Re: Selector 'release' sent to dealloced instance of class NSDeviceRGBColor ??
Re: Selector 'release' sent to dealloced instance of class NSDeviceRGBColor ??
- Subject: Re: Selector 'release' sent to dealloced instance of class NSDeviceRGBColor ??
- From: Andy Lee <email@hidden>
- Date: Sun, 2 Oct 2005 11:43:48 -0400
On Oct 2, 2005, at 6:11 AM, Jonathan del Strother wrote:
I thought that NSZombie prevented memory deallocation, so there
should be no way for targetRGB's memory to later get occupied by
something else.
You're right, my hypothesis was flawed. Yet you were showing the
symptoms of exactly that happening, so maybe I should have wondered
what could violate NSZombie's preservation of memory. Then I could
have asked if you happened to be calling any low-level memory
operations like free(). I might have been hailed as a genius! Too
bad it's only my hindsight that's 20/20. :)
I think I've found the problem. This helpful page - http://
www.cocoadev.com/index.pl?DebuggingAutorelease - pointed out the
malloc_history tool, which I used to eventually track down a free()
operation on an NSString. Ooops. The NSString used to be a
malloc'd wchar, I must've missed that line when I changed
everything over.
Anyway - mental note : using free() on an NSObject confuses the
hell out of the memory debug tools :)
Thanks for the suggestions, regardless.
I'm glad you solved the problem. I finally fixed a "logically
impossible" bug last night, so I can imagine your relief.
--Andy
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Cocoa-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden