Re: libjava test suite keeps getting stuck

[Bryce McKinlay <bryce at waitaki dot otago dot ac dot nz>]

Zack Weinberg wrote:

> On Thu, Apr 19, 2001 at 10:12:08AM -0700, Zack Weinberg wrote:
> >
> > If I run Divide_1 by hand it doesn't get stuck
>
> This is likely to be a long-standing bug in the Linux thread library.
> exit(3) from the main thread does not reliably terminate all other
> threads.  libjava may be assuming that it does.  The bug is more or
> less unfixable, and needs to be worked around.

Actually I doubt that this has anything much to do with threads and
exit(). I tested the threads code extensively when I was hacking on it
last year, including shutdown cases, and I'm pretty confident that it
works at least 99.9% of the time ;-)

I think that this probably has something to do with the recent EH
improvements. Divide_1 exercises the semantic differences between Java
ArithmeticException and ia32 FP exception, and it is a test case that has
in the past broken frequently. I'll bet something needs to be tweaked in
HANDLE_DIVIDE_OVERFLOW (in libjava/include/i386-signal.h) to work
correctly with the new EH.

regards

  [ bryce ]