On Tuesday, March 10 2020, Boyuan Yang wrote: > 在 2020-03-09一的 17:12 -0400，Sergio Durigan Junior写道： >> Control: severity -1 serious >> Control: tags -1 + patch >> >> On Thursday, March 05 2020, I wrote: >> >> > On Wednesday, January 29 2020, Boyuan Yang wrote: >> > >> > > Dear source-highlight maintainer, >> > > >> > > Currently source-highlight would FTBFS due to errors when running "make >> > > test" >> > > after build: >> > > >> > > https://buildd.debian.org/status/fetch.php?pkg=source-highlight&arch=arm64&ver=3.1.9-1&stamp=1579812231&raw=0 >> > > >> > > There are differences between the expected output value and the real >> > > value. >> > > Please investigate into it. >> > >> > I took some time to investigate this on amdahl.d.o, and I am not able to >> > reproduce this failure. I left "make check" running in a loop for >> > several minutes, and all of the runs successfully passed without a >> > problem. >> > >> > I'm almost sure this was a failure caused by either a non-deterministic >> > test, or by some broken dependency that got fixed later. I'm therefore >> > taking the liberty to downgrade the severity of this bug to normal, as >> > well as retitling it to remove the FTBFS part. Feel free to revert >> > these changes if you disagree. >> >> I stand corrected. The bug has not been fixed, and it's now impacting >> the latest build even on amd64: >> >> >> https://buildd.debian.org/status/fetch.php?pkg=source-highlight&arch=amd64&ver=3.1.9-1.1&stamp=1583753021&raw=0 >> >> I *think* I understand why the problem is manifesting. The issue here >> is that 'make check' is parallelized, which may cause a race condition >> due to buffering, redirections and file-writing commands. It's a mess. >> >> I can trigger a testsuite error when I locally run 'make check -j6', but >> the error doesn't happen on 'make check -j1'. For this reason, I'm >> proposing the patch below, which disables parallelization on >> dh_auto_test. >> >> I am not 100% sure of this analysis, though. But since I can't spend >> much more time on this, I think it's worth a try. > > This looks okay to me. Will go for the NMU and give it a try. > > However, I do not plan to close this bug; the failure of parallel build is > still a bug. Downgrading it to severity:important might be a better choice.
I don't have a problem with this, but I think the title should be updated to reflect that the failure happens when testing in parallel, then. Thanks, -- Sergio GPG key ID: 237A 54B1 0287 28BF 00EF 31F4 D0EB 7628 65FC 5E36 Please send encrypted e-mail if possible http://sergiodj.net/
signature.asc
Description: PGP signature
