Did the fix for recycled thread ids uncover another bug?

Did the fix for recycled thread ids uncover another bug?

[Paul Gilliam]

Forgive me if this is not the place to post this.  It's kind of a 
follow-up to
a thread that was here a few months ago.

If I set a breakpoint in a thread function, then when that breakpoint is 
hit,
things go bad.  If instead of a breakpoint, I hit cntl-c while the thread
function is active, things work right.

I have seen this on Intel and ppc.  (The Intel was recent, but not 
current.
The ppc was current).

The strange thing is that the bug shows up on 32-bit ppc, but not 64-bit 
ppc.

I will post a follow-up with the details if someone tells me this is the 
right
mail-list.


-=# Paul Gilliam #=-

Re: Did the fix for recycled thread ids uncover another bug?

[Daniel Jacobowitz]

On Mon, Jun 14, 2004 at 03:57:57PM -0700, Paul Gilliam wrote:
> Forgive me if this is not the place to post this.  It's kind of a 
> follow-up to
> a thread that was here a few months ago.
> 
> If I set a breakpoint in a thread function, then when that breakpoint is 
> hit,
> things go bad.  If instead of a breakpoint, I hit cntl-c while the thread
> function is active, things work right.
> 
> I have seen this on Intel and ppc.  (The Intel was recent, but not 
> current.
> The ppc was current).
> 
> The strange thing is that the bug shows up on 32-bit ppc, but not 64-bit 
> ppc.
> 
> I will post a follow-up with the details if someone tells me this is the 
> right
> mail-list.

You should probably use gdb@ instead.

-- 
Daniel Jacobowitz

Re: Did the fix for recycled thread ids uncover another bug?

[Michael Snyder]

Daniel Jacobowitz wrote:
> On Mon, Jun 14, 2004 at 03:57:57PM -0700, Paul Gilliam wrote:
> 
>>Forgive me if this is not the place to post this.  It's kind of a 
>>follow-up to
>>a thread that was here a few months ago.
>>
>>If I set a breakpoint in a thread function, then when that breakpoint is 
>>hit,
>>things go bad.  If instead of a breakpoint, I hit cntl-c while the thread
>>function is active, things work right.
>>
>>I have seen this on Intel and ppc.  (The Intel was recent, but not 
>>current.
>>The ppc was current).
>>
>>The strange thing is that the bug shows up on 32-bit ppc, but not 64-bit 
>>ppc.
>>
>>I will post a follow-up with the details if someone tells me this is the 
>>right
>>mail-list.
> 
> 
> You should probably use gdb@ instead.
> 

If you do, would you please include me in the Cc: list?
I'm sort of a semi-active threads maintainer.  I'm curious
about the details of how "things go bad".

Michael