Christian Hofstaedtler <z...@debian.org> writes: > * Russ Allbery <r...@debian.org> [140409 02:46]:
>> Hm. Those are timing-sensitive tests that can fail on extremely slow >> systems, but I did have them tweaked so that they pass on all of Debian's >> buildds. > Actually this should be quite a fast system. Yeah, I just reproduced. I should have checked first, sorry. This was a different problem. Looks like Linux changed the amount of data that it's willing to buffer yet again, so tests that have to write enough data so that a network write will actually block are not blocking again. I'll change the thresholds in this upload. >> Yeah, I was getting ready to make that change. (There's no way for >> this package to use the metapackage.) It currently builds against 1.9 >> and 2.0; I'll change that to build against 2.0 and 2.1 and upload >> without any changes to the tests and see if the buildds are still >> happy. If so, I think the FTBFS is spurious. > Note that there's also a ruby-all-dev metapackage. Maybe you can > make use of it... Unfortunately, Ruby's mkmf system doesn't, so far as I can tell, have any way of building for all installed versions, so I need to know what versions to build for so that I can iterate through the Ruby interpreters. If there are tools available to do this for a build system that uses mkmf directly, I'd be happy to switch over. (But I don't mind doing uploads for new Ruby versions; it's not hard, and I should be able to generally do so quickly.) -- Russ Allbery (r...@debian.org) <http://www.eyrie.org/~eagle/> -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
