The qsearch applet, doesn't find gtkhtml-sharp. Reproducible: Always Steps to Reproduce: 1. qsearch gtkhtml-sharp Actual Results: Empty output Expected Results: It should show the available gtkhtml-sharp versions user@host ~/ $ equery which gtkhtml-sharp /usr/portage/dev-dotnet/gtkhtml-sharp/gtkhtml-sharp-2.16.0.ebuild
I can reproduce this issue with "qsearch gtkhtml-sharp". Using "qsearch -c gtkhtml-sharp" at the contrary works fine. The reason is that DESCRIPTION is defined in an eclass, not the ebuild itself. I don't see any simple way to fix that. Making qsearch grep in the inherited eclasses would not even be enough, since here it would also need to do some bash vars evaluations: % grep DESCR /var/portage/tree/eclass/gtk-sharp-component.eclass DESCRIPTION="${GTK_SHARP_COMPONENT_BUILD} component of gtk-sharp" Maybe "qsearch -c" should simply be made the default behavior (with "q -m" being added to the rsync hook sure)?
(In reply to comment #1) > I can reproduce this issue with "qsearch gtkhtml-sharp". > Using "qsearch -c > gtkhtml-sharp" at the contrary works fine. The reason is that DESCRIPTION is > defined in an eclass, not the ebuild itself. Thanks for looking into this. > I don't see any simple way to fix that. Making qsearch grep in the inherited > eclasses would not even be enough, since here it would also need to do some > bash vars evaluations: > % grep DESCR /var/portage/tree/eclass/gtk-sharp-component.eclass > DESCRIPTION="${GTK_SHARP_COMPONENT_BUILD} component of gtk-sharp" > > Maybe "qsearch -c" should simply be made the default behavior (with "q -m" > being added to the rsync hook sure)? We can't really do that. On dev boxes we often use our CVS checkouts as PORTDIR and that wont ever contain a metadata/ dir, so a qsearch would end up returning nothing. Perhaps however maybe we could check if the metadata dir exists. If it does then default to CACHE_METADATA otherwise fallback to CACHE_EBUILD. in order to do that we however would need an override flag for all supporting applets that can use (cache || ebuilds)
Closing bug