Ciaran McCreesh wrote: > On Tue, 07 Aug 2007 09:14:17 -0400 > Richard Freeman <[EMAIL PROTECTED]> wrote: >> If the patches are conditional then nobody else is affected anyway. > > And that's the issue -- this claim is incorrect. With conditional > patching everyone is affected. A common example used to be selinux > patches that were applied with 'use selinux && epatch blah'. When > maintaining a package with one of these and doing a version bump, one > had the following options: [true things shortened]
Here's how things mostly work between the Gnome herd and FreeBSD team, for Gnome packages. 1) Roy reports breakage in gnome package, provides quick fix in bugzilla 2) we make that patch conditional on fbsd-* 3) everyone works with upstream to include a proper fix in said package 4) backport patch/wait for next release 5) rinse, repeat But it really depends on what the problem is. We've tried all the different ways described in this thread and they all work. 2 conclusions : - get devs to trust one another, I know that whenever Roy sends us a freebsd patch, it's for a good reason. - work with upstream. Harder but it's the best long term solution. Wishful thinking ... yeah I know :) Cheers, Rémi -- [EMAIL PROTECTED] mailing list
