quartz window: explicit release or not?

[SDLBugzilla]

This bug report was migrated from our old Bugzilla tracker.

Reported in version: HG 1.2
Reported for operating system, platform: Mac OS X (All), PowerPC

Comments on the original bug report:

On 2006-01-27 11:06:53 +0000, Ryan C. Gordon wrote:

http://www.libsdl.org/cgi/cvsweb.cgi/SDL12/src/video/quartz/SDL_QuartzVideo.m

Bob's patch in revision 1.41 disagrees with mine in revision 1.45...figure out which is actually right.

--ryan.

On 2006-01-27 11:23:29 +0000, Ryan C. Gordon wrote:

Setting Sam as "QA Contact" on all bugs (even resolved ones) so he'll definitely be in the loop to any further discussion here about SDL.

--ryan.

On 2006-03-22 13:29:46 +0000, Max Horn wrote:

All I can say is that I just verified that
[qz_window isReleasedWhenClosed]
returns true (at the end of QZ_SetVideoWindowed, that is).

On 2006-04-27 06:08:41 +0000, Sam Lantinga wrote:

It definitely sounds like Bob's is right. Can you verify?

On 2006-05-07 17:08:43 +0000, Sam Lantinga wrote:

Can someone confirm this for 1.2.10 release?

On 2006-05-07 18:52:27 +0000, Ryan C. Gordon wrote:

I added some debug printf()s ...

printf("pre close: %d\n", (int) [ qz_window retainCount ] );
[ qz_window close ];
printf("post close: %d\n", (int) [ qz_window retainCount ] );
[ qz_window release ];
printf("post release: %d\n", (int) [ qz_window retainCount ] );

This produces:

pre close: 1
post close: 1
objc: FREED(id): message retainCount sent to freed object=0x11371d0
Trace/BPT trap

The Apple docs say release-when-closed is ignored if the window has a controller...I guess we do in this case. So the explicit release, as it is in svn right now, is correct. No patch needed.

--ryan.