On 11/05/11 at 15:58 +0200, Vincent Fourmond wrote: > [putting back debian-ruby, I don't understand how it got lost] > > On 11/05/11 14:59, Lucas Nussbaum wrote: > > On 11/05/11 at 14:56 +0200, Vincent Fourmond wrote: > >> On 11/05/11 14:50, Lucas Nussbaum wrote: > >>> But for rmagick, it seems that the hooks only deal with documentation > >>> building. Wouldn't it be simpler to just run those scripts in the > >>> correct order in debian/rules? > >> > >> Of course, it is still possible, but it requires quite a bit of > >> cumbersome patching of the setup scripts, which I'd rather avoid. > > > > Why? You could call the hooks directly. > > No. They won't work, as the hooks are run with a series of > setup.rb-supplied configuration informations. Those need to be supplied, > probably once for each ruby version, and that makes that really painful. > > >> Moreover, these hooks are used both for documentation building and as a > >> very comprehensive test suite. It should build for every version of ruby > >> supported. > > > > That doesn't sound hard to do within the existing framework. > > ? I don't have a clue of how to do that, at least sanely. For one > thing, I don't want to hardcode the list of supported ruby versions in > debian/rules. > > Moreover, supporting the two main upstream build systems doesn't sound > like a stupid idea from the first place.
I don't think that setup.rb is mainstream anymore, given that everybody is using rubygems. I did look for gems using setup.rb hooks, and rmagick was the only one I could find. So if you do a rmagick-specific hack, I think it's fine. Lucas -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
