Re: Bug reports and severity
On Wed, Jun 15, 2005 at 04:29:05PM +0200, Aurelien Jarno wrote:
> Hi all,
>
> I wanted to have your opinion on the bugs reports we do for GNU/kFreeBSD
> and their severity. I think if we want to be able to release with Etch
> we need a more agressive policy.
>
> Currently I am reporting bugs with "Severity" wishlist (after all this
> is an unsupported architecture). However, I should admit that I don't
> have a lot of success (bugs waiting for months).
>
> I have seen that the Hurd guys are more agressive, they are reporting
> them with severity "Important", and they seems to have more success. For
> example with the gdm package, their patch was accepted in a short period
> of time, whereas we are *still* waiting for more than 300 days. Our bug
> is "Wishlist", their is "Important".
>
> I am therefore proposing the following policy, in the case a *patch is
> provided*:
> - severity "Normal" if the package was never built before on
> GNU/kFreeBSD
> - severity "Important" if the package was built before on GNU/kFreeBSD,
> ie there is a regression
>
> I think we should not fill a bug report specific to GNU/kFreeBSD without
> a patch or without a clear description of what to do, ie config update
> or libtool update (I am almost considering that as a patch), because we
> don't have a developper accessible machine. Being not enough agressive
> makes a lot patches sitting in the BTS. Being too agressive makes the
> port loosing credibility.
>
> Any thoughts on such a policy?
I assume we're talking about FTBFS bugs which theoricaly should be reported
as serious (we have a few GNU/kFreeBSD bugs in the style of #307475, too).
When Marcus Brinkmann started the GNU/Hurd port, he started using the "serious"
severity for FTBFS bugs, but this was disruptive wrt the release process (e.g.
britney) so they switched to important.
But now we have "${release}-ignore" tags. How about using "serious" with
"sarge-ignore" and/or "etch-ignore" tags?
I'm adding CC to debian-release.
--
Robert Millan
Reply to: