arne anka a écrit , Le 04/12/2009 23:29: >>>> # This one is OK >>>> log_level = INFO >>>> >>>> # This one isn't >>>> log_level = WARNING >>> >>> sounds certainly like one of those beloved race conditions. >> >> Please explain. > > well, race condition means two (or more) operations racing each other > for a bit of code -- ie non-determinitsic outcome. > so, using INFO means, far more operation cause output and every single > of these outputs caused a small delay compared to WARNING. > ie, it could mean an operation gets more time than intended and wins the > race over the one intended to be first. > > while writing it occurs to me it could simply be an issue of too short > timeouts because of slow hardware (or too fast sowftware, for that matter). > with INFO the hardware gets more time to do whatever necessary, > increasing the chance to get ready just in time.
It makes since since zhone fails with log_level settings WARNING, ERROR, and CRITICAL, while it succeeds with DEBUG and INFO. > in the latter case an increased timeout will probably help a lot. since > it didn't happen with older fso, the fso guys could probably shed some > light on that, telling if somewhere a timespan was decreased. Thanks, _Gilles.
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Smartphones-userland mailing list Smartphones-userland@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-userland