[AD] Fwd: OSX - lockup

[Peter Hull <peter.hull90@xxxxxxxxxx>]

Hello! Sorry, I sent this just to myself instead of to the list (what's 
the smiley for "stupid"?) Ignore the diff that relates to the docs, as 
I think that's been clarified already.

Cheers

Pete


Begin forwarded message:

> From: Peter Hull <peter.hull90@xxxxxxxxxx>
> Date: Thu Sep 8, 2005  9:57:42  pm Europe/London
> To: Peter Hull <peter.hull90@xxxxxxxxxx>
> Subject: Re: OSX - lockup
>
> On Thursday, September 1, 2005, at 06:59  am, Peter Hull wrote:
>
> > More of a note to myself, really, as I'm away this weekend. Solak has 
> > found a lock-up when using vsync() whilst the screen is acquired
> > http://www.allegro.cc/forums/view_thread.php?_id=524355
> >
> > acquire_bitmap(screen);
> >  vsync(); /* problem here */
> >  rect(10, 10, SCREEN_W - 10, SCREEN_H - 10, makecol(255, 255, 255));
> >  release_bitmap(screen);
> >  vsync();
> >
> >
> > I'll send a fix in soon; I think the most sensible course is that 
> > vsync should be a no-op if the screen is acquired. Any comments?
> OK, here is the proposed fix. If you call vsync() when the screen is 
> acquired, it will do nothing in release builds and assert in debug 
> builds. I think it's still possible to get a race condition that will 
> sometimes lock things up iff you're calling vsync from a different 
> thread to the drawing thread, but surely no-one would be daft enough 
> to do that, would they?
>
> Also, for discussion, is a change to the docs to remind people not to 
> do too much stuff while the screen is locked. The paragraph is still 
> rather DOS/Windows-centric though, should it be made more general? 
> (i.e. the bit about locking DirectDraw surfaces)
>
> Pete

Attachment: vsync.osx.diff
Description: Binary data

>

Attachment: vsync.docs.diff
Description: Binary data

>