Jeff Bailey wrote:
On Thu, Apr 04, 2002 at 12:52:13PM -0500, B. Douglas Hilton wrote:
I patched GNU Pth a.k.a. "Portable Threads" for Hurd a while back.
It has pretty dismal performance but you can fool around with it if
you want. Links to the source and patch are here:
Is gdb smart enough to know about libpth? (Like does `info threads'
yield useful results?)
Not sure. I can try it this weekend. I have since had to re-install
Hurd, and haven't done much with it other than testing Oskit-Mach
kernels since.
Since Pth provides a "pthread.h" and libpthread I don't see why it
wouldn't work. Like I mentioned though, its performance is not quite
what you would expect from a "real" pthread implementation, and I
really haven't tested it for correctness either. Also, it is
non-preemptive, so pthread apps can deadlock if they are not
written to provide for this situation.
The patch wasn't really a big deal, I hacked the aclocal.m4 to
hardcode in defaults which were used with the Linux kernel 2.0x
which also lacked sigstack and sigaltstack. I *think* it is
working correctly, but it is a real kludge if there ever was
one :-)
Gotta go. l8r...
- Doug
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
