list.
6566    
6567    This version still disagrees with Vincent's last proposed XSL, but I
6568    think my output is now preferable. Hopefully Josh has been thinking
6569    along the same lines.
6570
6571commit 91568d7070e2c2400dfdbdff515f51da74ca3701
6572Author: Jamey Sharp <jamey@minilop.net>
6573Date:   Fri Sep 22 15:37:49 2006 -0700
6574
6575    Make xcb_conn.c agree that XCBSetupReq is now XCBSetupRequest.
6576
6577commit 2538acd8c89d73e8c074849c63857cb899a29bbb
6578Author: Alan Coopersmith <alan.coopersmith@sun.com>
6579Date:   Fri Sep 22 11:53:21 2006 -0700
6580
6581    List xproto.xml path explicitly instead of relying on GNU Make $< expansion
6582    
6583    (Some non-gnu makes, such as Solaris make, only recognize $< in implicit
6584     suffix rules, not explicit ones like this.)
6585
6586commit 2c8b5994b3fbba343199ef555594a32e29d8bcee
6587Author: Jamey Sharp <jamey@minilop.net>
6588Date:   Fri Sep 15 01:51:05 2006 -0700
6589
6590    Shut down the connection in all "fatal" error cases.
6591
6592commit 79e3227022ae3d66f9f3806d231fdeec2a06cc6b
6593Author: Jamey Sharp <jamey@minilop.net>
6594Date:   Fri Sep 15 01:57:53 2006 -0700
6595
6596    Add a private connection shutdown method for error cases.
6597
6598commit 7f71bf9c0f30536e85907b2c991cb7001861e1d3
6599Author: Jamey Sharp <jamey@minilop.net>
6600Date:   Fri Sep 15 01:09:27 2006 -0700
6601
6602    Make all public functions do nothing on an error connection.
6603
6604commit 0aa96bfc7abe18889cd85bfaa05b05d53e572bb1
6605Author: Jamey Sharp <jamey@minilop.net>
6606Date:   Fri Sep 15 00:39:51 2006 -0700
6607
6608    Convert connection functions to return error objects.
6609
6610commit ac17ae62fe1a3a29991e36e35eeee838ac4acb00
6611Author: Jamey Sharp <jamey@minilop.net>
6612Date:   Fri Sep 15 00:29:39 2006 -0700
6613
6614    Provide a "has error" property for XCBConnection.
6615
6616commit df7fb77d6e22be76ca73f111c586db99a60178ae
6617Author: Jamey Sharp <jamey@minilop.net>
6618Date:   Thu Sep 21 12:13:37 2006 -0700
6619
6620    Refactor XCBPollForEvent with a shorter critical section.
6621    
6622    This simplifies the patch for bug #8208 later.
6623
6624commit b08ca2b4b451a94ece20207766cd5262fd55179b
6625Author: Jamey Sharp <jamey@minilop.net>
6626Date:   Mon Sep 18 16:52:00 2006 -0700
6627
6628    XCB*Id is a variable, not a type: adjust API conventions accordingly.
6629
6630commit 3cc0ddf0e6b72ba553ebfc2b9126d62ef4013bd3
6631Author: Jamey Sharp <jamey@minilop.net>
6632Date:   Wed Sep 13 12:30:11 2006 -0700
6633
6634    Fix bug #7261: events do not signal the end of replies for that sequence number.
6635
6636commit c912187f15c560c46768acb88aae30e67b0f78ce
6637Author: Jamey Sharp <jamey@minilop.net>
6638Date:   Mon Sep 18 00:25:26 2006 -0700
6639
6640    Handle names of constants during API conversion.
6641    
6642    Use an XSLT stylesheet to get a list of all the constant names.
6643
6644commit 06fba014435cfbdd1ff284d6d513d114503d02c2
6645Author: Jamey Sharp <jamey@minilop.net>
6646Date:   Sun Sep 17 23:38:57 2006 -0700
6647
6648    Complete rewrite of api_conv.pl.
6649    
6650    Now handles all API changes except constant names, which are treated like type names.
6651
6652commit bf41af718d6a83532d1c2f63ac16e6484e8e8b68
6653Author: Jamey Sharp <jamey@minilop.net>
6654Date:   Thu Sep 14 00:04:47 2006 -0700
6655
6656    Simplify the API conversion tool without functionality changes.
6657
6658commit cdffbdd7ef9dee58b3c36ca46bb88aa187b46526
6659Author: Thomas Coppi <thisnukes4u@gmail.com>
6660Date:   Wed Sep 13 23:50:23 2006 -0700
6661
6662    Prototype API conversion tool for upcoming lowercased XCB API.
6663
6664commit a92716f1da2741fca850b3c37299e80032726276
6665Author: Jamey Sharp <jamey@minilop.net>
6666Date:   Wed Sep 13 12:15:23 2006 -0700
6667
6668    Finish removing deprecated functions. Fixes build failure (oops).
6669
6670commit 476ccc1ba3d20f3a545b84089b6fbd40576e7bf5
6671Author: Jamey Sharp <jamey@minilop.net>
6672Date:   Tue Sep 12 13:43:17 2006 -0700
6673
6674    Ignore xproto.xml that now gets copied in while building.
6675
6676commit 0d7fb3afdd308d714a97144125a69a5f4976916a
6677Author: Jamey Sharp <jamey@minilop.net>
6678Date:   Tue Sep 12 13:25:49 2006 -0700
6679
6680    Remove deprecated public API functions in preparation for 1.0 release. Xlib-specific deprecation remains.
6681
6682commit d4e768fc32ff9ce79a4259d252f4a4a4b11d5db8
6683Author: Bart Massey <bart@cs.pdx.edu>
6684Date:   Tue Sep 12 00:42:11 2006 -0700
6685
6686    Cleaned up previous fix for GNU Make stuff.
6687    Got "make distcheck" to work (whatever that does).
6688
6689commit 0c3e528563f8e8db45c340e7d050a1b42e65cb73
6690Author: Bart Massey <bart@bartfan.localdomain>
6691Date:   Mon Sep 11 23:29:10 2006 -0700
6692
6693    Removed GNU make dependency.
6694    Made extensions directory persist.
6695
6696commit 81d6fd64c09255e90104b3c6e37c9023bd0aa6ec
6697Author: Jamey Sharp <jamey@minilop.net>
6698Date:   Sat Sep 9 15:52:37 2006 -0700
6699
6700    Handle XC-MISC request failure when allocating new XIDs. (fixes #7001)
6701
6702commit f6d4fc007fc0f4bdf8f06ab0798aadce87d468d1
6703Author: Thomas Hunger <hto@arcor.de>
6704Date:   Mon Aug 14 06:59:46 2006 -0700
6705
6706    Tutorial corrections.
6707
6708commit dbd2d9689305ce5bf24aebd894551bb6b590f6e4
6709Author: Josh Triplett <josh@freedesktop.org>
6710Date:   Sun Jul 30 22:02:37 2006 -0700
6711
6712    Implement error handling plan 7.
6713    
6714    Needs improvement: should not duplicate the code of every request function.
6715
6716commit 65ed274f05ba670eb02a55b098aed141fa9611ec
6717Author: Josh Triplett <josh@freedesktop.org>
6718Date:   Sun Jul 30 20:28:16 2006 -0700
6719
6720    Parameterize context in cookie-type and request-matching template.
6721
6722commit f74487e34fd61d11a501c07030f8fd7ed8caeb1c
6723Author: Jamey Sharp <jamey@minilop.net>
6724Date:   Sun Jul 30 14:55:14 2006 -0700
6725
6726    Fix Keith's 32-bit wrap fix.
6727    
6728    Issue 0, 1, or 2 syncs as needed and do not handle 16-bit wrap until
6729    absolutely necessary.
6730
6731commit 13896d8f658b917e891aa18e0ac4906d630881bb
6732Author: Josh Triplett <josh@freedesktop.org>
6733Date:   Sun Jul 30 15:09:03 2006 -0700
6734
6735    Fix typo: s/request/sequence/
6736
6737commit 685d8c76c25fd9f236fa1a74bae61699eaa78579
6738Author: Josh Triplett <josh@freedesktop.org>
6739Date:   Sun Jul 30 14:41:19 2006 -0700
6740
6741    Add XCBRequestCheck function needed for Plan 7.
6742
6743commit b3a2f83f1e4d3567fcd4494e7bae31e99645ef85
6744Author: Eric Anholt <anholt@FreeBSD.org>
6745Date:   Fri Jul 28 15:29:05 2006 -0700
6746
6747    Restore netinet/in.h include that was removed in the DECnet commit.
6748
6749commit b7c96681b2bfc968b198058122c93050feb7927e
6750Author: Eric Anholt <anholt@FreeBSD.org>
6751Date:   Tue Jul 25 09:13:59 2006 -0700
6752
6753    Replace HAVE_* tests with just checking for (__solaris__)
6754    
6755    The installed headers can't be relying on the presence of the internal
6756    config.h defines, and it was breaking the xcb build for me as well due to
6757    config.h not being included early enough.
6758
6759commit 14faffc326b9ab1f738edb3ec7febdbe0f5bc079
6760Author: TORRI Vincent <torri@doursse.(none)>
6761Date:   Tue Jul 25 08:24:59 2006 +0200
6762
6763    fix closing <b> tag
6764
6765commit 8ba93ebb757706ae8dcf5af57c6e3812f72de80e
6766Author: Ian Osgood <iano@quirkster.com>
6767Date:   Fri Jul 7 07:18:41 2006 -0700
6768
6769    Fix build on Solaris (use inttypes.h)
6770
6771commit e0574a617b9133cd728be8ea3618ef1312eeebbf
6772Author: Ian Osgood <iano@quirkster.com>
6773Date:   Sat Jul 1 00:55:08 2006 -0700
6774
6775    Finally remove X.h from xcb.h, fix broken image tests.
6776
6777commit db2504130bc450bd328830060cb3a243dd06b52f
6778Author: Keith Packard <keithp@neko.keithp.com>
6779Date:   Sat Jul 1 00:16:07 2006 -0700
6780
6781    Switch sequence comparisons to handle 32-bit sequence number wrap.
6782    
6783    Create a macro, XCB_SEQUENCE_COMPARE, that accepts two sequence numbers and
6784    a comparison operator and correctly handles 32-bit wrap around.
6785    Rewrite all ordered sequence number comparisons to use this macro.
6786    Also, caught one error where a sequence was stored in a signed int variable.
6787    Push out a GetInputFocus request when the sequence number does wrap at 32
6788    bits so that applications cannot see sequence 0 (as that is an error
6789    indicator).
6790
6791commit 75fead5b868a0dfdc9e6fd5ef0dd37eb71761312
6792Author: Josh Triplett <josh@freedesktop.org>
6793Date:   Thu Jun 15 03:03:13 2006 -0700
6794
6795    Remove arbitrary division between xcb_types and xproto by merging
6796    xcb_types.xml into xproto.xml.
6797
6798commit 62749d54fd79b12123607599d58add126ce5de6e
6799Author: TORRI Vincent <torri@doursse.(none)>
6800Date:   Thu Jun 8 06:59:47 2006 +0200
6801
6802    fix compilation with c++ compilers. Remove some trailing spaces
6803
6804commit 829188cdd6274b5b5271dff5612b30c978a61a38
6805Author: Ian Osgood <iano@quirkster.com>
6806Date:   Tue May 30 11:54:25 2006 -0700
6807
6808    Deprecate XCBSync, move to XCBAuxSync.
6809
6810commit 936077cbc87addc914d33ab79a7d066f0f51d3ad
6811Author: Jamey Sharp <jamey@minilop.net>
6812Date:   Sun May 14 22:49:18 2006 -0700
6813
6814    Use correct word offset when testing for GetFBConfigsSGIX VendorPrivate.
6815
6816commit 442730a9a25644e6d09065cdde2f1595ea65caf3
6817Author: Josh Triplett <josh@freedesktop.org>
6818Date:   Sun May 14 22:37:55 2006 -0700
6819
6820    In the GLX workaround, use !strcmp to check for equality with "GLX", not strcmp.
6821
6822commit 2e49f58e4cd670e6bd6a0006833277cfb1da60e6
6823Author: Josh Triplett <josh@freedesktop.org>
6824Date:   Wed May 10 14:22:27 2006 -0700
6825
6826    Stop overwriting CFLAGS in xcb/src/Makefile.am.
6827
6828commit 92456577dbbe3874d6e40a1ef26b63dd405c160c
6829Author: Donnie Berkholz <spyderous@gentoo.org>
6830Date:   Sun May 7 15:44:37 2006 -0700
6831
6832    Fix dependencies on libXCB.la to not walk down from $(top_builddir); this fixes parallel builds.
6833
6834commit 4a891c6f4e49a04866b2cac7bf8bb679a256d3b4
6835Author: Josh Triplett <josh@freedesktop.org>
6836Date:   Sat Apr 29 17:02:43 2006 -0700
6837
6838    Work around the unsetting of CFLAGS in src/Makefile.am, by passing CFLAGS to make via DEB_MAKE_INVOKE in debian/rules.
6839
6840commit 140c30e522c664f0675b63301bea22e4062cd155
6841Author: Josh Triplett <josh@freedesktop.org>
6842Date:   Sat Apr 29 16:28:01 2006 -0700
6843
6844    Remove unnecessary dirs files.
6845
6846commit 105c6d67e6fa3d636068a2a4407bfa6adfafe0cc
6847Author: Josh Triplett <josh@freedesktop.org>
6848Date:   Sat Apr 29 16:09:06 2006 -0700
6849
6850    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.
