Re: tiff 3.7.1-1 somehow migrated to testing
Andreas Barth <aba@not.so.argh.org> writes:
> Hi Jay,
>
> * Jay Berkenbilt (ejb@ql.org) [050124 18:30]:
>> Bottom line: 3.7.1-1 has an RC bug: the -dev package has incorrect
>> dependencies. This bug is fixed in 3.7.1-2 which was uploaded as
>> urgency low. How should we fix this? One possibility would be to
>> bump the urgency of 3.7.1-2 and let it go into testing. Another
>> possibility would be to somehow force 3.6.1-5 back into testing, but
>> it would be a real shame to have to add an epoch to the version for
>> this.
>
> Well, AFAICS, the _only_ bad effect is that some other packages may
> FTBFS till -2 reaches testing. As we didn't freeze testing till now, -2
> will almost definitly reach testing before sarge is released. So, I
> think the way to go is just to wait and see what happens. The only
> request is: Please don't upload any new version till -2 managed to make
> it to sarge for itself.
Looks like 3.7.1-2 is in testing now, so I guess someone must have
bumped the urgency.
>> Also, if 3.7.1-1 moved to testing because of the race condition I'm
>> speculating about, this should be fixed. Where should I report the
>> problem?
>
> debian-release is the right address for that.
I guess I mean that there may be a bug in the scripts that allow this
race condition to occur, and I'd like to know where to report that bug
so that it could potentially be fixed. Is that still debian-release?
Maybe this is a long known problem and it's only surprising to me
because I've never seen it before. There must be some gap between the
updates of wherever the testing scripts checks the bug counts and
where it checks what the most recent upload is and how long it's been
there. I notice that bugs from uploads get automatically closed very
quickly after an upload, but I don't have any window into the other
parts. (Maybe I'll just investigate by finding and reading the code.)
>From my limited and relatively uninformed perspective, I see three
possible courses of action: just let this be and notify debian-release
if this kind of accidental transition occurs, change the mechanism by
which bugs are closed to make it not happen until the new version
shows up in wherever the testing scripts are looking for it, or change
the testing scripts to be smarter about this type of condition. In
any case, in the absence of any specific response, I'm just going to
let this drop and be satisfied that the specific issue with tiff is
taken care of.
Thanks for the response, and thanks to whoever actually resolved the
problem.
--Jay
Reply to: