On Fri, Feb 22, 2013 at 09:52:38AM +0100, Piotr Ożarowski wrote: > > So it seems to me that passing 'skip-build=1' by default is an interface > > change that should be reverted, to maximize compatibility with existing > > build > > systems.
> You do realize it's a bug in distutils/setuptools/distribute and not in > pybuild, right? No, I don't realize that all. Distutils etc. may not be well designed, but the bug here is that pybuild is passing a non-default option that's incompatible with existing uses. I don't see how that's a distutils problem. OTOH, I also don't know why pybuild is passing skip-build=1, except as an optimization? > All I can do is to add yet another workaround (.pydistutils.cfg one is > already a dirty hack). I'm not sure if I want to, I'd rather suggest you > to add to debian/rules: > export PYBUILD_INSTALL_ARGS=build > or use PYBUILD_BEFORE_INSTALL / --before-install > Do you know how many packages it affects? Maybe it's worth making the > build longer if that's what many upstream authors rely on (I wouldn't be > surprised to be honest - distutils is a mess) There are 39 packages in unstable that build-depend on python-distutils-extra, and it looks like about half of them use build_i18n. The build_icons and build_help classes are also affected. So depending on why skip-build=1 is being set, it's possible that it's not worth pybuild being compatible out-of-the-box with distutils-extra. But it would be nice if the universal python builder could handle this case. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slanga...@ubuntu.com vor...@debian.org
signature.asc
Description: Digital signature