6851
6852commit 911ae3fbdd12bcfa39e6bbd85a7327ddd1cb9287
6853Author: Josh Triplett <josh@freedesktop.org>
6854Date:   Sat Apr 29 15:57:28 2006 -0700
6855
6856    Set DEB_MAKE_CHECK_TARGET=check in debian/rules.
6857
6858commit 3b7cf7ecf8ae614d7cb833ac4ef9fbcde3df05ae
6859Author: Josh Triplett <josh@freedesktop.org>
6860Date:   Sat Apr 29 15:34:11 2006 -0700
6861
6862    Add a Section field for the source package in debian/control.
6863
6864commit 4e26eefcd9a035c4f7ee1d5f267095edc7df7c89
6865Author: Josh Triplett <josh@freedesktop.org>
6866Date:   Sat Apr 29 11:15:47 2006 -0700
6867
6868    Drop library version from source package name.
6869
6870commit eb3521283f91a84ce149fcc416720289ca37e29f
6871Author: Josh Triplett <josh@freedesktop.org>
6872Date:   Sat Apr 29 10:58:23 2006 -0700
6873
6874    Set distribution to experimental.
6875
6876commit 2a18842f0f93516ff726a2dfa44348410bb55626
6877Author: Josh Triplett <josh@freedesktop.org>
6878Date:   Sat Apr 29 10:56:56 2006 -0700
6879
6880    Remove Bugs field in debian/control, so bugs go to the Debian BTS.
6881
6882commit e6132aff2fdcd0ab1ef5d7ac84a3c597dc07d370
6883Author: Ian Osgood <iano@quirkster.com>
6884Date:   Fri Apr 28 15:27:09 2006 -0700
6885
6886    Move the remainder of the constants in X.h into XML enumerations.
6887    Fix xcb_auth to use one of the new enumerations.
6888
6889commit 162c7593adf1577ca8aecfcd53fd5644b6182609
6890Author: Josh Triplett <josh@freedesktop.org>
6891Date:   Fri Apr 28 00:51:53 2006 -0700
6892
6893    Change libxcbxvmc0-dev Depends to libxcbxv0-dev, not libxv0-dev.
6894
6895commit 4f9b6556e2b68d1f05ed4df9d2cc82edb9868872
6896Author: Josh Triplett <josh@freedesktop.org>
6897Date:   Thu Apr 27 23:40:55 2006 -0700
6898
6899    Update debian packaging to create library, -dev, and -dbg packages for each new extension library.
6900
6901commit b7d77a8de79493919281c4f742b1a65c1328b1c4
6902Author: Josh Triplett <josh@freedesktop.org>
6903Date:   Thu Apr 27 17:32:20 2006 -0700
6904
6905    Fix the year in COPYING.
6906
6907commit 5c35ea63db8dcc30bd84eb1ae8f12f0969bb8a63
6908Author: Josh Triplett <josh@freedesktop.org>
6909Date:   Thu Apr 27 17:20:52 2006 -0700
6910
6911    Add an explanation of libtool -version-info to src/Makefile.am, and add an explicit -version-info 0:0:0 for all libraries.
6912
6913commit 08f5cc389275a71821a62c1674622ca85792b3fe
6914Author: Josh Triplett <josh@freedesktop.org>
6915Date:   Thu Apr 27 16:24:35 2006 -0700
6916
6917    Fix tutorial to use XCBSetup rather than the previous deprecated name XCBConnSetupSuccessRep.
6918
6919commit 34016bcdab9b8373b4a8f9520b01d3179d35a315
6920Author: Josh Triplett <josh@freedesktop.org>
6921Date:   Thu Apr 27 16:15:50 2006 -0700
6922
6923    Remove execute bit on tutorial.
6924
6925commit 135cda6e9dafeef36efc8c4be49ae99b417034f2
6926Author: Josh Triplett <josh@freedesktop.org>
6927Date:   Thu Apr 27 13:47:21 2006 -0700
6928
6929    Fix a comment for the renaming of XCBConnSetupSuccessRep to XCBSetup, and fix another comment which had a copy-paste error.
6930
6931commit 17c3448f58c1aaf1d89b706dfa00e697061115c7
6932Author: Josh Triplett <josh@freedesktop.org>
6933Date:   Thu Apr 27 12:33:41 2006 -0700
6934
6935    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.
6936
6937commit ef18582d83fac146b1a689bd60d580819dba0ffd
6938Author: Josh Triplett <josh@freedesktop.org>
6939Date:   Thu Apr 27 12:04:52 2006 -0700
6940
6941    Fix typo and rephrase checking message.
6942
6943commit 3e213a1becd3081e905e3711ff3fe3e4865cd7dd
6944Author: Jamey Sharp <jamey@minilop.net>
6945Date:   Thu Apr 27 10:34:22 2006 -0700
6946
6947    Declare arguments const in the various structure accessors, so XCBGetSetup is not so painful to use.
6948
6949commit dd932e025b2de0fa33b65971288c63679238e08a
6950Author: Jamey Sharp <jamey@minilop.net>
6951Date:   Wed Apr 26 23:19:16 2006 -0700
6952
6953    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.
6954
6955commit b825f3385becccfcde34626f872721301e82f960
6956Author: Josh Triplett <josh@freedesktop.org>
6957Date:   Wed Apr 26 11:23:32 2006 -0700
6958
6959    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.
6960    
6961    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.
6962
6963commit 1476dcecfd3e817236ae996314c10482aae8e3e9
6964Author: Ian Osgood <iano@quirkster.com>
6965Date:   Wed Apr 26 10:50:31 2006 -0700
6966
6967    Changes to makefile to build libXCBxtest
6968
6969commit bb8cf58015bae083dd77f8679f9a8299603c58a6
6970Author: Patrick Caulfield <patrick@tykepenguin.com>
6971Date:   Mon Apr 24 08:29:18 2006 -0700
6972
6973    Add support for DECnet. Still needs configure-script options to enable.
6974
6975commit 771761ccaad31d029d470dde84279e94494310b6
6976Author: Jamey Sharp <jamey@minilop.net>
6977Date:   Thu Apr 20 11:51:01 2006 -0700
6978
6979    Minor performance improvement: do not call _xcb_in_expect_reply unless it is needed. It is not often needed.
6980
6981commit 8953a14f2b1518042ed0745574e22fa2adfb6cac
6982Author: Jamey Sharp <jamey@minilop.net>
6983Date:   Wed Apr 19 22:48:10 2006 -0700
6984
6985    Add <sys/select.h> to xcb_in.c to fix bug #6122.
6986
6987commit 05a66af895442b9fceb96c9130e77694927eabaf
6988Author: Alan Coopersmith <alan.coopersmith@sun.com>
6989Date:   Wed Apr 19 21:51:33 2006 -0700
6990
6991    Bugfix: xcb_conn.c included <sys/fcntl.h> instead of the POSIX-standard <fcntl.h>.
6992
6993commit f705456744fe4beb193d27eb64fa9157102db753
6994Author: Alan Coopersmith <alan.coopersmith@sun.com>
6995Date:   Wed Apr 19 21:40:42 2006 -0700
6996
6997    Only set CWARNFLAGS to gnu flags if $GCC is set. Otherwise check for Sun compiler and use its enhanced warning flag.
6998
6999commit 6f369fde3c676e2b1a67bd71923b61942991d726
7000Merge: 922cb61 f090da9
7001Author: Ian Osgood <iano@quirkster.com>
7002Date:   Wed Apr 19 20:45:31 2006 -0700
7003
7004    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7005
7006commit f090da98f367ed869fd9277d2fef22555be0f91d
7007Author: Jamey Sharp <jamey@minilop.net>
7008Date:   Wed Apr 19 20:31:20 2006 -0700
7009
7010    Remove the last goto in XCB: XCBWaitForReply now permits multiple threads to force the same cookie.
7011
7012commit d5347485a55e58381781d803e19bfdd982a4685b
7013Author: Jamey Sharp <jamey@minilop.net>
7014Date:   Wed Apr 19 20:23:37 2006 -0700
7015
7016    Restructure XCBWaitForReply to eliminate two gotos.
7017
7018commit d5ab03b4b71648bf0a06e42e3c288a68c46ba497
7019Author: Jamey Sharp <jamey@minilop.net>
7020Date:   Wed Apr 19 20:15:15 2006 -0700
7021
7022    Fixed poll_for_reply, added comments, and refactored XCBWaitForReply to call poll_for_reply.
7023
7024commit 7667adbc631119ec39f3ef5a316aec42dbf5f393
7025Author: Jamey Sharp <jamey@minilop.net>
7026Date:   Wed Apr 19 16:49:32 2006 -0700
7027
7028    Add XCBPollForReply and deprecate XCBGetRequestRead and XCBGetQueuedRequestRead.
7029
7030commit 53971ea183d9d1dcfbaec18b135e49c9c118fabb
7031Merge: 71de16f 8275ac3
7032Author: Josh Triplett <josh@freedesktop.org>
7033Date:   Sun Apr 16 11:23:52 2006 -0700
7034
7035    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7036
7037commit 71de16fac2a145d5ef8069d2d28d7c32cea603cf
7038Author: Josh Triplett <josh@freedesktop.org>
7039Date:   Sun Apr 16 11:05:50 2006 -0700
7040
7041    Improve package descriptions.
7042
7043commit e92bde6e5152c6d0c4efa9240604e75178c1a3db
7044Author: Josh Triplett <josh@freedesktop.org>
7045Date:   Sun Apr 16 09:59:13 2006 -0700
7046
7047    Change Depends on x-dev to x11proto-core-dev.
7048
7049commit eca61f6b5e9df7321222499a1f660f6eb7e6112e
7050Author: Josh Triplett <josh@freedesktop.org>
7051Date:   Sun Apr 16 09:25:36 2006 -0700
7052
7053    Add Build-Depends on libxdmcp-dev.
7054
7055commit a0057d7a48b90b8f11fc9d5c82b5b8d800c34db5
7056Author: Josh Triplett <josh@freedesktop.org>
7057Date:   Sun Apr 16 09:24:01 2006 -0700
7058
7059      * Debian X11R7 transition:
7060        * Change Build-Depends on x-dev to x11proto-core-dev.
7061        * Install headers to /usr/include/X11, not /usr/X11R6/include/X11.
7062        * Pre-Depends: x11-common (>= 1:1.09).
7063
7064commit 11c62f7d9d65c10c796c2199c73c8f167e53f234
7065Author: Josh Triplett <josh@freedesktop.org>
7066Date:   Sat Apr 15 22:50:33 2006 -0700
7067
7068    Stop installing the libtool .la files.
7069
7070commit 057ae541a3a73cffd58533029292c1c721fa3162
7071Author: Josh Triplett <josh@freedesktop.org>
7072Date:   Sat Apr 15 20:44:11 2006 -0700
7073
7074    Update -dbg package handling to work with debhelper compat level 5.
7075    Increase minimum version on debhelper Build-Depends to 5.0.0.
7076
7077commit 8eedb4a487dcede0e52849e36c4da13bdf0c8b51
7078Author: Josh Triplett <josh@freedesktop.org>
7079Date:   Sat Apr 15 20:22:05 2006 -0700
7080
7081    Stop running autoreconf in debian/rules, and remove Build-Depends for
7082    autoconf, automake, and libtool.
7083
7084commit cc075990f4fc2ed09c708036569049ddd24605ac
7085Author: Josh Triplett <josh@freedesktop.org>
7086Date:   Sat Apr 15 20:21:22 2006 -0700
7087
7088    Add xcbint.h to noinst_HEADERS, so it gets distributed.
7089
7090commit cb6e1849b66c17f96d79598adb740ed16325a9c1
7091Author: Josh Triplett <josh@freedesktop.org>
7092Date:   Sat Apr 15 14:17:52 2006 -0700
7093
7094    Use screensaver.xml
7095
7096commit 91aeea2a3e72af16733ddcba037541440c0c4739
7097Author: Josh Triplett <josh@freedesktop.org>
7098Date:   Sat Apr 15 12:52:05 2006 -0700
7099
7100    Put EXTHEADERS and EXTSOURCES in order.
7101
7102commit ff38c17c48c271847d12c81cbf80142c6918dc78
7103Author: Josh Triplett <josh@freedesktop.org>
7104Date:   Sat Apr 15 12:26:03 2006 -0700
7105
7106    Split all non-essential extensions into their own separate libraries, named
7107    libXCBextname.  To use extension extname, include extname.h and link with
7108    -lXCBextname.  This allows extensions to change without bumping the main
7109    libXCB version.
7110    
7111    bigreq and xc_misc remain in libXCB, because XCB uses them internally to make
7112    big requests and to allocate XIDs, respectively.
7113
7114commit 8275ac3a4a23220b5c3b4f191a45befe2d34d6bd
7115Author: Eric Anholt <anholt@FreeBSD.org>
7116Date:   Sun Apr 9 19:51:10 2006 -0700
7117
7118    Retry a select() if it returns with EINTR.  Fixes IO errors in Xephyr, which is
7119    often interrupted by timers.
7120
7121commit 66a88ed0e556ca869ddc9df5a35e3d6446d12b02
7122Author: Eric Anholt <anholt@FreeBSD.org>
7123Date:   Sun Apr 9 19:19:12 2006 -0700
7124
7125    Remove unnecessary include.  Noticed by jamey.
7126
7127commit 922cb6137a12982ecd9e1c73ecefdcbc9e193eae
7128Author: Ian Osgood <iano@quirkster.com>
7129Date:   Wed Mar 22 17:57:57 2006 -0800
7130
7131    Work on the tutorial, and update xproto.xml to match:
7132    * Fixed grammar
7133    * Answered some TODO's and added some more
7134    * Updated X.h constants to those in xproto.h
7135    * Added enumerations used in tutorial to xproto.xml
7136    * Prefered XCBFlush to XCBSync
7137    * Corrected and refactored the "events" example
7138    
7139    I extracted the examples to test them. Where should I put them?
7140
7141commit fc577b81bfeb79bb78ee529278ed52d59d489f89
7142Author: Josh Triplett <josh@freedesktop.org>
7143Date:   Tue Mar 21 14:22:21 2006 -0800
7144
7145    Remove outdated fd.o-* entries from */debian/.gitignore (obsolete since before they came from .cvsignore).
7146
7147commit df5d8adc1f18776e4417a03b465dae9273511fb1
7148Author: Jamey Sharp <jamey@minilop.net>
7149Date:   Thu Mar 16 11:27:06 2006 -0800
7150
7151    The typedefs replacing Xmd.h conflict with Xmd.h. Here is a hacky workaround: FIXME!
7152
7153commit 5ccf7216d266f22755e40a8b9858ebd60fd48f40
7154Author: Ian Osgood <iano@quirkster.com>
7155Date:   Wed Mar 15 22:41:08 2006 -0800
7156
7157    Change <bit> syntax to <bit>n</bit> as Jamey and Josh suggested
7158     and add more button, key, and graphics enumerations.
7159
7160commit 35ecaf45be75cb78fe18dd4ea9d564b03f6cea67
7161Author: Ian Osgood <iano@quirkster.com>
7162Date:   Wed Mar 15 10:09:47 2006 -0800
7163
7164    Remove proto/X11 dependencies from xcb-demo.
7165
7166commit 4142e34695310c20a8a30af64e0b40842bfb5dbe
7167Author: Ian Osgood <iano@quirkster.com>
7168Date:   Tue Mar 14 18:23:37 2006 -0800
7169
7170    Remove xcb-util dependency on proto/X11
7171    by moving many defs from X.h to <enum>s in xproto.xml
7172
7173commit f5c4956a81ed5aee3ea46e91785ec14f82e205d5
7174Author: Ian Osgood <iano@quirkster.com>
7175Date:   Tue Mar 14 11:53:01 2006 -0800
7176
7177    Fix XCBGetSetup to match prototype.
7178
7179commit 9472c251ae426bb496a3112d0ae390f1f1c35515
7180Merge: 71ddf29 0d648ac
7181Author: Ian Osgood <iano@quirkster.com>
7182Date:   Tue Mar 14 11:44:43 2006 -0800
7183
7184    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7185
7186commit 71ddf29d00ad8ff2323faf2cc6c354cbf3eca16c
7187Author: Ian Osgood <iano@quirkster.com>
7188Date:   Tue Mar 14 10:18:22 2006 -0800
7189
7190    Add an expression construct <bit bit="n"/>
7191    for mask enumerations such as CW flags.
7192    Replaces the C-specific use of 1<<n in xproto.xml
7193
7194commit 0d648ac0ab3a2d457284644e677fd2ed612f3e7c
7195Author: TORRI Vincent <torri@doursse.(none)>
7196Date:   Tue Mar 14 07:40:39 2006 +0100
7197
7198    doxygen documentation of the API
7199
7200commit a810d1ffe4825b9359f13ebb395f5f681961fc89
7201Author: Ian Osgood <iano@quirkster.com>
7202Date:   Mon Mar 13 17:49:17 2006 -0800
7203
7204    Remove dependencies on Xmd.h and X.h
7205    (Still including X.h until defs are moved to xproto.xml
7206     and xcb-util and xcb-demo are fixed to use them.)
7207
7208commit be1302b6efb33967bce5356af58e3e0ae3b19363
7209Author: Ian Osgood <iano@quirkster.com>
7210Date:   Mon Mar 13 10:36:13 2006 -0800
7211
7212    Remove last deprecation warning.
7213
7214commit d8de2c7c2a9b8bf59c8f1fcece5faed0b6cbf6f4
7215Merge: 5e0cfa8 e757673
7216Author: Ian Osgood <iano@quirkster.com>
7217Date:   Sun Mar 12 23:03:24 2006 -0800
7218
7219    Merge branch 'master' of git+ssh://iano@git.freedesktop.org/git/xcb
7220
7221commit e7576738c33e73fb4f29c1426c2ec49257564129
7222Author: Ian Osgood <iano@quirkster.com>
7223Date:   Sun Mar 12 23:02:45 2006 -0800
7224
7225    Restructure to remove most deprecation warnings.
7226
7227commit 5e0cfa84bec3e04f3a1991baaa29da09c7c4a02c
7228Author: Jamey Sharp <jamey@minilop.net>
7229Date:   Sun Mar 12 17:15:50 2006 -0800
7230
7231    Use libXdmcp, if available with XdmcpWrap, for XDM-AUTHORIZATION-1. Closes bug #6106.
7232
7233commit d69c403cba9bdebd1bd41b62ae7e28f5852248d4
7234Merge: b83f18a 5cdc02e
7235Author: Jamey Sharp <jamey@minilop.net>
7236Date:   Sun Mar 12 13:36:33 2006 -0800
7237
7238    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7239
7240commit b83f18a4cc2303dfda59807d56e16bbc5c18b09d
7241Author: Jamey Sharp <jamey@minilop.net>
7242Date:   Sun Mar 12 13:20:29 2006 -0800
7243
7244    Only _xcb_conn_wait calls _xcb_out_write now, so move it to xcb_conn.c and make it static.
7245
7246commit fb61c94d685a254ef0702a2e2093b8cdda02d514
7247Author: Jamey Sharp <jamey@minilop.net>
7248Date:   Sun Mar 12 12:40:34 2006 -0800
7249
7250    Remove c->out.vec. Pass iovecs directly down the call tree. Add _xcb_out_flush_to, refactor other functions, make write_block static.
7251
7252commit 5cdc02e3441dc623e91ebc5d7f735565b83f8435
7253Author: Jamey Sharp <jamey@minilop.net>
7254Date:   Sat Mar 11 20:32:04 2006 -0800
7255
7256    Portability fixes. Should help DragonFly and Solaris, and should not hurt anything else. Tested only on linux so far though.
7257
7258commit 83e652f566671f96ffc53a3c0099a84a1606c695
7259Author: Jamey Sharp <jamey@minilop.net>
7260Date:   Thu Mar 9 00:02:42 2006 -0800
7261
7262    Move c->out.vec refs out of _xcb_conn_wait up to _xcb_out_flush.
7263
7264commit 5b0ae3421dd373a8575b7a0d60989edfc056cf71
7265Merge: fd1f9cb 1261a41
7266Author: Jamey Sharp <jamey@minilop.net>
7267Date:   Wed Mar 8 22:50:48 2006 -0800
7268
7269    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7270
7271commit fd1f9cb13f9ab9a6bb6aa0c894d4891809bece8a
7272Author: Jamey Sharp <jamey@minilop.net>
7273Date:   Wed Mar 8 22:49:59 2006 -0800
7274
7275    Unify autofoo .gitignore bits into one top-level file, and remove resulting redundancies.
7276
7277commit 1261a4150c8bd39a6fb6887df52abbbb446f7ffb
7278Author: Ian Osgood <iosgood@Titania.local>
7279Date:   Wed Mar 8 22:26:20 2006 -0800
7280
7281    Add symbols for error and reply response_types,
7282    and use XCBKeymapNotify from xcb-proto.
7283
7284commit 6fb661f3ff2fca342b4ea76d5a583c5e4f53e076
7285Author: Ian Osgood <iosgood@Titania.local>
7286Date:   Wed Mar 8 21:56:57 2006 -0800
7287
7288    Bugfix: null-terminate the path in sockaddr_un before using it.
7289    This may fix itermittant connect failures.
7290
7291commit 621f891c49cbf4beba1e20fb9b6fb1be576d42f3
7292Author: Jamey Sharp <jamey@minilop.net>
7293Date:   Wed Mar 8 14:21:16 2006 -0800
7294
7295    Move c->out.vec refs out of _xcb_out_write up to _xcb_conn_wait.
7296
7297commit c491eeb9a9f670f7d4869d7dae7a5adce4565998
7298Author: Jamey Sharp <jamey@minilop.net>
7299Date:   Tue Mar 7 21:19:58 2006 -0800
7300
7301    Fix off-by-one error that kept the last byte(s) of the output queue from being used.
7302
7303commit 522a6e0eac9adeaac533a5b700f42d85d46e2dd7
7304Author: TORRI Vincent <Vincent.Torri@iecn.u-nancy.fr>
7305Date:   Tue Mar 7 10:25:23 2006 -0800
7306
7307    Use the GCC 4 visibility extension to mark everything in xcbint.h hidden.
7308
7309commit 5437032c7dee85da99b612a707fa94d012d40282
7310Merge: 2ad5450 d1cfd4d
7311Author: Jamey Sharp <jamey@minilop.net>
7312Date:   Tue Mar 7 00:26:38 2006 -0800
7313
7314    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7315
7316commit d1cfd4d0a21feaa5ccf0a1fd693327b8c5718abf
7317Author: Jamey Sharp <jamey@minilop.net>
7318Date:   Mon Mar 6 01:10:20 2006 -0800
7319
7320    Off-by-one error in the sequence-wrapping proof, and therefore in the corresponding code.
7321
7322commit 2ad5450ff3566a6e378769d7ff68f148c992b468
7323Author: TORRI Vincent <torri@doursse.(none)>
7324Date:   Sun Mar 5 09:49:02 2006 +0100
7325
7326    wrong spaces
7327
7328commit 81ea81c6d9a67fdf664fc502fe3610b37aea8354
7329Author: TORRI Vincent <torri@doursse.(none)>
7330Date:   Sun Mar 5 09:41:11 2006 +0100
7331
7332    fix some font sizes
7333
7334commit 66364da372dc1a2e1d75729a1265d632ae642ad6
7335Author: TORRI Vincent <torri@doursse.(none)>
7336Date:   Sun Mar 5 09:40:15 2006 +0100
7337
7338    fixes + valid css
7339
7340commit 01656220ef275a811a284420121d37bf9aa85f70
7341Author: TORRI Vincent <torri@doursse.(none)>
7342Date:   Sun Mar 5 09:22:03 2006 +0100
7343
7344    fixes + valid css
7345
7346commit e5458e477db95e1e064e46ca28245ecd51b5b524
7347Author: Jamey Sharp <jamey@minilop.net>
7348Date:   Sun Mar 5 00:20:50 2006 -0800
7349
7350    Implement provably-correct sequence wrap handling. Add flag XCB_REQUEST_DISCARD_REPLY.
7351
7352commit 6659c8c63b82329468b249cc99181d72ec26c698
7353Author: TORRI Vincent <torri@doursse.(none)>
7354Date:   Sun Mar 5 09:05:21 2006 +0100
7355
7356    lots of fixes. Thanks to Indan Zupancic
7357
7358commit 6e4745bbd0c924d846550496a072b0f2b0d4482c
7359Author: TORRI Vincent <torri@doursse.(none)>
7360Date:   Sun Mar 5 07:54:20 2006 +0100
7361
7362    lots of fixes. Thanks to Indan Zupancic
7363
7364commit df9c7cb4f9b05caf07900f3c15be379d3f9979fb
7365Author: Josh Triplett <josh@freedesktop.org>
7366Date:   Sat Mar 4 18:35:53 2006 -0800
7367
7368    Add *.tar.{bz2,gz} to .gitignore for the benefit of "make distcheck".
7369
7370commit 29f9fe0fc805a1ec6860f167a45664cc1cf0c769
7371Author: Jamey Sharp <jamey@minilop.net>
7372Date:   Fri Mar 3 11:08:10 2006 -0800
7373
7374    API/ABI change: XCBSendRequest returns the sequence number instead of using an out-parameter. Now 0 is a special sequence number indicating failure.
7375
7376commit 255c21b17f61147388bab6e1d42623a008a4a8d2
7377Author: Jamey Sharp <jamey@minilop.net>
7378Date:   Fri Mar 3 01:45:00 2006 -0800
7379
7380    Add 32-bit full_sequence fields to generic errors and events, so callers can always correlate events and errors with requests correctly.
7381
7382commit 622b599c8fcf5d677f76ca03f3241a23dba58712
7383Author: Jamey Sharp <jamey@minilop.net>
7384Date:   Thu Mar 2 23:39:38 2006 -0800
7385
7386    Tweak to previous API change: Require that spare iovecs fall before vector[0]. Leave vector in well-defined state.
7387
7388commit 87905f0579e749ac6d92843780af246160318eca
7389Author: Jamey Sharp <jamey@minilop.net>
7390Date:   Thu Mar 2 23:31:35 2006 -0800
7391
7392    assert() that XCBSendRequest was handed enough space to set opcodes and short length fields.
7393
7394commit 45e1cc0935ca1d7a138e1c700db229f50205a556
7395Author: Jamey Sharp <jamey@minilop.net>
7396Date:   Thu Mar 2 23:12:21 2006 -0800
7397
7398    Bugfix: The fixed-length part needs two iovecs reserved, just like all the other parts.
7399
7400commit b46953f46c7d0fd8cbb73f922524968c1ac86fa4
7401Author: Jamey Sharp <jamey@minilop.net>
7402Date:   Thu Mar 2 23:05:08 2006 -0800
7403
7404    Use XPath position() function instead of xsl:number, to evaluate array indices while generating code for easier readability.
7405
7406commit 5e115e2441ed32f5fa495370b36b01c03bbff66d
7407Author: Jamey Sharp <jamey@minilop.net>
7408Date:   Thu Mar 2 15:35:31 2006 -0800
7409
7410    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.
7411
7412commit ed823bf65192a72f8c3060698c9bded9f77d49c2
7413Author: Jamey Sharp <jamey@minilop.net>
7414Date:   Thu Mar 2 13:49:00 2006 -0800
7415
7416    Separate notion of request-completed from current-request, and mark requests completed more aggressively. Detects some usage errors that are otherwise undetectable.
7417
7418commit c05ae15b661bff6b95deb3abc7f48abe16892ac8
7419Author: Jamey Sharp <jamey@minilop.net>
7420Date:   Mon Feb 27 12:12:33 2006 -0800
7421
7422    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.
7423
7424commit 8f991bdd389f6c419cb18cdaea966304529de165
7425Author: Jamey Sharp <jamey@minilop.net>
7426Date:   Mon Feb 27 11:03:13 2006 -0800
7427
7428    Add XCB_REQUEST_RAW flag for XCBSendRequest.
7429
7430commit 5e350126a728f3c0c3bc8d8673e5ad67dc174f79
7431Author: Jamey Sharp <jamey@minilop.net>
7432Date:   Mon Feb 27 10:53:08 2006 -0800
7433
7434    Bugfix: how about *not* leaking all pending_replies when no reply matches, as often happens with Xlib?
7435
7436commit 731c85762d8994119f4eaf572cc59c9adbb7abd9
7437Author: Jamey Sharp <jamey@minilop.net>
7438Date:   Mon Feb 27 02:57:36 2006 -0800
7439
7440    Move test for other writing threads *before* allocating a sequence number.
7441
7442commit 8ab4dcde9db8175ef0093123bd04c72471e205ff
7443Author: Jamey Sharp <jamey@minilop.net>
7444Date:   Mon Feb 27 02:14:48 2006 -0800
7445
7446    _xcb_list is no longer used, so remove it. Simplify _xcb_map.
7447
7448commit 76ad79a7adbb315f7c7d5f08d6f42553210868cd
7449Author: Jamey Sharp <jamey@minilop.net>
7450Date:   Mon Feb 27 01:43:07 2006 -0800
7451
7452    Replace readers generic list with hand-written typesafe version.
7453
7454commit fa71b9b835da6360c374dcb4ab93be3a27dbd737
7455Author: Jamey Sharp <jamey@minilop.net>
7456Date:   Mon Feb 27 01:06:15 2006 -0800
7457
7458    Rename struct XCBReplyData to reader_list to follow my naming conventions and better reflect its purpose.
7459
7460commit ff665b57266b9e7e9b9a366272c2115bbd516173
7461Author: Jamey Sharp <jamey@minilop.net>
7462Date:   Mon Feb 27 00:54:59 2006 -0800
7463
7464    In struct XCBReplyData, change void *data to pthread_cond_t *data. That was dumb.
7465
7466commit 50acfeae36f5f67e8b92fd7610141c489ee348c5
7467Author: Jamey Sharp <jamey@minilop.net>
7468Date:   Sun Feb 26 23:43:44 2006 -0800
7469
7470    Delete unused xcb_list functions and refactor others.
7471
7472commit 86ce18c22c8daebe8796d3c38e8a1d496ab6e8de
7473Author: Jamey Sharp <jamey@minilop.net>
7474Date:   Sun Feb 26 22:58:17 2006 -0800
7475
7476    Replace current_reply generic queue with hand-written version. No generic queues remain so delete their implementation.
7477
7478commit 0f130b4d945a27fd2b4655c351ebe70d61cac598
7479Author: Jamey Sharp <jamey@minilop.net>
7480Date:   Sun Feb 26 18:28:50 2006 -0800
7481
7482    Replace events generic queue with hand-written typesafe version.
7483
7484commit ff7b6c9124e7caf26381cc7a10fba9eaf1875652
7485Author: Jamey Sharp <jamey@minilop.net>
7486Date:   Sun Feb 26 15:45:08 2006 -0800
7487
7488    API/ABI break: Add flags to XCBSendRequest, first for error-checked requests.
7489    There's no more race condition between event and reply handling.
7490    The *RequestBlind and *RequestChecked functions are not yet implemented.
7491
7492commit 7875040fa179d8ca3fa4ec046c1b8c2a351a6621
7493Author: Jamey Sharp <jamey@minilop.net>
7494Date:   Sun Feb 26 02:00:03 2006 -0800
7495
7496    Replace pending_replies generic queue with a hand-implemented typesafe version.
7497
7498commit a1eff0c49a2dbb328a79edbeb2b851fa1dfe9e61
7499Author: Jamey Sharp <jamey@minilop.net>
7500Date:   Sun Feb 26 01:27:01 2006 -0800
7501
7502    Replace my old generic map ADT with a growable array for the extension cache.
7503
7504commit 46a754998149c5f4a1670787b3ea36731caf6506
7505Author: Jamey Sharp <jamey@minilop.net>
7506Date:   Sat Feb 25 23:27:47 2006 -0800
7507
7508    XCB has not had tracing features for a long time: remove the remnants.
7509
7510commit b6cbe837748df5819d00682330d34e2c7d24d349
7511Author: Jamey Sharp <jamey@minilop.net>
7512Date:   Sat Feb 25 23:26:55 2006 -0800
7513
7514    Rearrange an if statement that's been bothering me.
7515
7516commit f27166f49b9ef6bdcce78429bffc724d1e4fb360
7517Author: Jamey Sharp <jamey@minilop.net>
7518Date:   Fri Feb 24 02:17:36 2006 -0800
7519
7520    Coalesce _xcb_writev into _xcb_out_write and simplify.
7521
7522commit 9463653b1e6dc0a9054266aa3eecb0839129b991
7523Author: Jamey Sharp <jamey@minilop.net>
7524Date:   Fri Feb 24 01:56:18 2006 -0800
7525
7526    Take advantage of Requires and *.private fields for a more accurate pkg-config file.
7527
7528commit bae98d36040d0cda1862839410cf92bac3927280
7529Author: Jamey Sharp <jamey@minilop.net>
7530Date:   Fri Feb 24 01:50:48 2006 -0800
7531
7532    Move _xcb_set_fd_flags to xcb_conn.c and make it static. xcb_util.c now has only public functions.
7533
7534commit 67b2649dc4b6726c6d11fb0e41429ae5de82b0e8
7535Author: Jamey Sharp <jamey@minilop.net>
7536Date:   Fri Feb 24 01:40:45 2006 -0800
7537
7538    Move _xcb_read_block to xcb_in.c and make it static. Change calls in xcb_conn.c to _xcb_in_read_block instead.
7539
7540commit 838317f4d305e07f35ddd99e2ebccfa8fac680c6
7541Author: Jamey Sharp <jamey@minilop.net>
7542Date:   Fri Feb 24 01:17:03 2006 -0800
7543
7544    Remove XCB_CEIL and use a simpler definition for XCB_PAD.
7545
7546commit 1b50d2ee1ef532429674126eace88ac73d51ec23
7547Author: Jamey Sharp <jamey@minilop.net>
7548Date:   Fri Feb 24 01:04:48 2006 -0800
7549
7550    Quit using "-include config.h": use #ifdef HAVE_CONFIG_H etc. like everyone else.
7551
7552commit e1c2777abac0479ebdf06f2005120d25fff53935
7553Author: Jamey Sharp <jamey@minilop.net>
7554Date:   Fri Feb 24 00:59:08 2006 -0800
7555
7556    Coalesce readn into _xcb_in_read and simplify.
7557
7558commit 3f8d0bd5322749132626e8f203017b6da6448fd0
7559Author: Jamey Sharp <jamey@minilop.net>
7560Date:   Fri Feb 24 00:48:18 2006 -0800
7561
7562    Move _xcb_readn to xcb_in.c and make it static. Minor change to _xcb_read_block to not depend on _xcb_readn.
7563
7564commit cdf362f33a7bc4d412231de32e37ffb65a06d3f1
7565Author: Jamey Sharp <jamey@minilop.net>
7566Date:   Fri Feb 24 00:25:34 2006 -0800
7567
7568    Bugfix: protect the output queue from being written while another thread is flushing it.
7569
7570commit 4e91ae275e46e76de64ab089d6f13c1e9b6ddee6
7571Author: Jamey Sharp <jamey@minilop.net>
7572Date:   Fri Feb 24 00:02:43 2006 -0800
7573
7574    Simplify: Always use writev. (In _xcb_out_flush, convert the output queue to a single iovec if needed.)
7575
7576commit 7f0bc778c88ab2f565cc05d5d3d5ee4c8d1388a1
7577Author: Jamey Sharp <jamey@minilop.net>
7578Date:   Thu Feb 23 22:41:59 2006 -0800
7579
7580    Factor padding out of _xcb_out_write_block and into its callers, XCBSendRequest and write_setup.
7581    This requires dynamically allocating memory in XCBSendRequest, but this
7582    malloc/free pair turns out to cause a 30% speed hit for the 'x11perf -noop'
7583    test -- so for the moment I use alloca where available and fall back to malloc
7584    on other platforms. Later I think I'll change the contract of XCBSendRequest
7585    so the caller is responsible for memory allocation, because the caller ought
7586    to always be able to stack-allocate here.
