On Fri, Aug 01, 2003 at 10:06:39PM +0200, Arnaud Vandyck wrote: > Matt Zimmerman <[EMAIL PROTECTED]> wrote: > > On Fri, Aug 01, 2003 at 07:50:15PM +0200, Arnaud Vandyck wrote: > [...] > > > If there are RC bugs to packages that 'release-status-sarge' depends > > > on, it won't go to testing... > > > > Of course it would, unless it had a versioned dependency that could > > not be met. And how would you know in which version the bug would be > > fixed? > > 'release-status-sarge' is just a package to monitor the work to be done > to have a stable release. > > It does not matter to know in which version the bug will be fixed. What > I want for sarge is emacs21 ( >= 21.2 ) so if every RC bugs are closed > with 21.3 or 21.4, the dependency >=21.2 is ok.
And what if the version in testing has an RC bug? "release-status-sarge" says everything is OK. > What I think is interresting with my proposal is that the release > happens when packages we want for the next stable release are ready, > stable. I am saying that the reality of the situation is more complex than is accounted for in this approach. > Don't you agree with a way of monitoring the steps to be done to the > next stable release? > > Maybe you exactly know where Debian goes and what we are waiting for > (yes I saw the mails about gnome2, kde3, gcc3.3, etc...)? I do not. I do not think that version number milestones are important for a release. I think that having a well-integrated, high-quality distribution is important for a release, and this is not so easily monitored. -- - mdz