Mailing-List: contact stdcxx-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: stdcxx-dev@incubator.apache.org
Received-SPF: pass (athena.apache.org: local policy)
Message-ID: <46B25631.3050609@roguewave.com>
Date: Thu, 02 Aug 2007 16:09:53 -0600
From: Martin Sebor <sebor@roguewave.com>
Organization: Rogue Wave Software
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US;
 rv:1.8.1.2) Gecko/20070221 SeaMonkey/1.1.1
MIME-Version: 1.0
To: stdcxx-dev@incubator.apache.org
Subject: Re: many builds in DATA status
References: <46B22BEA.7050101@roguewave.com> <46B23644.1040000@roguewave.com>
In-Reply-To: <46B23644.1040000@roguewave.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Andrew Black wrote:
> Greetings Martin
> 
> I observed the Windows issues earlier today when I glanced at the
> results, but I didn't spot the Solaris failures when I was making that
> pass.  The reason for the GCC builds on windows are processed correctly
> is because they use the UNIX infrastructure, rather than the Windows
> infrastructure.
> 
> The cause of the failure on Solaris appears to be a lack of
> functionality in the /usr/bin/awk, specifically the lack of a toupper()
> builtin.  Our options on this platform appear to be a choice between
> using /usr/bin/nawk (aka /usr/xpg4/bin/awk),

The latter is the POSIX awk. That's the one we need if we're using
POSIX features that aren't available in the Solaris implementation
of awk. I suggest you do that as a quick stopgap solution until we
settle on the final format of exec output.

> and rewriting the parser to
>  avoid the use of the toupper() (and tolower()) builtins.  A third
> option I see (altering the exec utility) I'll go into more detail
> elsewhere (specifically in the bulletproofing thread).

I think we will do that regardless. Relying on strings in all caps
seems too fragile. I agree that the bulletproofing thread is the
right place to discuss these details.

> 
> I'm less certain about the cause of the failure on windows, but I
> suspect it to be basically the same as the cause of failure on Solaris -
> that of the start/stop parsing tags not being picked up.

We need to do the same thing there.

Martin

> 
> --Andrew Black
> 
> Martin Sebor wrote:
>> There are large number of builds in DATA status in recent build
>> results: essentially, all Solaris and Windows builds (with gcc
>> on Windows being the notable exception). I suspect they might
>> have been caused by our recent changes to the run/publish script.
>>
>> Andrew, can you please confirm and if so look into it ASAP?
>>
>> Thanks
>> Martin