On Sun, 2011-02-20 at 17:20 +0000, Adam D. Barratt wrote: > [Executive summary: britney2's attempt to avoid continually > installability checking the entire package set means it can increase > installability when removing packages] [...] > The patch does add some overhead to each migration operation as more > packages need to be checked but if this is a problem then it could be > mitigated by making the recursive check conditional on the operation in > question being a removal rather than upgrade or addition of a package.
As a very unscientific test, I set up my development instance of b2 with the live data shortly before tonight's run and ran it with the patch applied. Both runs produced the same result in equivalent time (in fact the run with the patch was ~15 seconds faster, presumably due to external influences, phases of the moon, etc.) This might be more of an issue with packages with a large number of r-deps, but I think it may be worth resolving that if/when it arises. > Comments / questions / cluesticks welcome. Regards, Adam -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
