В Fri, 16 Mar 2018 14:57:53 +0100, Riccardo Mottola написа: > The crashes happen either when the pool is released in line 224 and > NSData or in the setter of one of the properties, as the first stack > or with a crash at line 181, which is still inside this method.
You haven't posted a backtrace. I assume you cannot reproduce? > I Don't see at a glance a gross error in CWImapCacheManager - Hmm, this is perhaps the buggiest part of the library, or at least the most fragile. I remember ~15 years ago Ludo was advising users to just delete the cache if something goes wrong. A shot in the dark, but maybe there are some type mismatches or broken pointer arithmetic (or both) here or in the io.c function definitions. I would suggest uncommenting these NSLog statements (or converting them permanently to the appropriate NSDebug*Log) to see what's going on. Don't forget to adjust the specifiers for the expected types. If the crashes occur on a 64-bit machine, it would probably help to verify if they are reproducible on 32-bit. _______________________________________________ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
