I agree that that would be a bug, but I can't reproduce it.
Running on my MacBook Pro with 10.5.2 and a fresh build of Nu from
sources:
[Xenon:~] tim% nush
decarbonization.nu
2008-03-18 23:35:32.440 nush[16414:807] doSomething
2008-03-18 23:35:33.440 nush[16414:807] doSomething
2008-03-18 23:35:34.440 nush[16414:807] doSomething
2008-03-18 23:35:35.440 nush[16414:807] doSomething
2008-03-18 23:35:36.440 nush[16414:807] doSomething
2008-03-18 23:35:37.440 nush[16414:807] doSomething
^C
[Xenon:~] tim% cat
decarbonization.nu
(class TestObject is NSObject
(- (void)doSomething is
(NSLog "doSomething")))
(set object (TestObject new))
(NSTimer scheduledTimerWithTimeInterval:1.0 target:object
selector:"doSomething" userInfo:nil repeats:YES)
((NSApplication sharedApplication) run)
[Xenon:~] tim%
It also works for me with a Nu installed from the 0.3.0 dmg.
Would you confirm that you get this?
[Xenon:~] tim% nush
Nu Shell.
% YES
1
I recently moved the definition of YES from nu/
cocoa.nu to the
compiled-in symbol table initialization near the end of objc/
operators.m.
Tim