Le Mer 3 Mai 2006 14:02, Alexander Sack a écrit : > Anyway, I would have preferred to roll this out in a different > fashion: > > 1. release it as a package so all maintainers can use the tool if > they like. I intended it like that at the begining. sadly, the tool uses a local mirror of the debbugs .summary files (takes 450Mb of disk) + some specific things. The project bts-link inherited from had this approach, but it meant that: usertags were user specified (so to see gnome remote bugs status you would have to guess which user they chose, and then again for kde crew, ...) > 2. if this shows that indeed it works as expected, launch a central > service one can opt-in. Sure, if one maintainer is too bothered, I can deactivate the bugzilla from his upstream, and he won't be disturbed at all. I expect some small rant, but mostly, when you "forward" a bug, it's because you don't have the time/power/... to fix it, so you usually are *really* happy if someones helps you in triaging. I do not touch any other kind of bugs *AT ALL*. and soon, control processings will be archived on bts-link-upstream@lists.alioth.debian.org, so that you can grep easily what it did to your package. I also have the logs. So I'll even revert spurious changes myself if needed, to avoid any loss of time for the developpers. I agree with your remark, and it's also how I wanted to do things in the first place, but that would only have delayed the tool achievement, with no real gain in QA. So it appeared (and still does) to be acceptable the way I did. -- ·O· Pierre Habouzit ··O madcoder@debian.org OOO http://www.madism.org
Attachment:
pgpl1o8v96SUX.pgp
Description: PGP signature