Re: [g-a-devel]regression test fails

[Michael Meeks <michael ximian com>]

Hi Benny,

On Mon, 2002-01-28 at 18:32, Bill Haneman wrote:
> > i'm getting the same blocking that was happening before the rebuild.

	Grief - as I remember last time you posted a trace the blocking
happened before the key code, and it dumped core :-) so this is a
different problem.

> Hmm, it looks as though the regression test is blocking while waiting
> for  a key event to be synthesized (and then intercepted).

	I've seen this locally; it's possibly a nasty race somewhere - I can
replicate it if I run test-simple and then do a lot of WM work - such as
switching desktop, etc. etc. but I've no idea what is going on, it's not
clear why it shouldn't work always - and the XGrabKey has to complete
before the method returns; that is - wait, possibly we need to do a
gdk_flush after our XGrabKey to ensure that the grab has got to the
server before we return and then do our test key poking ...

	What do you think Bill ? sounds reasonable.

	Regards,

		Michael.


-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot