Hi :) I ported Pawel Jakub Dawidek's POSIX file system test suite to the Hurd:
https://teythoon.cryptobitch.de/fstest/ The code can be found here: http://darnassus.sceen.net/gitweb/teythoon/fstest.git/ I haven't looked closely at any failures yet, but if the results have any relation to the quality of the translators, we are doing okay and - perhaps surprisingly - tmpfs is doing better than ext2fs: ext2fs: 1797/1962 tests passed. tmpfs: 1863/1962 tests passed. When run using the prove utility included in the perl distribution, the test suite crashes an tmpfs translator. This doesn't happen with my script, not sure why. However, my script has a --fuzz option, that executes tests in a random order until the translator being tested crashes. With the current Hurd code, this can be used to trigger an assertion failure in tmpfs. I like the idea of running automated tests. I have for example a tool that updates (as in aptitude full-upgrade) a Debian/Hurd system and checks whether it still boots and works (as determined by a couple of tests). This tool notified me in advance that updating libparted0debian1 would render the system unusable. I notified Pino and he notified debian-hurd@ http://lists.debian.org/debian-hurd/2013/08/msg00024.html I've been pondering how to make this process more efficient. In essence, I'm looking for a way to distribute logs of such tools to people interested in them and don't mind the amount of automated mails. I think that a mailing list could serve that purpose well. In addition, the mailing list archive would be an easy accessible and search-able way to archive the logs. Thoughts? Justus
