[Xcb] [RFC] [PATCH 0/4] pthread cancellation

[Josh Triplett]

On Thu, Jan 07, 2010 at 10:17:33PM +0200, Rémi Denis-Courmont wrote:
> Le jeudi 7 janvier 2010 21:48:53 Jamey Sharp, vous avez écrit :
> > Finally, another point Ian raised: How can we tell whether you've made
> > every call safe? Is there any good way to either test this or be
> > convinced by inspection? Massive bonus points if you add tests to
> > XCB's (rather pathetic) test suite.
> 
> AFAICT, resource leak is the most likely consequence of bug in that regard. 
> Testing those is not easy, unless you can instrument the functions doing 
> resource allocation and releasing. Namely one'd need its own malloc/free, 
> getaddrinfo/freeaddrinfo, socket/close functions to do accounting :/

Sounds like something we could potentially test with valgrind.

- Josh