| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Updated separately to allow the previous patch to be used for multiple
releases.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In Glibc 2.34 and later, SIGSTKSZ may not be a compile-time constant.
It is no longer possible to statically allocate the alternate signal
stack for the main thread, as we've been doing for the last 25 years.
This commit implements dynamic allocation of the alternate signal stack
even for the main thread. It reuses the code already in place to allocate
the alternate signal stack for other threads.
The alternate signal stack is freed when the main OCaml code / an OCaml thread
stops.
(partial back-port of PR#10266 and PR#10726)
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some `Long_val` conversions were missing.
This was setting the wrong values for the `custom_` parameters,
causing the major GC to work too much, slowing down some programs.
Add regression test.
Fixes: #9326
(cherry picked from commit 78321e6ac59e18a54cd9209aa57012f8b47908c9)
|
|
|
|
|
|
|
|
|
| |
The variables are caml_debug_info and caml_atom_table.
The multiple definitions look like a cut-and-paste error.
They cause problems with C compilers that don't follow the "common" model.
Fixes: #9144
(cherry picked from commit 53327d777ddcacf829c77c0bd3569a8ade2c72b8)
|
|
|
|
|
|
|
| |
The latest mingw-w64 _mingw.h header now automatically sets __USE_MINGW_ANSI_STDIO=1 for C99 and later. mingw-w64 always defines snprintf but the inline definition when __USE_MINGW_ANSI_STDIO=0 temporarily #undef's any snprintf macro so works in harmony with ours. The other case does not do this so we get declaration errors.
Fixes: #9938 .
(cherry picked from commit a03b6035ebe4682950018e0e1337bab18771fd81)
|
|
|
|
|
|
| |
Don't assume . in AWKPATH
(cherry picked from commit d4ace8c347a9a4f8baa8f0b1738a5b4436fcd027)
|
| |
|
|
|
|
|
| |
(cherry picked from commits 7156b9cd1511cc2eb79a60856208269e5be1edc4
and 3bb1886b6564a4774ae41f4743628435633852ac)
|
|
|
|
|
|
| |
Check for definition of AT_SECURE before using it
This fixes compilation on kfreebsd.
|
|
|
|
|
|
| |
Windows supports make reconfigure now
(cherry picked from commit c71997a167c3670d202c6ecaf830c6a25b4b95b8)
|
|
|
|
|
|
|
| |
In --disable-installing-bytecode-programs mode, the .opt version of the
tools is installed, but the symlink for the tool itself is not created.
(cherry picked from commit 705739fa54260b7a0e6cbba0b5a99e52c79f9c09)
|
|
|
|
| |
Improve detection of stat supported options
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Fixes: #8841.
Cherry-pick of commit 5e4b55d3b on trunk.
|
| |
|
|\
| |
| | |
Small fixes in configure.ac
|
| | |
|
| | |
|
|/
|
|
|
|
|
|
|
| |
Variables host (i.e. host type canonicalized by config.sub) and
host_alias (i.e. the argument to --host) are subtly different, the
right one to use as prefix is $host_alias.
By the way, ac_tool_prefix is defined the same way, maybe it should be
used instead of toolpref?
|
|\
| |
| | |
Fix spelling errors reported by Lintian
|
| |
| |
| |
| | |
No change entry needed
|
|/
|
|
| |
No change entry needed
|
| |
|
| |
|
| |
|
|\
| |
| | |
Backport #8818 (Dynlink/packing issue in bytecode) for 4.08+rc2
|
|/
|
|
|
|
| |
Fix #8816 (Dynlink/packing issue in bytecode with 4.08)
(cherry picked from commit e14a61158caa2249eef02fafa244bf1572fbec67)
|
|
|
|
|
|
|
|
| |
* Fix typographic issues in the doc of new modules
No change entry needed
(cherry picked from commit 6a6f34e48306d573bf3036681cb304ed84415957)
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
A few tests have actually been broken by commit
b2a0f3b659a8d7e2f44c337718c1ea9373479d7d
The present commit fixes all the tests that had not been fixed meanwhile.
(cherry picked from commit 1a798abff0cbf135da5526375e62f88ccfdd07de)
|
|
|
|
|
|
|
|
| |
This is a follow-up to commit 7077b60 that fixed a lack of 8-alignment for the frame table on ADM64, as reported in #7591.
A similar issue was reported in #7887 for ARM64 and is fixed here.
For good measure, explicit alignment was added to PPC64 as well, although there was probably no issue there.
Closes: #7887.
|
| |
|
|
|
|
|
| |
Pervasives.input has never been a primitive (it was always unsafe_input)
- string_of_int has the benefit of being declared external.
|
| |
|
|
|
|
|
| |
manual: restore two missing chapters
(cherry picked from commit 047f231f4c0ea97e3b877ef3d302cc4647385bd9)
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* Don't generate illegal Pfield's when compiling alias coercions
* Simplify lambda code when compiling packs
* Add regression test for pr8769
* Add Changes entry
|
| |
|
| |
|
|
|
|
|
|
| |
amd64: align data section to word boundary
(cherry-picked from 40cda4a1d61d4d07645b4a33d13f4f080e5c390a)
|
|
|
|
| |
(cherry picked from commit 466950b48f698cb2d9fbb5571a3ed203a8765044)
|
|
|
|
|
|
|
|
|
|
|
|
| |
Makefile.dev is only included if a Git clone is detected.
make evaluates macros in target specifications when the Makefile is
read, which meant that the build-all-asts target caused `git ls-files`
to be called on every invocation of make. This tweaks it to be a
recursive call to make which puts the $(AST_FILES) macro in the recipe
where it is only evaluated when the target is requested.
(cherry picked from commit 93ee6b43b05783d56e3d8d984fb2503f5870807d)
|
|
|
|
|
|
| |
Fix Option.bind doc comment
(cherry picked from commit 7e6618a4213dba1e41d71a5355d29794b0f5e402)
|