On Tue, Jun 23, 2015 at 06:51:51PM +0200, Emilio Pozuelo Monfort wrote: > On 23/06/15 15:34, Antonio Terceiro wrote: > > On Tue, Jun 23, 2015 at 02:41:36PM +0200, Emilio Pozuelo Monfort wrote: > >> On 22/06/15 15:59, Antonio Terceiro wrote: > > [...] > >>> qdbm > >>> remctl > >>> rrdtool > >>> ruby-fcgi > >>> ruby-filesystem > >>> ruby-god > >>> ruby-narray > >>> ruby-odbc > >>> ruby-rmagick > >>> ruby-sdl > >>> ruby-taglib2 > >>> ruby-uconv > >>> stfl > >>> hyperestraier > >>> libguestfs > >>> mapserver > >>> ruby-hdfeos5 > >>> ruby-mpi > >>> ruby-netcdf > >>> ruby-passenger > >>> ruby-redcarpet > >>> thin > >> > >> Scheduled. > > > > Thanks > > Looks like a few of these were also done in round 1. Please be careful there > so > we don't waste buildd time. > > Anyway, those are mostly done.
That's weird since I compiled that list starting from the packages listed as bad in the transition page itself. > >>> All others either FTBFS, or just use the default ruby (which should > >>> be fine), or don't use the supported versions reported by > >>> ruby-defaults/gem2deb. They will need to be looked at individually. > >>> > >>> I should be able to workaround a large part of the FTBFS by adding a > >>> dependency to gem2deb, and after that I will be able to file FTBFS bugs. > >>> > >>> Additionally, can you please adjust the ben file to remove false > >>> positives on arch:all packages? > >>> > >>> Affected: .architecture ~ /any/ & .depends ~ /libruby/ > >> > >> Why? Aren't the arch:all ones something that should be dealt with as well? > >> Not > >> through binNMUs obviously, but the tracker lets you know when they have > >> been > >> fixed or how many stuff is still using other ruby versions. > >> > >> Actually from a closer look I see that those two depend on libruby, but > >> not on > >> libruby2.X. So they shouldn't be tracked because of that. I've fixed the > >> is_affected regex to look for /libruby2/, which fixed that. > > > > Your solution is indeed better because it would catch any arch:all > > packages that depend on specific versions (but shouldn't). Thanks again. > > I added another tweak to is_bad as it was reporting packages that had > > Depends: [...], libruby2.1 (>= 2.1.0), libruby2.2 (>= 2.2.0~1), [...] > > as bad. Those are now marked as good. Thanks. I should have a next round soon. -- Antonio Terceiro <terce...@debian.org>
signature.asc
Description: Digital signature
