Re: [bug-gettext] Continous integration on Solaris

[Dagobert Michelsen]

Hi Daiki,

Am 12.01.2014 um 04:15 schrieb Daiki Ueno <address@hidden>:
> Dagobert Michelsen <address@hidden> writes:
> 
>> I change the configuration so the test-suite.log are preserved and
>> linked from the box in the waterfall display.
> 
> Thanks!
> 
>> Unfortunatelyl currently the build itself fails and therefore there
>> are no test results. The failing line is
> 
> Well, I wonder if the error was transient, as I can now see the log files:
> 
> https://buildfarm.opencsw.org/buildbot/builders/ggettext-solaris10-sparc/builds/5/steps/shell_4/logs/tools_test-suite.log
> 
>  FAIL: msgfmt-qt-1
>  =================
> 
>  msgfmt: Cannot convert from "ISO-8859-2" to "UTF-8". msgfmt relies on 
> iconv(). This version was built without iconv().
> 
> and looking at the failure line:
> 
>>> /opt/csw/bin/gmsgfmt: address@hidden: warning: PO file header missing or 
>>> invalid
>>>                                  warning: charset conversion will not work
>                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> I guess maybe libiconv is not available on the build environment?

I have adjusted the configure options and I am now down to 2 failed tests on 
sparc:
  
https://buildfarm.opencsw.org/buildbot/builders/ggettext-solaris10-sparc/builds/8

That is

> FAIL: xgettext-python-1
> =======================
> 
> ./xgettext-python-1: line 37: 28620 Segmentation Fault      (core dumped) 
> ${XGETTEXT} --add-comments --no-location -o xg-py-1.tmp xg-py-1.py 2> 
> xg-py-1.err

and

> FAIL: lang-c++
> ==============
> 
> In file included from /usr/include/sys/time.h:421:0,
>                  from ../../gnulib-lib/sys/time.h:38,
>                  from /usr/include/sys/select.h:23,
>                  from ../../gnulib-lib/sys/select.h:36,
>                  from /usr/include/sys/types.h:616,
>                  from ../../gnulib-lib/sys/types.h:27,
>                  from ../../gnulib-lib/stdio.h:58,
>                  from ../../gnulib-lib/wchar.h:71,
>                  from /opt/csw/include/c++/4.8.2/cwchar:44,
>                  from /opt/csw/include/c++/4.8.2/bits/postypes.h:40,
>                  from /opt/csw/include/c++/4.8.2/iosfwd:40,
>                  from /opt/csw/include/c++/4.8.2/ios:38,
>                  from /opt/csw/include/c++/4.8.2/ostream:38,
>                  from /opt/csw/include/c++/4.8.2/iostream:39,
>                  from l-c++-prog.cc:5:
> ../../gnulib-lib/stdio.h:1025:1: error: 'char* gets(char*)' conflicts with 
> previous using declaration 'char* std::gets(char*)'
>  _GL_WARN_ON_USE (gets, "gets is a security hole - use fgets instead");
>  ^


On i386 I see two more failing tests (leaving aside the flaws of our gij Java 
compiler):
  
https://buildfarm.opencsw.org/buildbot/builders/ggettext-solaris10-i386/builds/9

> FAIL: lang-bash
> ===============
> 
> 1c1
> < «Votre commande, s'il vous plait», dit le garçon.
> ---
> > �Votre commande, s'il vous plait�, dit le gar�on.

and

> FAIL: lang-vala
> ===============
> 
> 1,2c1,2
> < �Votre commande, s'il vous plait�, dit le gar�on.
> < EUR remplace FF.
> ---
> > 'Your command, please?', asked the waiter.
> > FF is replaced by EUR.

Maybe these are locale-specific?


Best regards

  — Dago

-- 
"You don't become great by trying to be great, you become great by wanting to 
do something,
and then doing it so hard that you become great in the process." - xkcd #896

smime.p7s
Description: S/MIME cryptographic signature