search

Perl / perl5

🐪 The Perl programming language
https://dev.perl.org/perl5/
Other
1.93k stars 549 forks source link

[PATCH] Time-HiRes: Fix generating Makefile when it does not exist #16533

Open p5pRT opened 6 years ago

p5pRT commented 6 years ago

Migrated from rt.perl.org#133153 (status was 'open')

Searchable as RT133153$

p5pRT commented 6 years ago

From @pali

More times when compiling/updating perl codebase I'm getting following compile error​:

./miniperl -Ilib make_ext.pl lib/auto/Time/HiRes/HiRes.so MAKE="make" LIBPERL_A=libperl.a LINKTYPE=dynamic Makefile.PL​: The "xdefine" exists\, skipping the configure step. Use "/home/pali/perl/miniperl Makefile.PL --configure" or​: "/home/pali/perl/miniperl Makefile.PL --force to force the configure step. Warning​: No Makefile! make[1]​: Entering directory '/home/pali/perl/dist/Time-HiRes' make[1]​: *** No rule to make target 'all'. Stop. make[1]​: Leaving directory '/home/pali/perl/dist/Time-HiRes' make[1]​: Entering directory '/home/pali/perl/dist/Time-HiRes' make[1]​: *** No rule to make target 'all'. Stop. make[1]​: Leaving directory '/home/pali/perl/dist/Time-HiRes' Unsuccessful make(dist/Time-HiRes)​: code=512 at make_ext.pl line 570. makefile​:582​: recipe for target 'lib/auto/Time/HiRes/HiRes.so' failed make​: *** [lib/auto/Time/HiRes/HiRes.so] Error 2

Reason is that Makfile.PL does not generate Makefile if --configure or --force is not specified. And it does not generate it even when Makefile does not exist.

It is really ridiculous that Makefile.PL does not generate Makefile without specifying some non-standard flags which even make_ext.pl does not support.

This patch ensures that Makefile is always generated when Makefile.PL is called despite which arguments were passed. So it fixes above problem with compilation.

p5pRT commented 6 years ago

From @pali

0001-Time-HiRes-Fix-generating-Makefile-when-does-not-exi.patch ```diff From 1cbf6f0d9e6e3f60774bdc01635c1dc63600ebcb Mon Sep 17 00:00:00 2001 From: Pali Date: Wed, 25 Apr 2018 13:19:42 +0200 Subject: [PATCH] Time-HiRes: Fix generating Makefile when it does not exist More times when compiling/updating perl codebase I'm getting following compile error: ./miniperl -Ilib make_ext.pl lib/auto/Time/HiRes/HiRes.so MAKE="make" LIBPERL_A=libperl.a LINKTYPE=dynamic Makefile.PL: The "xdefine" exists, skipping the configure step. Use "/home/pali/perl/miniperl Makefile.PL --configure" or: "/home/pali/perl/miniperl Makefile.PL --force to force the configure step. Warning: No Makefile! make[1]: Entering directory '/home/pali/perl/dist/Time-HiRes' make[1]: *** No rule to make target 'all'. Stop. make[1]: Leaving directory '/home/pali/perl/dist/Time-HiRes' make[1]: Entering directory '/home/pali/perl/dist/Time-HiRes' make[1]: *** No rule to make target 'all'. Stop. make[1]: Leaving directory '/home/pali/perl/dist/Time-HiRes' Unsuccessful make(dist/Time-HiRes): code=512 at make_ext.pl line 570. makefile:582: recipe for target 'lib/auto/Time/HiRes/HiRes.so' failed make: *** [lib/auto/Time/HiRes/HiRes.so] Error 2 Reason is that Makfile.PL does not generate Makefile if --configure or --force is not specified. And it does not generate it even when Makefile does not exist. It is really ridiculous that Makefile.PL does not generate Makefile without specifying some non-standard flags which even make_ext.pl does not support. This patch ensures that Makefile is always generated when Makefile.PL is called despite which arguments were passed. So it fixes above problem with compilation. --- dist/Time-HiRes/Makefile.PL | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/dist/Time-HiRes/Makefile.PL b/dist/Time-HiRes/Makefile.PL index e0f7dd9a2f..bab8276929 100644 --- a/dist/Time-HiRes/Makefile.PL +++ b/dist/Time-HiRes/Makefile.PL @@ -1068,9 +1068,9 @@ sub main { } else { init(); } - doMakefile; - doConstants; } + doMakefile; + doConstants; my $make = $Config{'make'} || "make"; unless (exists $ENV{PERL_CORE} && $ENV{PERL_CORE}) { print <
p5pRT commented 6 years ago

From csjewell@cpan.org

On Wed\, 25 Apr 2018 04​:30​:36 -0700\, pali@​cpan.org wrote​:

More times when compiling/updating perl codebase I'm getting following compile error​:

./miniperl -Ilib make_ext.pl lib/auto/Time/HiRes/HiRes.so MAKE="make" LIBPERL_A=libperl.a LINKTYPE=dynamic Makefile.PL​: The "xdefine" exists\, skipping the configure step.

The patch is not the correct way to fix the problem - it only symptom-fixes\, it does not fix the cause of the problem. The cause is that the "xdefine" file still exists between builds. I would ask why a 'make realclean' or 'make clean' (whatever you are doing between builds that prepares it for rebuilding) is not removing the "xdefine" file (which is the flag that says 'I configured already')\, instead.

p5pRT commented 6 years ago

From [Unknown Contact. See original ticket]

On Wed\, 25 Apr 2018 04​:30​:36 -0700\, pali@​cpan.org wrote​:

More times when compiling/updating perl codebase I'm getting following compile error​:

./miniperl -Ilib make_ext.pl lib/auto/Time/HiRes/HiRes.so MAKE="make" LIBPERL_A=libperl.a LINKTYPE=dynamic Makefile.PL​: The "xdefine" exists\, skipping the configure step.

The patch is not the correct way to fix the problem - it only symptom-fixes\, it does not fix the cause of the problem. The cause is that the "xdefine" file still exists between builds. I would ask why a 'make realclean' or 'make clean' (whatever you are doing between builds that prepares it for rebuilding) is not removing the "xdefine" file (which is the flag that says 'I configured already')\, instead.

p5pRT commented 6 years ago

From @bulk88

On Wed\, 25 Apr 2018 09​:17​:47 -0700\, csjewell@​cpan.org wrote​:

The patch is not the correct way to fix the problem - it only symptom- fixes\, it does not fix the cause of the problem. The cause is that the "xdefine" file still exists between builds. I would ask why a 'make realclean' or 'make clean' (whatever you are doing between builds that prepares it for rebuilding) is not removing the "xdefine" file (which is the flag that says 'I configured already')\, instead.

I agree\, why is there a build product on disk in the first place? .gitignore problem or makefile clean target problem?

-- bulk88 ~ bulk88 at hotmail.com

p5pRT commented 6 years ago

The RT System itself - Status changed from 'new' to 'open'

p5pRT commented 6 years ago

From @pali

make_ext.pl in more cases remove Makefile file and then expect that calling Makefile.PL without any additional argument would generate it again. This is working fine for all subparts except Time-HiRes which does not generate Makefile again.

If you think that it is incorrect\, then (re)define Makefile.PL API under which conditions is Makefile being generated and fix scripts around which calls Makefile.

p5pRT commented 5 years ago

From @tonycoz

On Thu\, 26 Apr 2018 07​:12​:31 -0700\, pali@​cpan.org wrote​:

make_ext.pl in more cases remove Makefile file and then expect that calling Makefile.PL without any additional argument would generate it again. This is working fine for all subparts except Time-HiRes which does not generate Makefile again.

If you think that it is incorrect\, then (re)define Makefile.PL API under which conditions is Makefile being generated and fix scripts around which calls Makefile.

I think the correct behaviour here would be to always probe\, rather than making it conditional on xdefine being missing.

Tony

p5pRT commented 5 years ago

From @pali

On Tuesday 22 January 2019 19​:44​:02 Tony Cook via RT wrote​:

I think the correct behaviour here would be to always probe\, rather than making it conditional on xdefine being missing.

This should work.