On Sun, Jul 01, 2007 at 03:35:36AM -0500, Przemys?aw Pawe?czyk wrote:
> My
> post "was marked" (in reply) as "it's an application bug", and went
> into....oblivion.

An application calling pthread_cond_wait() with an unlocked mutex is wrong.
Why this happens needs to be examined inside that application - it may
be a secondary effect of a bug in NetBSD, but by a very high percentage
more likely (this is a polite understatement) is a flaw in the application
logic.

If you are unable to examine this problem yourself further, you should report
it to the RapidSVN developers. The diagnostic is precise, it should be easy
for them to analyze the problem. It might be a problen in wxGTK instead of
RapidSVN itself, but to know that we'd at least have to see a stack trace
from the fault - and the RapidSVN people sure can forward the problem if they
find out it is.

Martin
(happy svn command line user)