Control: tags -1 moreinfo Ian
On Wed, Sep 27, 2023 at 10:33:32PM +0100, Ian Jackson wrote: > Mark Hindley writes ("Bug#1052942: insserv: FTBFS: insserv: Could not read > script nolsbheader: No such file or directory"): > > Thanks for this. However, I am currently unable to repoduce this > > failure in my customary pbuilder setup. And it doesn't appear at > > reproducible builds either[1] > > I just tried this myself with my usual sbuild setup and it succeeded > there too[1]. Thanks for your confirmation. > Lucas, I think something from your rebuild environment > (a chroot of some kind I presume) must be triggering this failure. Is > there some way we can reproduce it more precisely (eg, a buildinfo > file?) Yes, I agree a buildinfo file might give a hint. > I looked at the build log > http://qa-logs.debian.net/2023/09/25/insserv_1.24.0-1_unstable.log > and compared it to the one from my sbuild, using diff. There are a > lot of changes to the "furniture" but also there are noise changes to > the output of the insserv test suite, including ordring changes of > passing tests. This seemed surprising to me. > > Mark, is the insserv test suite supposed to produce deterministic > output ? I have never had the need to look at the testsuite since I started looking after the package. A quick look now doesn't immediately reveal something that would obviously change the order of the tests. I tried again locally with make -j8 and still could not reproduce any failure. Mark