Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Perl: porters

Win32 problems with Encode and EUMM

 

 

Perl porters RSS feed   Index | Next | Previous | View Threaded


rurban at x-ray

Apr 25, 2012, 1:51 PM

Post #1 of 6 (152 views)
Permalink
Win32 problems with Encode and EUMM

I got much more failures with MSVC9 and mingw, both DEBUGGING threaded
after a cpan session, which sets
the PERL_JSON_BACKEND and PERL_YAML_BACKEND env vars.

[perl #112612] is required to fix a Encode encxs problems with
perlglob (CORE only).
miniperl.exe is not able to call perlglob.exe, perl.exe is. I guess
it's a path issue in the system call.
Broken with both mingw and MSVC.

MB+EUMM needs unset the ENV for JSON::XS and YAML backends if run from
a win32 makefile. See the attached log.

Pod/Simple/Text.pm still has a wrong \E at line 74

mingw/msvc9 with empty PERL_JSON_BACKEND and PERL_YAML_BACKEND:
Test Summary Report
-------------------
op/fork.t
(Wstat: 0 Tests: 25 Failed: 1)
Failed test: 16
../cpan/CGI/t/tmpdir.t
(Wstat: 0 Tests: 9 Failed: 0)
TODO passed: 3-9
Files=2329, Tests=528604, 4321 wallclock secs (55.20 usr + 4.53 sys =
59.73 CPU)
Result: FAIL

Anyone else?

BTW: I'm halfway through getting rid of the win32 special make- and
config files for mingw, msys and msvc
(requiring mingw bash, using Configure as with cygwin), but this will
need some more time.
--
Reini Urban
http://cpanel.net/   http://www.perl-compiler.org/
Attachments: log.test-mingw (8.49 KB)


tony at develop-help

Apr 25, 2012, 5:15 PM

Post #2 of 6 (144 views)
Permalink
Re: Win32 problems with Encode and EUMM [In reply to]

On Wed, Apr 25, 2012 at 03:51:59PM -0500, Reini Urban wrote:
> I got much more failures with MSVC9 and mingw, both DEBUGGING threaded
> after a cpan session, which sets
> the PERL_JSON_BACKEND and PERL_YAML_BACKEND env vars.
>
> [perl #112612] is required to fix a Encode encxs problems with
> perlglob (CORE only).
> miniperl.exe is not able to call perlglob.exe, perl.exe is. I guess
> it's a path issue in the system call.
> Broken with both mingw and MSVC.
>
> MB+EUMM needs unset the ENV for JSON::XS and YAML backends if run from
> a win32 makefile. See the attached log.
>
> Pod/Simple/Text.pm still has a wrong \E at line 74
>
> mingw/msvc9 with empty PERL_JSON_BACKEND and PERL_YAML_BACKEND:
> Test Summary Report
> -------------------
> op/fork.t
> (Wstat: 0 Tests: 25 Failed: 1)
> Failed test: 16
> ../cpan/CGI/t/tmpdir.t
> (Wstat: 0 Tests: 9 Failed: 0)
> TODO passed: 3-9
> Files=2329, Tests=528604, 4321 wallclock secs (55.20 usr + 4.53 sys =
> 59.73 CPU)
> Result: FAIL
>
> Anyone else?

I haven't seen any problems building Encode with blead with MSVC 9 free.

The op/fork.t failure you list was fixed in 253649d9.

> BTW: I'm halfway through getting rid of the win32 special make- and
> config files for mingw, msys and msvc
> (requiring mingw bash, using Configure as with cygwin), but this will
> need some more time.

Consolidating the build processes is a nice idea... but I don't think
the existing Win32 are going to get removed if there's a requirement
for the builder to install bash.

Tony


rurban at x-ray

Apr 26, 2012, 11:27 AM

Post #3 of 6 (144 views)
Permalink
Re: Win32 problems with Encode and EUMM [In reply to]

On Wed, Apr 25, 2012 at 7:15 PM, Tony Cook <tony [at] develop-help> wrote:
> On Wed, Apr 25, 2012 at 03:51:59PM -0500, Reini Urban wrote:
>> BTW: I'm halfway through getting rid of the win32 special make- and
>> config files for mingw, msys and msvc
>> (requiring mingw bash, using Configure as with cygwin), but this will
>> need some more time.
>
> Consolidating the build processes is a nice idea... but I don't think
> the existing Win32 are going to get removed if there's a requirement
> for the builder to install bash.

Sure, purely optional. cmd.exe is still the default option.
--
Reini


perl.p5p at rjbs

May 2, 2012, 7:21 PM

Post #4 of 6 (138 views)
Permalink
Re: Win32 problems with Encode and EUMM [In reply to]

* Reini Urban <rurban [at] x-ray> [2012-04-25T16:51:59]
> I got much more failures with MSVC9 and mingw, both DEBUGGING threaded
> after a cpan session, which sets
> the PERL_JSON_BACKEND and PERL_YAML_BACKEND env vars.
>
> [...]
>
> MB+EUMM needs unset the ENV for JSON::XS and YAML backends if run from
> a win32 makefile. See the attached log.
>
> mingw/msvc9 with empty PERL_JSON_BACKEND and PERL_YAML_BACKEND:

Is there a ticket about this? Or can you explain more thoroughly? I'm not
sure what you're suggesting needs to be actually done.

--
rjbs
Attachments: signature.asc (0.48 KB)


perl.p5p at rjbs

May 2, 2012, 7:23 PM

Post #5 of 6 (135 views)
Permalink
Re: Win32 problems with Encode and EUMM [In reply to]

* Reini Urban <rurban [at] x-ray> [2012-04-25T16:51:59]
> Pod/Simple/Text.pm still has a wrong \E at line 74

In blead, line 74 of Pod/Simple/Text.pm is:

unless $_[0]{'Thispara'} =~ /\b\Q$link->{to}/;

I'm not sure what you mean about a \E.

--
rjbs
Attachments: signature.asc (0.48 KB)


fraserbn at gmail

May 2, 2012, 7:40 PM

Post #6 of 6 (139 views)
Permalink
Re: Win32 problems with Encode and EUMM [In reply to]

On Wed, May 2, 2012 at 11:21 PM, Ricardo Signes
<perl.p5p [at] rjbs>wrote:

> * Reini Urban <rurban [at] x-ray> [2012-04-25T16:51:59]
> > I got much more failures with MSVC9 and mingw, both DEBUGGING threaded
> > after a cpan session, which sets
> > the PERL_JSON_BACKEND and PERL_YAML_BACKEND env vars.
> >
> > [...]
> >
> > MB+EUMM needs unset the ENV for JSON::XS and YAML backends if run from
> > a win32 makefile. See the attached log.
> >
> > mingw/msvc9 with empty PERL_JSON_BACKEND and PERL_YAML_BACKEND:
>
> Is there a ticket about this? Or can you explain more thoroughly? I'm not
> sure what you're suggesting needs to be actually done.
>
>
I had this issue when building blead on WinXP, so let me expand on this
bit. If something previously set PERL_JSON_BACKEND or PERL_YAML_BACKEND --
in my case, installing Strawberry Perl -- then the MB & EUMM tests fail
spectacularly; Unsetting those variables makes everything work fine. So the
test failures that Reini sent are with those two vars unset.

I didn't get the fork.t failure over here, but I built a couple of days
ago. Maybe something changed?

Perl porters RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.