On Thu, Dec 2, 2010 at 5:20 PM, Dave Reisner <[email protected]> wrote:
> On Thu, Dec 02, 2010 at 05:13:56PM +0100, Cédric Girard wrote: > > On Thu, Dec 2, 2010 at 4:59 PM, Xyne <[email protected]> wrote: > > > > > > > > Packages that are built from vcs but which are based on some form of > > > upstream > > > "release" should not include the tag in the package name. > > > > > > I think the simplest rule of thumb would be that if the same PKGBUILD > > > generates > > > different binary packages depending on when makepkg was run, then it > should > > > include the suffix in the name. > > > > > > > > These two rules are not the same. For instance the package xbmc-svn [1] > is > > based on fixed svn version that does not corresponds to any "release" > > upstream. It is just tested svn revisions (by the packager) as not every > > revisions are usable. > > > > [1] http://aur.archlinux.org/packages.php?ID=20156 > > > > So doesn't that just mean that we have some packages currently in > existance which break the guideline we're trying to establish? I propose > that this particular package is named incorrectly, and would be better > off as xbmc-devel. > > dave > Yes you are right I misread Xyne message and understood the two rules quoted above as different. One is just broader than the other but there is no contradiction between them. -- Cédric Girard
