Control: severity -1 serious Hi Jonas
Happy Holidays! On Sun, 27 Dec 2020 at 19:12, Jonas Smedegaard <jo...@jones.dk> wrote: > I tightened the test checking - enabled them during build and made them > more verbose (and explicitly passed environment variable "srcdata" which > might have previously failed). > > Those changed was finalized and released today as 5.8.0-2, and I hope it > can shed some light on what more detailed fails when it does. Thanks for taking care of this package. > To my surprise, the armhf failures are evidently no longer persistent > even _before_ this latest release. I therefore relax severity of this > bug. That was surprising. :-) On armhf [1], I see 142 failures for version 5.8.0-1 and one pass on 2020-12-27 02:04:17 UTC. I also see failures of 5.8.0-2 at least on amd64 [2], arm64 [3] and i386 [4]. As this is unsuitable for regression testing, I'm raising the severity of this bug. It's not clear to me whether it's just the test that is broken, in which case skipping the multi-thread test is an option, or whether the multi-threaded spell check feature is really broken in 5.8.0, in which case reverting to 5.7.0 would be better. Regards Graham [1] https://ci.debian.net/packages/l/link-grammar/testing/armhf/ [2] https://ci.debian.net/packages/l/link-grammar/unstable/amd64/ [3] https://ci.debian.net/packages/l/link-grammar/unstable/arm64/ [4] https://ci.debian.net/packages/l/link-grammar/testing/i386/