On Wed, Nov 25, 2015 at 02:31:18PM +0100, John Paul Adrian Glaubitz wrote: > On 11/25/2015 01:59 PM, Antonio Terceiro wrote: > > according to > > https://packages.debian.org/search?searchon=contents&keywords=fast_xs. > so&mode=filename&suite=unstable&arch=any > > > > > > > sh4 is the only architecture where ruby-hpricot ships that file. > > It affects m68k as well though: > > drwxr-xr-x root/root 0 2015-11-23 00:38 > ./usr/lib/m68k-linux-gnu/ruby/vendor_ruby/ > drwxr-xr-x root/root 0 2015-11-23 00:46 > ./usr/lib/m68k-linux-gnu/ruby/vendor_ruby/2.2.0/ > -rw-r--r-- root/root 9476 2015-11-23 00:46 > ./usr/lib/m68k-linux-gnu/ruby/vendor_ruby/2.2.0/fast_xs.so > > > > https://buildd.debian.org/status/fetch.php?pkg=ruby-hpricot&arch=m68k&ve > r=0.8.6-5&stamp=1448236109 > > No idea why the database doesn't list the file. > > > there is something clearly broken with it, because the fast_xs.so > > file is removed during the tests (debian/ruby-tests.rb), instead > > of in debian/rules after the build; was the original ruby-hpricot > > build on sh4 done with tests disabled? > > > > Also, ruby-hpricot is deprecated and we want to remove it from > > unstable as soon as possible. If you can "fix" this by rebuilding > > it on sh4, please do it, because I don't expect to gain any > > maintainance other than removal. > > Yes, we usually build with "nocheck nobench" on slower architectures. > That shouldn't really have any effects on the resulting debs. Please > make sure other ruby packages don't show similar behaviour as to avoid > such issues in the future.
this is not a common practice at all, seems to have been done in this package as a quick hack without much thought. > I will rebuild ruby-hpricot on m68k and sh4 with test enabled in the > mean time to have this issue resolved. cool -- Antonio Terceiro <[email protected]>
signature.asc
Description: PGP signature
