On Fri, Nov 30, 2012 at 11:28:12AM +0000, Stuart Henderson wrote: > On 2012/11/30 05:47, Brad Smith wrote: > > ----- Original message ----- > > > On 2012/11/28 23:14, Brad Smith wrote: > > > > On Thu, Nov 29, 2012 at 12:14:33AM +0600, Alexandr Shadchin wrote: > > > > > On Mon, Nov 19, 2012 at 11:33:09PM +0600, Alexandr Shadchin wrote: > > > > > > Hi, > > > > > > > > > > > > This update package deadbeef to the latest release 0.5.6. > > > > > > Tested on amd64. > > > > > > > > > > > > Split package on -main, -gtk2 and -gtk3. > > > > > > Please make review this change. > > > > > > > > > > > > Comments ? OK ? > > > > > > > > > > > > > > > > Fix BUILD_DEPENDS. > > > > > > > > The gtk2 sub-package should have a @pkgpath marker so upgrading > > > > from the older package to the new main + gtk2 package set will > > > > have the expected result. > > > > > > good luck getting that to work ;) > > > > Am I crazy for expecting that to actually work? I was fairly > > certain it would but maybe it is another thing with the pkg tools > > that doesn't work. > > The gtk2 package is named deadbeef-gtk2-0.5.6, so it isn't considered > as a replacement for deadbeef-0.5.5, so pkg_add -u doesn't consult the > deadbeef-gtk2 package at all to even see the @pkgpath line. > > One thing that could be done would be to name the -main package > something like deadbeef-core-0.5.6, name the -gtk2 package deadbeef-0.5.6, > remove @pkgpath in PLIST-main, add @pkgpath in PLIST-gtk2, and add @conflicts > as necessary.
Ewww. That's pretty awful. > More than this would need Quirks.pm to be able to do version comparisons > rather than the fast hash lookups, but doing those checks for every > updated package will slow down pkg_add -u .. Or better yet get rid of the nonsense of splitting up the port and not having to worry about this stuff at all. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
