Boost logo

Boost Interest :

Subject: Re: [Boost-cmake] [PATCH] building Boost.Python fails
From: troy d. straszheim (troy_at_[hidden])
Date: 2009-09-16 21:58:23


Ingmar Vanhassel wrote:
> Excerpts from troy d. straszheim's message of Wed Sep 16 17:08:44 +0200 2009:
>> Thanks for the patches. I tested and pushed to 1.40.0 branch of cmake repo.
>>
>> -t
>
> Will these patches land in boost SVN & its next release too?
>

Well, there was some discussion about this over on the -dev list. I
would like to have cmake removed completely from boost svn to avoid
confusion. Vladimir Prus has pointed out that the locations of the
cmakefiles violate some code-organization standards.

Due to the way boost does its releases, the trunk version would have to
be actively maintained by somebody. Apparently one can't just commit to
release without committing to trunk, as it confuses a crucial
merge-tracking tool that the release managers use, and then the cmake
effort holds up the release process and inflicts pain on people who
neither know nor care about cmake. This is bad. It is also clear that
it wouldn't be productive to further discuss version control, testing,
buildsystems, modularization etc. on the -dev list... I'm not about to
publish a "manifesto" (if you don't get the reference, don't worry) and
be a troll.

OTOH if somebody wants to take maintainership of the cmakefiles in the
boost SVN tree, great. I'll continue to push tweaks to the git repo
where you can click-click-download patchfiles at will; if the tweaks
make it to SVN first, no problem, git doesn't care.

To me, what makes most sense is, with the help of those on this list, to
just tune up the cmakefiles shortly after the release is frozen, and
release under separate cover.

-t


Boost-cmake list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk