On Sun, 2016-01-31 at 18:02 +0100, Axel Beckert wrote: > I was totally surprised to get this mail already today after 5 days > instead of after 10 days: > https://packages.qa.debian.org/l/lynx/news/20160131T163912Z.html
This is actually the expected behaviour, although sometimes surprising; see below for more details. > src:lynx migrated to testing within 5 days, despite: > > * Explicit urgency=low. I consider ignoring this setting a quite > important issues. I really wanted more exposion for this package as > it contained some changes which might break things. > > * New source package name, i.e. it's the first upload to unstable > since the package went through NEW (two uploads to experimental > inbetween, though), In general, the urgency used is the highest of all uploads to the archive with versions higher than that of the version in testing. (The urgencies list provided to britney via dak doesn't list suites, hence the comparison is in practice based on all uploads.) For a "new" package (i.e. one that is not currently in testing), the urgency used is the longer of the default urgency (currently "medium") and the urgency computed as mentioned in the previous paragraph. Looking at the excuses for yesterday gives us: <li><a id="lynx" name="lynx">lynx</a> (- to 2.8.9dev8-4) <ul> <li>Maintainer: Debian Lynx Packaging Team <li>Too young, only 4 of 5 days old <li>Ignoring high urgency setting for NEW package The "high" would have come from some combination of: adsb@franck:~$ grep "lynx .*high" /srv/release.debian.org/britney/var/data-b2/testing/Urgency lynx 2.8.4-1 high lynx 2.8.4.1b-1 high lynx 2.8.5-2sarge1 high lynx 2.8.5-2sarge2 high lynx 2.8.5-2sarge1.2 high lynx 2.8.6-2 high giving the package a "desired" urgency of "high", which was then pushed back to "medium" by britney. Regards, Adam
