threads.xs: Fix warning "implicit declaration of function '_setjmpex'" using MinGW-w64

[twata1]

When building threads 2.28 using at least gcc-8.5.0, gcc-11.3.0 and gcc-12.1.0 of MinGW-w64, the following warning will appear.

Generating a gmake-style Makefile
Writing Makefile for threads
gmake[1]: Entering directory 'C:/Perl64-gcc10/perl-5.37.2/dist/threads'
Running Mkbootstrap for threads ()
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" -MExtUtils::Command -e chmod -- 644 "threads.bs"
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" -MExtUtils::Command::MM -e cp_nonempty -- threads.bs ..\..\lib\auto\threads\threads.bs 644
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" "C:\Perl64-gcc10\perl-5.37.2\dist\ExtUtils-ParseXS\lib\ExtUtils/xsubpp"  -typemap C:\Perl64-gcc10\perl-5.37.2\lib\ExtUtils\typemap  threads.xs > threads.xsc
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" -MExtUtils::Command -e mv -- threads.xsc threads.c
gcc -c   -DWIN32 -DWIN64 -DPERL_TEXTMODE_SCRIPTS -DMULTIPLICITY -DPERL_IMPLICIT_SYS -DUSE_PERLIO -D__USE_MINGW_ANSI_STDIO -fwrapv -fno-strict-aliasing -mms-bitfields -O2   -DVERSION=\"2.28\" -DXS_VERSION=\"2.28\"  "-I..\..\lib\CORE"   threads.c
threads.xs: In function 'S_jmpenv_run':
threads.xs:18:23: warning: implicit declaration of function '_setjmpex'; did you mean '_setjmp3'? [-Wimplicit-function-declaration]
   18 | #    define setjmp(x) _setjmpex((x), mingw_getsp())
      |                       ^~~~~~~~~
threads.xs:18:23: note: in definition of macro 'setjmp'
   18 | #    define setjmp(x) _setjmpex((x), mingw_getsp())
      |                       ^~~~~~~~~
..\..\lib\CORE/iperlsys.h:1083:33: note: in expansion of macro 'Sigsetjmp'
 1083 | #  define PerlProc_setjmp(b, n) Sigsetjmp((b), (n))
      |                                 ^~~~~~~~~
..\..\lib\CORE/cop.h:122:26: note: in expansion of macro 'PerlProc_setjmp'
  122 |         cur_env.je_ret = PerlProc_setjmp(cur_env.je_buf, SCOPE_SAVES_SIGNAL_MASK);  \
      |                          ^~~~~~~~~~~~~~~
threads.xs:504:5: note: in expansion of macro 'JMPENV_PUSH'
  504 |     JMPENV_PUSH(jmp_rc);
      |     ^~~~~~~~~~~
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" -MExtUtils::Mksymlists \
     -e "Mksymlists('NAME'=>\"threads\", 'DLBASE' => 'threads', 'DL_FUNCS' => {  }, 'FUNCLIST' => [], 'IMPORTS' => {  }, 'DL_VARS' => []);"
g++ threads.def -o ..\..\lib\auto\threads\threads.dll -shared -s -L"c:\perl\lib\CORE" -L"C:\winlibs-x86_64-posix-seh-gcc-10.3.0-llvm-12.0.0-mingw-w64-9.0.0-r2\mingw64\lib" -L"C:\winlibs-x86_64-posix-seh-gcc-10.3.0-llvm-12.0.0-mingw-w64-9.0.0-r2\mingw64\x86_64-w64-mingw32\lib" -L"C:\winlibs-x86_64-posix-seh-gcc-10.3.0-llvm-12.0.0-mingw-w64-9.0.0-r2\mingw64\lib\gcc\x86_64-w64-mingw32\10.3.0" threads.o   "..\..\lib\CORE\libperl537.a" -lmoldname -lkernel32 -luser32 -lgdi32 -lwinspool -lcomdlg32 -ladvapi32 -lshell32 -lole32 -loleaut32 -lnetapi32 -luuid -lws2_32 -lmpr -lwinmm -lversion -lodbc32 -lodbccp32 -lcomctl32 -Wl,--enable-auto-image-base
"C:\Perl64-gcc10\perl-5.37.2\miniperl.exe" "-I..\..\lib" -MExtUtils::Command -e chmod -- 755 ..\..\lib\auto\threads\threads.dll

This patch was proposed by @sisyphus.
We have also confirmed that the warning has been suppressed in the following environments.

[Environment]
Windows 8.1 64 bit
Perl 5.37.2
gcc-8.5.0
gcc-11.3.0
gcc-12.1.0

Fixes #20096

[thesamesam]

I don't think it's really a suppression, it's actually fixing the warning.

[twata1]

I changed the word "Suppress" to "Fix" in the commit message and the pull requests title.

[khwilliamson]

Thank you for this patch. I'm ready to merge it, but we need an entry for you as a Perl author. Traditionally this has been a real name and an email address. But we can use just @twata1 if that is what you'd like.

[demerphq]

On Sun, 4 Dec 2022 at 14:01, Karl Williamson ***@***.***> wrote: Thank you for this patch. I'm ready to merge it, but we need an entry for you as a Perl author. Traditionally this has been a real name and an email address. But we can use just @twata1 <https://github.com/twata1> if that is what you'd like.

Please just run Porting/updateAUTHORS.pl It will create the required changes based on your git config. You should review them and then commit them as part of your PR. You can also ask the script for --help and --man. Note it may take a few seconds if git hasn't operated on the perl repo recently, so either just wait, or you can add --from=v5.37.6 to get it to do a little less work and finish quicker. cheers, Yves

…

-- perl -Mre=debug -e "/just|another|perl|hacker/"

[twata1]

Thank you for your response.
I am trying to do Porting/updateAUTHORS.pl now.

[demerphq]

Dont forget to commit and push the results. :-)

[demerphq]

Something weird about how you did that. I think you were not rebased onto the latest version of the code. I will find some time later to resolve the conflicts here and figure out what is up. If I run updateAUTHORS.pl against blead right now it does not find anything to update, so i do not know why when you ran it picked up the other people.

[twata1]

I apologize for not rebasing to the latest version of the code.

[demerphq]

Not your fault. I should have added an "--add-me" to the tool which would avoid this.

[demerphq]

Ok, fixed. You were sufficiently behind blead that you were using an old version of the tool. The new version would have done it different. I have rebased ontop of blead, and will merge once CI finishes. Thanks for the fix, and sorry you had a bad experience for your first PR.

[twata1]

No problem.
Thank you very much for the information about Porting/updateAUTHORS.pl!

[demerphq]

On Tue, 6 Dec 2022 at 03:32, twata1 ***@***.***> wrote: No problem. Thank you very much for the information about Porting/updateAUTHORS.pl! You are welcome. I am trying to make the contributor recognition process

easier, thanks to your experience I will push a patch so the next person has it even easier. No idea why I didnt the --add-me option, there is something very similar for update-me. Bah, blind spots. :-) Yves

…

-- perl -Mre=debug -e "/just|another|perl|hacker/"

[demerphq]

On Tue, 6 Dec 2022 at 09:48, demerphq ***@***.***> wrote: On Tue, 6 Dec 2022 at 03:32, twata1 ***@***.***> wrote: > > No problem. > Thank you very much for the information about Porting/updateAUTHORS.pl! > > You are welcome. I am trying to make the contributor recognition process easier, thanks to your experience I will push a patch so the next person has it even easier. No idea why I didnt the --add-me option, there is something very similar for update-me. Bah, blind spots. :-)

BTW, I noticed your patch needs a bump to the version of threads. I was assuming you would take care of that, but i can if you wish. Yves

…

-- perl -Mre=debug -e "/just|another|perl|hacker/"

[twata1]

I wish someone else can take care of this as I am having trouble reproducing it in my Windows environment.

[demerphq]

Ok. Ill take care of it.

…

On Fri, 9 Dec 2022 at 17:42, twata1 ***@***.***> wrote: I wish someone else can take care of this as I am having trouble reproducing it in my Windows environment. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

-- perl -Mre=debug -e "/just|another|perl|hacker/"

[demerphq]

Ok, done. It just needed threads.pm updated. You should have been able to see the test failure with make test_porting.

[twata1]

Thank you very much. It is very helpful.

I have just looked at the dist/threads/lib/threads.pm that you have updated for me.
Indeed, I think we should update the version once we fix the module.
I was having trouble with t/porting/cmp_version.t PASSing on my Windows environment without versioning it up, and trying to re-build the blead but the build failed.

Anyway, Thank you for taking care of me!

[demerphq]

My pleasure. Thanks for contributing! Yves

…

On Sat, 10 Dec 2022 at 03:51, twata1 ***@***.***> wrote: Thank you very much. It is very helpful. I have just looked at the dist/threads/lib/threads.pm that you have updated for me. Indeed, I think we should update the version once we fix the module. I was having trouble with t/porting/cmp_version.t PASSing on my Windows environment without versioning it up, and trying to re-build the blead but the build failed. Anyway, Thank you for taking care of me! — Reply to this email directly, view it on GitHub <#20116 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAZ5RZ2YQJTO6ITFRAACZTWMPV3LANCNFSM5652JRNA> . You are receiving this because you commented.Message ID: ***@***.***>

-- perl -Mre=debug -e "/just|another|perl|hacker/"

[demerphq]

And merged. Thanks again.