Howard Chu wrote:

michael@stroeder.com wrote:

...

P.S.: I had this case more than once: A test script exited with error when being run within 'make test' but then I can't reproduce the issue.

Yeah, that's the nature of things. We should probably add a "ulimit -c unlimited" to the run script so we can always get a core file.

Already added this to my build script which also invokes 'make test'.

Problems of this type that occur during "make test" but not when running the individual script seem to have a dependency on the memory pressure from other processes. (I.e., when "make test" is running there are more processes than when you simply run an individual script.)

Well, I thought running OpenOffice, Seamonkey and an additional virtual machine on my ancient 1 GB laptop is enough "memory pressure". But it seems to happen rarely.

In this case, it's most likely ITS#5733, which still hasn't been tracked down.

Anyway I was able to obtain a stack-trace. See my other posting.

Ciao, Michael.