Re: Strange testing migration excuse for xzgv
On Thu, Jan 02, 2014 at 05:13:02PM +0000, Adam D. Barratt wrote:
> Urgencies are "sticky", meaning that for any given migration the
> effective urgency used is the highest of all uploads with version
> numbers higher than the package in testing - i.e. if 1.0-1 is
> uploaded at "high", and -2 at "medium", then -2 will be eligible to
> migrate after two days. In the case of a "new" package, there is no
> version of the package in testing, so all versions in the urgency
> record are considered (as britney has no concept of what version
> used to be in testing).
OK, so the issue is that since there is currently _no_ xzgv package in
testing, it's picking up the urgency=high upload from 2006-08-31
(0.8-3sarge1) to stable-security. That seems... counterintuitive.
Is there a reason why britney doesn't just use the default upload
priority for "new" packages, and not even try to caluclate the highest
of all uploads in this case? I understand that's the same net result
as the current algorithm (which is to calculate the highest priority,
and then ignore it), but it results in a very confusing explanation of
what is going on in the testing migration excuses listing on the
packages.qa.debian.org page.
In any case, thanks for explaining what's going on. I appreciate
understanding how things in britney works "under the covers". If
people agree that this is worth fixing, I'm happy to file a bug, but
if everyone thinks it's not worth fixing, since this corner case
happens relatively rarely, I'll drop it.
Thanks!!
- Ted
Reply to: