On 2013-01-26 22:37, Dave Steele wrote: > Provide better visibility to problems with broken symlinks in packages > by raising the condition from an issue to an error on piatti.
Another reason to run two different tests for sid: * one for serious problems (e.g. (un-)installability problems, no leftovers or broken symlinks), this needs to have a very high coverage, i.e. low dependency-failed-testing count currently 73 failed, 23 dep-failed * one for important problems (leftover files, broken symlinks) currently with leftovers check: 138 failed, 244 dep-failed > Per > http://lists.alioth.debian.org/pipermail/piuparts-devel/2013-January/003424.html > , the failed count will rise by no more than 500 due to this change, > even if nothing is done to fix the errors. I just added another section ... sid-bl (--fail-on-broken-symlinks --warn-on-leftovers-after-purge): failed: 597 failed with broken symlinks (and no other problems): 508 failed with broken symlinks and other problems: 51 dep-failed: 3846 (tested with python-support.pth being ignored) > The rise will be gradual, > taking about 6 months for all of the packages to recycle. there is always grep -l ": Broken symlink" *.log | xargs -r rm -v > Since both blockers and block-ees will migrate only as the packages > are upgraded or recycled, it should be fairly easy to keep the > dependency-failed count due to this error to well under 1000. If we want to make this error visible, we should enable it at some point *and* recycle the logs immediately *and* have a mass-bug-filing prepared (i.e. discussed+agreed on debian-devel) beforehand, and someone should be ready to file the bugs. Andreas -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
