Re: Bug reports and the incoming queue
On Mon, Mar 24, 2003 at 06:53:44PM -0500, John Ashley Burgoyne wrote:
> Evolution having been non-upgradeable/uninstallable in sid for a few
> days now because of a missing dependency on libgal23, I looked to see
> whether anybody had submitted the bug. I saw that somebody has indeed
> submitted it, and moreover, it has been resolved. The "resolution"
> message from the maintainer, however, is a curt request never to submit
> such bugs because libgal23 is already in the incoming queue.
I agree with Henrique; the maintainer is just unlucky and needs to
accept such bugs. Furthermore, the maintainer ought not to have uploaded
evolution built against libgal23 knowing that libgal23 was still stuck
in NEW.
> On the other hand, I am not a Debian developer and had half forgotten
> that the incoming queue existed. The last time I found it, it was after
> an hour of searching on Google. This time it only took a single search
> from which, I *think* I found it (http://incoming.debian.org). I don't
> see libgal23 there anywhere, though, which leads me to believe that
> maybe I haven't found it after all. I also seem to remember from
> my last search for the incoming queue that there is some point in the
> life cycle of a package upgrade during which the developer has released
> the new package but it is hidden from normal users so that they don't
> jump the gun and install it before all of its dependencies exist.
Not quite. Packages with entirely new names (as opposed to new versions
of existing packages) go into queue/new (sometimes called just NEW)
until an ftpmaster has manually checked them. They are unreadable to
everyone except ftpmasters (even most developers) until this check has
happened, at which point they appear in http://incoming.debian.org/
until the next daily archive update. This unreadability is a legal
requirement of having cryptographic software in the main archive.
> What is one supposed to do, then, to be a conscientious bug submitter? I
> always check the other bug reports first, and I would be happy to check
> the incoming queue, too, if I could get a straight answer about how
> exactly I can (or can't) do that.
The submitter did the right thing. Checking incoming.debian.org is
useful where possible, but maintainers should not flame people who don't
do this.
Cheers,
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: