Re: Interesting difference in memory management of MacOS Foundation and GNUstep

[Mathias Bauer]

Hi,

in case somebody else is also interested in this: it seems that Apple's 
runtime "protects" the developer by ignoring changes to the retain count 
as soon as the object entered its deallocate method. Wrong decision, IMHO.

Regards,
Mathias

Am 14.02.14 10:53, schrieb Mathias Bauer:
> Hi dear list members,
>
> I have ported most parts of the CocoaHTTPServer from Robbie Hanson
> including GCDAsyncSocket to Linux and found an interesting difference in
> memory management.
>
> Here's the dealloc function of GCDAsyncSocket:
>
> > - (void)dealloc
> > {
> >     LogInfo(@"%@ - %@ (start)", THIS_METHOD, self);
> >
> >     if (dispatch_get_specific(IsOnSocketQueueOrTargetQueueKey))
> >     {
> >         [self closeWithError:nil];
> >     }
> >     else
> >     {
> >         dispatch_sync(socketQueue, ^{
> >             [self closeWithError:nil];
> >         });
> >     }
>
> This method crashes on Linux when the dealloc function in not called in
> the socket or target queue as the block that calls closeWithError is
> executed in this case. Catching "self" in it increments and after block
> execution decrements the retain count of "self". Doing this in a dealloc
> method of course leads into disaster.
>
> But the same code does not crash on MacOS or iOS, even if the block
> containing closwWithError is executed.
>
> I don't understand that, but maybe somebody knowing the differences
> better than I can explain that.
>
> Best regards,
> Mathias
>
> _______________________________________________
> Discuss-gnustep mailing list
> Discuss-gnustep@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep