On Sun, May 05, 2002 at 13:48:18 +0100, Richard Kimber wrote:
> Can anyone point me to a definition of what a "release critical bug" is

A bug that makes a package unsuitable for release, i.e. that has severity
"critical", "grave" or "serious".

> and who can determine whether a bug is one?

Determine whether it qualifies for one of those severities; see
http://www.debian.org/Bugs/Developer#severities .

> The number of release critical bugs has gone up by about 20 in the last
> week, yet some of these do not seem to me to be so important that they
> should delay the official release of 3.0

A release-critical bug is a bug that makes a package unsuitable for release.
That does not automatically imply that such a bug holds up the release. The
release manager can, and does occasionally, remove packages with
release-critical bugs from the candidate release tree.

> Is it really so critical that (to take #145845) gnome-chess won't build on
> an ia64?

Yes. First of all we don't treat non-x86 as second class citizens. Second,
that problem wasn't ia64-specific. It affected source builds on every
platform, including x86. A fixed version is now in Incoming.

> Or that gnomesword (#145262) has a dependency that Woody doesn't satisfy?

IMO yes, but that opinion isn't shared by the release manager (see
http://lists.debian.org/debian-devel/2002/debian-devel-200205/msg00206.html).

> There will always be some problems with some applications, but my idea of
> "release critical" pertains more to the essential functioning of the
> system and to the major applications that a high proportion of users
> actually use. These examples make me feel that the definition is perhaps a
> little too broad.

Yes, and these discussions remind me of back-seat drivers.
http://www.debian.org/devel/join/ is your friend.

Ray
-- 
"A.O.L.. C.I.A.. NSA. Whatever. They all have three letters. They all
collect information. And they all screw the public."
        Evil Crud Puppy in
        http://www.userfriendly.org/cartoons/archives/00feb/20000210.html


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to