Re: MINGW PPL build fix reported upstream ?

[Arnaud Lacombe <lacombar at gmail dot com>]

On Sun, Aug 8, 2010 at 12:26 PM, Yann E. MORIN
<yann.morin.1998@anciens.enib.fr> wrote:
> Arnaud, All,
>
> On Sunday 08 August 2010 17:06:00 Arnaud Lacombe wrote:
>> On Sun, Aug 8, 2010 at 9:16 AM, Yann E. MORIN
>> <yann.morin.1998@anciens.enib.fr> wrote:
>> > What exactly is 'likely currently broken' wrt Mingw32 ?
>> > As I don't do Windows, mingw32 is Terra Incognita for me, and I
>> > have to rely on users to report issues, and test fixes.
>> PPL ?
>> If there was a patch, there should have been a build failure.
>
> Yes, there might have been a build failure, indeed. But since we moved
> from using shared libraries to using static ones, the problem disapeared.
> And I am not fixing a bug we're not seeing anymore.
>
> Using shared libs meant we had to have a wrapper to all the tools in
> the toolchain, so we could set LD_LIBRARY_PATH to point to our own
> companion libraries (GMP, PPL, et al.).
>
> The wrapper used to be a simple POSIX shell script, but that was not
> working on shell-less systems, such as mingw32. So I made a C wrapper,
> that was portable, but much more complex than its shell counterpart.
>
> Although the wrapper was unavoidable to use shared companion libs, it
> was ugly. So we now build the companion libs staticaly, and use that.
> Thus, the issue with PPL and mingw32 is now moot.
>
I thought you were not doing MINGW32 ? How can you know if Bart's bug
disappeared or not ?

> So, three things:
> - I'm not fixing a bug we're not seeing;
> - I'm not pushing fixes upstream for that bug;
> - I'll be nuking the shared complibs and the wrapper for the next stable.
>
humm, looks there will be yet-another-harcoded-thing in the next
stable to come ;-)

> I hope this is a sufficient explanation. :-)
>
Well, I didn't ask you any explanation at first... The original mail
was not to you, but to Bart. You were CC'ed for reference, nothing
more...

 - Arnaud

--
For unsubscribe information see http://sourceware.org/lists.html#faq