@synchronized and @try/@catch cause stack corruption (self set to nil for callers)

[GoogleCodeExporter]

Compile and run the attached code files.

Expected output:
2007-08-19 11:25:43.509 test[4164] a. <TestObject: 0x302ef0> firstCall
2007-08-19 11:25:43.509 test[4164] a. <TestObject: 0x302ef0> secondCall
2007-08-19 11:25:43.509 test[4164] b. <TestObject: 0x302ef0> secondCall
2007-08-19 11:25:43.509 test[4164] c. <TestObject: 0x302ef0> secondCall
2007-08-19 11:25:43.509 test[4164] b. <TestObject: 0x302ef0> firstCall

Actual output:
2007-08-19 11:25:22.513 test[8202:d03] a. <TestObject: 0x100ed0> firstCall
2007-08-19 11:25:22.523 test[8202:d03] a. <TestObject: 0x100ed0> secondCall
2007-08-19 11:25:22.526 test[8202:d03] b. (null) (null)
2007-08-19 11:25:22.528 test[8202:d03] c. (null) (null)
2007-08-19 11:25:22.530 test[8202:d03] b. <NSThread: 0x1019e0>{name = (null), 
num = 1} 
(null)
zsh: bus error  /test

Original issue reported on code.google.com by kiji.ro...@gmail.com on 12 Sep 2007 at 3:08

Attachments:

• synchronized.m
• try-catch.m

[GoogleCodeExporter]

Possibly related to issue 10. Will fix that one and then verify.

Original comment by nightwat...@gmail.com on 12 Sep 2007 at 3:15

• Changed state: Waiting

[GoogleCodeExporter]

Original comment by nightwat...@gmail.com on 12 Sep 2007 at 3:16

• Changed state: Depends

[GoogleCodeExporter]

Reproduced.

Original comment by nightwat...@gmail.com on 12 Sep 2007 at 6:46

• Changed state: Started

[GoogleCodeExporter]

Fixed. Had forgot to merge the JBLEN fixes. This should probably be reported to 
LLVM too...

Original comment by nightwat...@gmail.com on 12 Sep 2007 at 6:55

• Changed state: Fixed