7587
7588commit 6e29e5f2ee2e6158f1a9480a83e4f906ab9c04d1
7589Author: Jamey Sharp <jamey@minilop.net>
7590Date:   Thu Feb 23 17:50:53 2006 -0800
7591
7592    Add XCBGetQueuedRequestRead for Xlib that does no syscalls, just returns whatever XCB already knows about.
7593
7594commit 26ac6292ba0535ac3747d23f2f4d284c4b5f42ba
7595Author: Jamey Sharp <jamey@minilop.net>
7596Date:   Thu Feb 23 15:29:40 2006 -0800
7597
7598    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.
7599
7600commit a736674943295245ccb90865c13d2b75dc6ecccb
7601Author: Jamey Sharp <jamey@minilop.net>
7602Date:   Thu Feb 23 14:57:46 2006 -0800
7603
7604    Minor performance fix: Only rearrange buffers for BIG-REQUESTs.
7605
7606commit e866bed9348bf3ffd480fd6ec329fd8b9581c328
7607Author: Jamey Sharp <jamey@minilop.net>
7608Date:   Thu Feb 23 14:32:11 2006 -0800
7609
7610    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.
7611
7612commit 1b83f8f8f326eca9d8852c82dd36696f81a720dc
7613Author: Jamey Sharp <jamey@minilop.net>
7614Date:   Thu Feb 23 14:30:08 2006 -0800
7615
7616    Remove _xcb_assert_valid_sequence. One test is trivially true, and the other may be temporarily violated without anything bad happening.
7617
7618commit 55c1842686d2e668708cd106b5e08847df0184c3
7619Author: Jamey Sharp <jamey@minilop.net>
7620Date:   Thu Feb 23 12:48:27 2006 -0800
7621
7622    Move _xcb_write and _xcb_writev to xcb_out.c and make them static, since only _xcb_out_write calls them.
7623
7624commit 213b5725928ccf8aedc807cc40a261b2d5431247
7625Author: Jamey Sharp <jamey@minilop.net>
7626Date:   Thu Feb 23 12:34:08 2006 -0800
7627
7628    Ignore test byproducts.
7629
7630commit 6149c7a6b57a193bc132fbc35b1772b75e5c7ca7
7631Author: Jamey Sharp <jamey@minilop.net>
7632Date:   Thu Feb 23 12:15:09 2006 -0800
7633
7634    More return value changes, and make _xcb_in_read_packet static since it is not called from outside xcb_in.c.
7635
7636commit 5b1d39e27b1a966df537ead248da5a57e7d9de97
7637Author: Jamey Sharp <jamey@minilop.net>
7638Date:   Thu Feb 23 11:50:12 2006 -0800
7639
7640    More return value changes.
7641
7642commit 662479760c42fc38c458381ee3eaed92e2c8b733
7643Author: Jamey Sharp <jamey@minilop.net>
7644Date:   Thu Feb 23 10:17:40 2006 -0800
7645
7646    Make the return value of _xcb_conn_wait boolean, instead of syscall-like.
7647
7648commit f8a8b465e080faf5479c0b38dc67cc9b9ae24927
7649Author: Jamey Sharp <jamey@minilop.net>
7650Date:   Wed Feb 22 23:38:16 2006 -0800
7651
7652    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.
7653
7654commit e0a35783837e0d4974cfa20c6a7a073698f3b0a5
7655Author: Jamey Sharp <jamey@minilop.net>
7656Date:   Wed Feb 22 23:11:36 2006 -0800
7657
7658    Remove USE_THREAD_ASSERT compile option. _xcb_conn_wait can no longer be re-entered, since we have no callback hooks now.
7659
7660commit a7d749ec3fd3303a4e7ace9d4f0f1672f9310ef2
7661Author: Jamey Sharp <jamey@minilop.net>
7662Date:   Wed Feb 22 23:09:29 2006 -0800
7663
7664    Remove vestiges of adjacent request combining implementation. It is very dead.
7665
7666commit 06c788932bb635da79076728927b7b93f64037c1
7667Author: Jamey Sharp <jamey@minilop.net>
7668Date:   Mon Feb 20 00:41:39 2006 -0800
7669
7670    Control debugging, optimization, and warning flags all independently.
7671
7672commit 3f98dae4a8274f2430c4967f8354356ef7c45024
7673Merge: e346f8c f825a32
7674Author: Jamey Sharp <jamey@minilop.net>
7675Date:   Sat Feb 18 23:57:18 2006 -0800
7676
7677    Merge branch 'master' of git+ssh://git.freedesktop.org/git/xcb
7678
7679commit e346f8c8efe6f931aa43026087d0ea406258c094
7680Author: Jamey Sharp <jamey@minilop.net>
7681Date:   Sat Feb 18 23:57:14 2006 -0800
7682
7683    Update .gitignores for .o files and autofoo stuff.
7684
7685commit f825a32135fa613dd6b6dd8553e366d8125bd7a4
7686Author: Eric Anholt <anholt@FreeBSD.org>
7687Date:   Sat Feb 18 23:05:32 2006 -0800
7688
7689    Add missing entries to .gitignore files.
7690
7691commit 9e013b564cc4479802c51f79ea5054d5f25edacc
7692Merge: d16efb5 7474473
7693Author: Eric Anholt <anholt@FreeBSD.org>
7694Date:   Sat Feb 18 22:57:42 2006 -0800
7695
7696    Merge branch 'master' of git+ssh://anholt@git.freedesktop.org/git/xcb
7697
7698commit d16efb51915d9661788c96140635514cf6314d46
7699Author: Eric Anholt <anholt@FreeBSD.org>
7700Date:   Sat Feb 18 22:56:03 2006 -0800
7701
7702    Move .cvsignore to .gitignore.
7703
7704commit 7474473f3312597f6de736b8f25cedc433427092
7705Author: Jamey Sharp <jamey@minilop.net>
7706Date:   Sat Feb 18 20:03:35 2006 -0800
7707
7708    Quit defining _XOPEN_SOURCE. We never needed it...
7709
7710commit 1d977c7b2d519e22f319075563e7faf4bbd31d58
7711Author: Jamey Sharp <jamey@id.minilop.net>
7712Date:   Sat Feb 18 19:23:33 2006 -0800
7713
7714    Remove the old ChangeLog file. GIT makes that obsolete.
7715
7716commit 5b7182c659391160239467f1041a1d755db45bd3
7717Author: Jamey Sharp <jamey@minilop.net>
7718Date:   Sat Feb 18 18:12:40 2006 -0800
7719
7720    Workaround X server bug, fd.o #3210: if a GLXGetFBConfigs request is sent, fix the length field in the reply.
7721
7722commit 47ceed78612d48dcda62cc2686fc527d61abe38b
7723Author: Josh Triplett <josh@freedesktop.org>
7724Date:   Sat Feb 18 16:49:41 2006 -0800
7725
7726    Remove xcl and CVSROOT.
7727