Mailing-List: contact stdcxx-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: stdcxx-dev@incubator.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Message-ID: <46C4D227.80307@roguewave.com>
Date: Thu, 16 Aug 2007 16:39:35 -0600
From: Martin Sebor <sebor@roguewave.com>
Organization: Rogue Wave Software, Inc.
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US;
 rv:1.8.1.6) Gecko/20070802 SeaMonkey/1.1.4
MIME-Version: 1.0
To: stdcxx-dev@incubator.apache.org
Subject: Re: Windows build results in the pink
References: <46B9EAEB.60607@roguewave.com> <46B9F67A.1020700@roguewave.com>
 <46BA46AB.1080404@roguewave.com> <46C07DBA.7030103@roguewave.com>
 <46C4C108.40902@roguewave.com> <46C4CEF1.9050606@roguewave.com>
In-Reply-To: <46C4CEF1.9050606@roguewave.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Actually, the test case is even simpler:

$ echo "1.a" | /usr/xpg4/bin/awk '$1 == "1.A" { print $1 }'
1.a

Looks like the POSIX awk on Solaris/x86 doesn't correctly
handle strings with a numeric prefix.

Martin Sebor wrote:
> Andrew and I looked into it some more and it turned out to
> be a bug in the Solaris/x86 implementation of the POSIX awk
> (/usr/xpg4/bin/awk):
> 
> $   uname -srp && echo "1.abc" \
>   | /usr/xpg4/bin/awk '$1 == toupper($1) { print $1 }'
> SunOS 5.10 i386
> 1.abc
> $
> 
> $   uname -srp && echo "1.abc" \
>   | /usr/xpg4/bin/awk '$1 == toupper($1) { print $1 }'
> SunOS 5.10 sparc
> $
> 
> I opened a bug at http://bugs.opensolaris.org/.
> 
> Martin
> 
> Andrew Black wrote:
>> Greetings Martin, all.
>>
>> After arguing with the nightly testing system to get it to keep usable
>> post-build archives around, I finally managed to track down the cause of
>> the failure on Solaris AMD64. The problem, perhaps not surprisingly, is
>> the awk implementation used.  (See
>> http://mail-archives.apache.org/mod_mbox/incubator-stdcxx-dev/200707.mbox/%3c46AD0F69.6090706@roguewave.com%3e 
>>
>> , which is continued as
>> http://mail-archives.apache.org/mod_mbox/incubator-stdcxx-dev/200708.mbox/%3c46B25BE0.5010607@roguewave.com%3e 
>>
>>  for reference)
>>
>>> [ablack@sunspot source-buildspace]$ awk -f parse_runlog.awk 
>>> runall.log | tail -7
>>> warnings=0
>>> tests.run=0
>>> tests.good=0
>>> examples.run=0
>>> examples.good=0
>>> locales.run=0
>>> locales.good=0
>>> [ablack@sunspot source-buildspace]$ /usr/xpg4/bin/awk -f 
>>> parse_runlog.awk runall.log | tail -7
>>> warnings=0
>>> tests.run=0
>>> tests.good=0
>>> examples.run=135
>>> examples.good=129
>>> locales.run=166
>>> locales.good=162
>>> [ablack@sunspot source-buildspace]$ nawk -f parse_runlog.awk 
>>> runall.log | tail -7
>>> warnings=75
>>> tests.run=341
>>> tests.good=275
>>> examples.run=135
>>> examples.good=129
>>> locales.run=166
>>> locales.good=162
>>
>> At this time we are using /usr/xpg4/bin/awk on both sparc and AMD64.
>> Any thoughts?
>>
>> --Andrew Black
>>
>> Andrew Black wrote:
>>> Greetings Martin, all.
>>>
>>> I had time on Friday to investigate the problem with the windows parser
>>> script, and checked in a change which resolved the issue.  I have not
>>> looked into the issue with result parsing on solaris 10 for AMD64.
>>>
>>> --Andrew Black
>>>
>>> Andrew Black wrote:
>>>> Greetings Martin
>>>>
>>>> I am currently occupied with another high-priority project, and hope to
>>>> finish it in the next couple days.  I'll look into these problems once
>>>> the other project has been dealt with.
>>>>
>>>> --Andrew Black
>>>>
>>>> Martin Sebor wrote:
>>>>> A similar problem exists in Solaris 10/AMD64 build results. There,
>>>>> examples and locales look like they are being reported correctly
>>>>> but tests are not.
>>>>>
>>>>> Martim
>>>>>
>>>>> Martin Sebor wrote:
>>>>>> Our latest Windows results are still having problems. I suspect it's
>>>>>> the fallout of the exec changes last week and the parsing script not
>>>>>> having been (fully) updated to adjust. Andrew, this is rather urgent
>>>>>> for us. Can you please look into it and let us know when we can
>>>>>> expect a fix?
>>>>>>
>>>>>> Thanks
>>>>>> Martin
>>>>>>
>