| /src/usr.bin/make/unit-tests/ |
| varname-dot-objdir.exp | 1 : purge-cache was reached.
|
| opt-keep-going.mk | 8 # status was nevertheless 0, which was wrong since the main targets could
9 # not be made. This was only wrong in -k mode combined with compat mode.
|
| deptgt-end-fail-all.mk | 7 # Until 2020-12-07, the .END node was made even if the main nodes had failed.
8 # This was not intended since the .END node had already been skipped if a
10 # themselves had failed. This inconsistency was not worth keeping. To run
|
| counter-append.mk | 7 # 6 words, even though the NEXT variable was only accessed 3 times.
8 # The cause for this surprising behavior was that the ::= variable modifiers
|
| counter.mk | 7 # 4 words, even though the NEXT variable was only accessed 3 times.
8 # The cause for this surprising behavior was that the ::= variable modifiers
|
| var-scope-cmdline.mk | 39 # was not entirely equivalent to the simple rule though. The mistake was
40 # that when a cmdline variable with '$$' in its name was added, a global
41 # variable was deleted, but not with the exact same name as the cmdline
42 # variable. Instead, the name of the global variable was expanded one more
44 # have a '$$' in their name, it was implemented correctly all the time.
46 # The bug was added in var.c 1.183 on 2013-07-16, when Var_Set called
50 # row. This bug was fixed in var.c 1.835 from 2021-02-22.
52 # Another bug was the wrong assumption that "deleting a cmdline variable is
55 # temporary loop variable after finishing the loop. It was probably no
[all...] |
| deptgt-begin-fail-indirect.mk | 8 # status was even 0.
|
| deptgt-end-fail-indirect.mk | 6 # and exited with status 0. The cause for this was a missing condition in
|
| opt-debug-graph3.mk | 6 # Before compat.c 1.222 from 2021-02-02, there was no debug output despite
|
| unexport.mk | 12 # was not reachable though. At that point, backslash-newline has already
|
| varname-dot-objdir.mk | 15 : ${.TARGET} was reached.
|
| directive-for-if.mk | 9 # Before that change, the following adventurous code was possible:
16 # A more practical usage of the .for loop that often led to surprises was the
30 # make complained about a "Malformed conditional", which was surprising since
35 # ${:Uvalue}, this surprising behavior was documented in the code, and a
36 # workaround was implemented that allowed bare words when they are followed
69 _!= echo "${var}" 1>&2; echo # In 2005, '.info' was not invented yet.
73 # Before for.c 1.39 from 2008-12-21, a common workaround was to surround the
77 # .for loop contained '"', which was effectively a code injection.
82 # loop was usually close to the .if clause, this was not a problem i
[all...] |
| deptgt-end-fail.mk | 7 # even though this was intended. The cause for this was a missing condition
66 # Until 2020-12-07, '.END' was even made if 'all' failed, but if a dependency
67 # of 'all' failed, it was skipped. This inconsistency was not needed for
|
| depsrc-use.mk | 10 # Before make.h 1.280 from 2021-12-28, a .USEBEFORE target was accidentally
12 # target was not.
|
| deptgt-silent-jobs.mk | 10 # was set. This was fixed in job.c 1.452 from 2022-02-12.
|
| include-main.mk | 7 # was included from", while before 2020-09-05 it was the "filename in which
|
| varmod-to-abs.mk | 8 # Between 2016-06-03 and 2020-11-14, it was possible to trick the :tA modifier
14 # This unintended behavior was caused by cached_realpath using a GNode for
|
| varmod-gmtime.mk | 47 # Before var.c 1.1050 from 2023-05-09, it was not possible to pass the
54 # Before var.c 1.631 from 2020-10-31 21:40:20, it was possible to pass
110 # Because this modifier is implemented using strtoul, the parsed time was
123 # Before var.c 1.631 from 2020-10-31, there was no error handling while
124 # parsing the :gmtime modifier, thus no error message was printed. Parsing
125 # stopped after the '=', and the remaining string was parsed for more variable
127 # variable value was discarded and thus not printed.
136 # followed by the next modifier, without a ':' separator. This was the same
|
| /src/sys/external/gpl2/dts/dist/arch/arm/boot/dts/ |
| omap3-beagle-xm-ab.dts | 9 /* HS USB Port 2 Power enable was inverted with the xM C */
|
| /src/common/lib/libc/quad/ |
| muldi3.c | 7 * This software was developed by the Computer Systems Engineering group
196 u_int prodh, prodl, was; local in function:__lmulq
229 was = prodl;
231 prodh -= HHALF(mid) + (prodl > was);
233 was = prodl;
235 prodh += HHALF(mid) + (prodl < was);
239 was = prodl;
241 prodh += HHALF(low) + (prodl < was);
|
| /src/sys/arch/m68k/060sp/dist/ |
| changes | 41 1) "movep" emulation where data was being read from memory
42 was reading the intermediate bytes. Emulation now only
45 2) "flogn", "flog2", and "flog10" of "1" was setting the
50 mode was pre-decrement or post-increment and the address register
51 was A0 or A1, the address register was not being updated as a result
|
| /src/lib/libc/arch/powerpc64/gen/ |
| swapcontext.S | 44 std %r3,SF_PARAM(%r1) # must save oucp (was 8)
45 std %r4,SF_PARAM+8(%r1) # must save ucp (was 12)
|
| /src/lib/libc/arch/vax/gen/ |
| __sigsetjmp14.S | 60 tstl (_JBLEN*4)(%r0) # test if "savemask" was set
|
| /src/sys/external/gpl2/dts/dist/arch/arm64/boot/dts/allwinner/ |
| sun50i-a64-pinephone-1.2.dts | 21 * PWM backlight circuit on this PinePhone revision was changed since 1.0,
24 * quite a bit between individual boards, so the lowest value here was
|
| /src/usr.bin/indent/ |
| README.md | 5 by James Gosling @ CMU. At some point in the 1970s or even 1980s, it was
13 In 2021, indent was updated to handle C99 comments and the code got a major
|
