clone 462678 -1 reassign -1 kghostview retitle -1 Dependency should be updated from gs to ghostscript-x severity -1 important thanks
On Sunday 27 January 2008, brian m. carlson wrote: > Ghostscript's x11 devices are in ghostscript-x. gs-gpl and gs-esp both > depend on both ghostscript and ghostscript-x, so there's no backward > compatibility issue. > > It looks like kghostview depends on gs, which is a very old package > name (it's what gs-gpl used to be called, and what gs-gpl provided), and > what ghostscript provides now. kghostview probably shouldn't be using > that name anymore, and should depend on the new packages. Thanks for the info. Installing ghostscript-x solved the issue. I agree that gs is old, but it was still a valid dependency for Etch, so the transition to ghostscript should ensure validity as well. I've analyzed the problem, and this is what happened: - I upgraded to ghostscript without problems and got both ghostscript and ghostscript-x installed - I later removed the transition packages, and for some reason removed ghostscript-x too (probably thinking that it was a viewer for X and thus not needed as I already had kghostview) There seem to be two problems: 1) kghostview should indeed update its dependency to ghostscript-x 2) the "ghostscript Provides gs" is wrong because ghostscript does not provide the full functionality of gs: it is missing the x11 output device support I'm cloning this BR to kghostview for 1). However 2) is something that should be fixed in the new ghostscript packages. IMO ghostscript-x should be providing gs, gs-esp and gs-gpl, and not ghostscript. If that had been the case, I would not have been allowed to remove ghostscript-x. Cheers, FJP P.S. Would it make sense for the ghostscript maintainers to do an inventory of which packages still depend on the (very) old packages and file BRs with a request to change their deps?
signature.asc
Description: This is a digitally signed message part.