On Sun, Jan 03, 2016 at 11:36:18AM +1300, Andrew Bartlett wrote: > On Sat, 2016-01-02 at 22:22 +0000, Jelmer Vernooij wrote: > > On Sun, Jan 03, 2016 at 11:00:44AM +1300, Andrew Bartlett wrote: > > > On Sat, 2016-01-02 at 21:51 +0000, Jelmer Vernooij wrote: > > > > On Sun, Jan 03, 2016 at 10:38:12AM +1300, Andrew Bartlett wrote: > > > > > On Thu, 2015-12-31 at 18:53 +1300, Andrew Bartlett wrote: > > > > > > Just a quick note to say that the LDB build failure isn't a > > > > > > regression > > > > > > - whatever is going on here has always been going on, it just > > > > > > wasn't > > > > > > tested before ldb 1.1.23. > > > > > > > > > > > That said, fixing this is going to require gdb on a s390. Is > > > > > > the > > > > > > porterbox where that should be done? > > > > > > > > > > > > https://db.debian.org/machines.cgi?host=zelenka > > > > > > > > > > > > How do I get on such a machine? > > > > > > > > > > I've asked (but not yet got, which is reasonable over the > > > > > holidays) > > > > > access to this machine. > > > > FWIW this issue is reproducible on other platforms than s390 too. > > > > > > Can you get me details? > > Sorry, I don't really have any other than that I've managed to hit it > > a few times while building locally on my amd64 laptop. > > Can we re-submit it and see if it flaps to success this time? > > Given it doesn't show up under valgrind or under multiple runs (Fedora > x64, git master) here, this is going to be difficult.
I think it was consistently failing when I encountered it (on my laptop, not in a chroot), and both s390 buildd builds (and the ubuntu build) failed too so it is not likely to be flaky there: https://buildd.debian.org/status/logs.php?pkg=ldb&ver=2%3A1.1.24-1&arch=s390x Unfortunately it seems to pass on my laptop now, though I am not sure what has changed to make it pass there. Cheers, Jelmer
signature.asc
Description: PGP signature
