[RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[Joel Brobecker]

Hello,

Michael had a very clever idea on how to avoid the assumption that
the block thread is the thread with the highest ID. We just get a
backtrace from all threads, and check that one of them matches the
backtrace that would be expected from the thread that would be blocked.

2004-05-18  J. Brobecker  <brobecker@gnat.com>

        * gdb.threads/pt.c: New file.
        * gdb.threads/pthread_cond_wait.exp: New test.

OK to apply?

Thanks,
-- 
Joel

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2ndversion)

[Michael Snyder]

Joel Brobecker wrote:
Hello,

Michael had a very clever idea on how to avoid the assumption that
the block thread is the thread with the highest ID. We just get a
backtrace from all threads, and check that one of them matches the
backtrace that would be expected from the thread that would be blocked.
2004-05-18  J. Brobecker  <brobecker@gnat.com>

        * gdb.threads/pt.c: New file.
        * gdb.threads/pthread_cond_wait.exp: New test.
OK to apply?
Oh, and you were gonna change the name of the source file.

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[Joel Brobecker]

> >2004-05-18  J. Brobecker  <brobecker@gnat.com>
> >
> >        * gdb.threads/pt.c: New file.
> >        * gdb.threads/pthread_cond_wait.exp: New test.
> >
> >OK to apply?
> 
> Oh, and you were gonna change the name of the source file.

Aaahhhh, yes, I forgot to update the ChangeLog but I confirm the name
has changed to pthread_cond_wait.c.

Thanks for spotting this.
-- 
Joel

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2ndversion)

[Michael Snyder]

Michael Snyder wrote:
Joel Brobecker wrote:

Hello,

Michael had a very clever idea on how to avoid the assumption that
the block thread is the thread with the highest ID. We just get a
backtrace from all threads, and check that one of them matches the
backtrace that would be expected from the thread that would be blocked.
2004-05-18  J. Brobecker  <brobecker@gnat.com>

        * gdb.threads/pt.c: New file.
        * gdb.threads/pthread_cond_wait.exp: New test.
OK to apply?


Oh, and you were gonna change the name of the source file.
Oops, sorry, that was meant to be a private reply.

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[Joel Brobecker]

[sigh, with the patch this time...]

Michael had a very clever idea on how to avoid the assumption that
the block thread is the thread with the highest ID. We just get a
backtrace from all threads, and check that one of them matches the
backtrace that would be expected from the thread that would be blocked.

2004-05-18  J. Brobecker  <brobecker@gnat.com>

        * gdb.threads/pt.c: New file.
        * gdb.threads/pthread_cond_wait.exp: New test.

OK to apply?

-- 
Joel
Attachment:
pthread_cond_wait.c
Description: Text document
Attachment:
pthread_cond_wait.exp
Description: Text document

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[Daniel Jacobowitz]

On Tue, May 18, 2004 at 02:57:30PM -0700, Joel Brobecker wrote:
> /* A small multi-threaded test case.
> 
>    Copyright 2003

Welcome to the present :)  Even though this was posted in 2003
originally, it was not released, so it should just say 2004.


-- 
Daniel Jacobowitz

Re: [RFA/testsuite/threads] backtrace past pthread_cond_wait() (2nd version)

[Joel Brobecker]

> > /* A small multi-threaded test case.
> > 
> >    Copyright 2003
> 
> Welcome to the present :)  Even though this was posted in 2003
> originally, it was not released, so it should just say 2004.

I think I'm never going to get this one right. Let me finish what I am
doing right now, take an aspirin, and resubmit everything properly.

:-)

-- 
Joel