On Tuesday 19 May 2009 14:56:25 James Harper wrote: > > > The calls to start_server are already protected by P() and V() so > > > everything in start_server is fine. > > > > The problem is in the test (line 357), which is not protected by a > > mutex. > > > Therefore your fix does not definitively resolve the problem. > > Line 357 is in "start_server". start_server is called in the following > places: > . line 284, where it is protected by the mutex acquire on line 249 > . line 329, where it is protected by the mutex acquire on line 311 > . line 433, where it is protected by the mutex acquire on line 390
Hmm. Very good point that I missed. > > Either I'm missing something obvious, or I just don't understand what's > going on (in which case I'm probably not qualified to write the > 'correct' patch you requested :) I think you are well qualified to write the patch. I had missed the fact that the mutex was set before calling start_server, which means that the test I was referring too *is* indeed covered by the lock, and thus your patch is good. > > More likely I'm just missing something obvious, in which case please > enlighten me (it's nearly 11pm here and I probably should be in bed). Sorry, but you are already enlightened, please go ahead and apply your patch either tonight or tomorrow :-) Kern ------------------------------------------------------------------------------ Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects _______________________________________________ Bacula-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/bacula-devel
