Peter Makholm skrev: > Jonas Smedegaard <[EMAIL PROTECTED]> writes: >=20 >> Looks sane to me, generally. But... >> >>> +ifneq ($(DEB_BUILDDIR),$(DEB_SRCDIR)) >>> +$(error DEB_BUILDDIR and DEB_SRCDIR must be the same for Perl builds= ) >>> +endif >> If I understand Fr=3DC3=3DA9d=3DC3=3DA9ric correctly, the above restri= ction is not >> necessary when using Build directly. I haven't tested, so this is just=
>> speculation for now... >=20 > Wasn't sure about this either. So I just left it in for both systems. Ah, ok. Makes sense :-) >> Another thing is worrying about existing packages - do we not risk >> breaking some current use of CDBS by changing this now? >> >> If so, it might be better to check for Build.PL but instead of setting= >> DEB_PERL_BUILDSYSTEM only warn about it. >=20 > I would really hate having a long term solution prefering a > compatability layer over using Module::Build directly. Agreed, but we must remain backwards compatible: # TODO: Favor Module::Build over MakeMaker DEB_PERL_BUILDSYSTEM ?=3D MakeMaker $(and $(filter MakeMaker,$(DEB_PERL_BUILDSYSTEM)),$(wildcard Build.PL),$(warning use of MakeMaker is deprecated, please set DEB_PERL_BUILDSYSTEM=3DModule::Build) (the last 3 lines above as a single line). When (or if) it is some day decided to bump the cdbs ABI, we can favor Module::Build and only use MakeMaker as fallback. Is that acceptable? (do my reasoning make sense at all?) - Jonas --=20 * Jonas Smedegaard - idealist og Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ - Enden er n=E6r: http://www.shibumi.org/eoti.htm
signature.asc
Description: OpenPGP digital signature