I've just made a fairly major rewrite of NSUserDefaults code in gnustep-base (at svn revision 32454), and this needs testing, particularly be people using different languages.
The purpose of the rewrite was to make the code more maintainable and less fragile, in particular to try to be sure it's thread-safe but avoids deadlocking issues. I had found a small thread safety issue which had been introduced a while back by a patch intended to prevent deadlocks, but with the code in the state it was in I could not see how to fix the thread safety problem without re-introducing the possibility of deadlock ... hence the need to restructure/simplify. The old code was horribly complicated because of the way it tried to use different methods for determining the appropriate language to use if the NSLanguages user default was not set up... this resulted in a web of mutually recursive method calls, where any modification to the existing code was likely to result in infinite recursions. The rewrite simplifies all that, but my worry is that in doing so I may have broken something for people using machines with different locales ... So I'd like to encourage people to try out the latest code and report any broken language related behavior please. _______________________________________________ Discuss-gnustep mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnustep
