Boost logo

Boost Testing :

From: Nicola Musatti (Nicola.Musatti_at_[hidden])
Date: 2007-07-02 17:39:34

John Maddock wrote:
> Nicola Musatti wrote:
>> David Abrahams wrote:
>>> Hi John and Roland,,
>>> w.r.t., what do we know exactly?
>>> Are we sure test wasn't killed by the regression system's timeout
>>> mechanism. Given the -l3600 argument, the test would have to have
>>> run for over an hour, right? Does it do anything like that?
>> FWIW, regex_regress_threaded has been failing intermittently on my PC
>> for ages. This is on Windows with different compilers.
> Oh, blast :-( What are the symptoms?

This is one instance with Borland's bcc32 5.8.2 compiler (BCB2006):

Assertion failed: *(s_data.cont.back().second) == k, file
..\boost/regex/pending/object_cache.hpp, line 106

Abnormal program termination

Assertion failed: attached_thread_count == 0, file
..\libs\thread\src\tss_hooks.cpp, line 157

Abnormal program termination


The same error appears with other releases of the Borland compiler, but
it's not deterministic: some runs fail, some succeed. I thought I had
seen it also with some version of MSVC, but I couldn't find an instance
in any of the numerous test results I have laying around, so I wouldn't
swear on it.

Nicola Musatti

Boost-testing list run by mbergal at