< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index Next in Thread >

Re: [reSIProcate] failures in testParseBuffer


On 2/27/14 8:20 AM, Daniel Pocock wrote:
On 27/02/14 17:03, Byron Campen wrote:
On 2/27/14 5:07 AM, Daniel Pocock wrote:
Some people have noticed failures in testParseBuffer, e.g.

https://launchpadlibrarian.net/167715579/buildlog_ubuntu-trusty-amd64.resiprocate_1.9.1-1_FAILEDTOBUILD.txt.gz


That is with gcc 4.8 (most of my own testing is with gcc 4.7 and travis
uses 4.6.3 and the problem hasn't appeared in either of those
environments)

Has anybody else seen issues relating to this recently or in the past?



_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel
    I've never seen this kind of behavior. Seems very troubling to me.
Is it reproducible?
I haven't tried to reproduce it myself yet, but the same failure has
been observed on a Debian unstable 64 bit build:

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=740232


but the build farm doesn't show any sign of this problem in any other
platform running Debian unstable/gcc 4.8
https://buildd.debian.org/status/package.php?p=resiprocate

(the Hurd failure in that report is the testLogger issue I mentioned in
an earlier email)


It would be nice if we could determine what the result actually was. My suspicion is that the moment we add any logging in the case of failure, the bug will go away. We need to try to catch this in a debugger and find out what's really going on (both what's in memory, and the actual instructions the compiler generated).

Best regards,
Byron Campen