Rainer Müller wrote: > Does your custom ports tree not just provide a custom implementation of > this port group? Or do the ports using this modified port group not need > any further modifications, but you do not want to copy all Portfiles to > your ports tree?
I indeed have custom Qt5 and CMake PortGroups (and a few others), and indeed I do not want to copy all ports that depend on either Qt5 or CMake (or both) into my own port tree. My own port tree only contains my own ports, and a number of ports that have local patches, updates or have been kept at an older version. Evidently, I do want all ports that build using CMake to use my PortGroup version, idem for Qt5 (and Qt4 and KDE4). So in the end I copy my PortGroup files into the rsync-based default port tree, and I don't run selfupdate anymore (I cherry-pick upgrades from the svn port tree instead). I have no idea how long I'll be able to keep that up ... Mojca: > Most often I was getting errors saying that a required variable (which > I just added to the PortGroup) was not defined. ... > Next time when I hit a problem I'll try to investigate in the spirit > of tree where the PortGroup lives. I think you're > The lookup strategy is also the same for mirror/archive sites, How many ports need to change those? > variant descriptions, livecheck Those are mostly defined in ports, I'd say, or in some cases in PortGroups. > . I am not sure whether it would make sense to > change this for all of them (for example archive sites are definitely > only local to a ports tree) or to introduce a special lookup just for > port groups... Can you give examples? I'm having trouble imagining why there would be any need to change the lookup strategy for those. R _______________________________________________ macports-dev mailing list macports-dev@lists.macosforge.org https://lists.macosforge.org/mailman/listinfo/macports-dev
