develooper Front page | perl.perl5.porters | Postings from February 2003

Re: Cygwin smoke

From:
alian
Date:
February 24, 2003 16:08
Subject:
Re: Cygwin smoke
Message ID:
3E5AB3D1.9000204@jupiter.alianet
>>I've a new w2000 at home, so I will use it for cygwin os and
> uname -a?

$ uname -a
CYGWIN_NT-5.0 mercure 1.3.20(0.73/3/2) 2003-02-08 12:10 i686 unknown 
unknown Cygwin

>>Failures:
>>
>>cygwin      [stdio/perlio]-Dcf_email=alian@cpan.org
>>cygwin      [stdio/perlio]-DDEBUGGING -Dcf_email=alian@cpan.org
>>cygwin      [stdio/perlio]-Dcf_email=alian@cpan.org -Dusemultiplicity
>>cygwin    	  [stdio/perlio]-DDEBUGGING -Dcf_email=alian@cpan.org
>>               (cont) -Dusemultiplicity
>>     lib/Test/Harness/t/strap-analyze.....FAILED at test 2
>>     lib/warnings.........................FAILED at test 414
>>../lib/Test/Harness/t/strap-analyze....dubiousD. FAILED test 2
>>
>>
>>This failure are well known from HM Brand report:
>>http://www.alianwebserver.com/perl/smoke/smoke_db.cgi?os_fil=cygwin&last_smoke_fil=18500&failure=1
> 
> 
> Yep, and it's hard to trace where it *realy* hides, but it has to do with a
> CRNL/NL mismatch. That's all. We would all be gratefull if you'd be able to
> silence it :)

I think I find something ...
I find a shortcut to see them:
PERLIO=perlio ./perl -Ilib lib/warnings.t
...
ok 412
ok 413
...
EXPECTED:
Insecure dependency in chdir while running with -T switch at - line 5.
GOT:
Insecure dependency in chdir while running with -T switch at - line 5.
not ok 414
...

It's very strange as perlio is the default for perl-5.9 no ?
This can help to track the real bug ....the failure go avay with:
PERLIO="perlio :crlf" ./perl -Ilib lib/warnings.t
So if a patch is done, I think this is with Test-Smoke that he must be done.

I presume that Cygwin build use a default value of perlio :crlf as make 
test didn't fail in command line without setting this var.
(Why perl didn't set $ENV{PERLIO} to his real value ?)

So in this case, I think we must use for cygwin smoke layers ('stdio', 
'perlio :crlf')  as for win32 smoke.

For add another comment, I think that use :crlf as default for cygwin is 
a bug. I use the env var CYGWIN=binmode, and see cr is not what I want. 
Gerrit what you think about this ?

Else I see in log file a lot of (like 200):
Use of uninitialized value in concatenation (.) or string at 
../lib/CGI.pm line 3459

>>Else I try to build with -Duseithreads, but no success. I see that 
>>nobody has send in past a result for this config.
> 
> 
> If you've got spare time, please try. Cygwin has promised to support threads,
> but it would also need time from the perl porters front.
> 
> 
>>In log, I've this:
>>
>>t/op/getpid..........................FAILED at test 0
>>panic: MUTEX_INIT.
>>With lots of others.

>>alian@mercure ~/perl/perl-current
>>$ gdb ./perl
>>(gdb) run -Ilib lib/Test/Harness/t/strap-analyze.t
>>Starting program: /home/alian/perl/perl-current/perl.exe -Ilib 
>>lib/Test/Harness/t/strap-analyze.t
>>panic: MUTEX_INIT.
>>
>>Program received signal SIGSEGV, Segmentation fault.
>>0x1000b594 in S_my_exit_jump () from 
>>/home/alian/perl/perl-current/cygperl5_9_0.dll
>>(gdb) where
>>#0  0x1000b594 in S_my_exit_jump () from 
>>/home/alian/perl/perl-current/cygperl5_9_0.dll
>>#1  0x1000b4c3 in Perl_my_failure_exit () from 
>>/home/alian/perl/perl-current/cygperl5_9_0.dll
>>#2  0x10061e29 in Perl_vcroak () from 
>>/home/alian/perl/perl-current/cygperl5_9_0.dll
....

If someone need more info, I can send them, but I really I don't know 
how debug this and where start ! I can report regular smoke, but find 
why it's fail ...

HTH,
--
Alain BARBET




nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About