Source: xapian-core Severity: serious Justification: (Occasionally) fails to autobuild
We've been seeing occasional FTBFS from xapian-core for some time, where tracing back the origin of the exception, it appears to be due to connect() or read() returning ECHILD, when neither is documented to. Looking at the number of failures compared to the number of attempts, it looks like it happens a few percent of the time. The failure has been seen on multiple architectures - mipsel, hppa, alpha and i386 so far. I have only ever seen this on Debian buildds. Rebuilds on porter boxes for failed cases have always worked for me, and I have never seen this when running the testsuite myself (which I do a lot, particularly on x86-64, as I also do upstream development). The rarity and inability to reproduce this where I can directly investigate makes it rather hard to debug what's going on, but we really should have a bug to track this, so I'm filing one. -- System Information: Debian Release: wheezy/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 3.2.0-2-amd64 (SMP w/8 CPU cores) Locale: LANG=en_NZ.UTF-8, LC_CTYPE=en_NZ.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash
signature.asc
Description: Digital signature
