On Thu, Oct 9, 2008 at 5:29 PM, Ruediger Pluem <[EMAIL PROTECTED]> wrote:


> Does anybody see problems with this or are we still too worried about
> the correct handling of signed vs. unsigned variables by apr_atomic_XXXX
> to use this in a non experimental MPM?


signed vs unsigned doesn't bother me.  However, we should consider the lack
of APR native atomic support on various platforms by default.  I think we
still have to say --enable-nonportable-atomics to get a native compare &
swap type operation on x86 + gcc, maybe also on SPARC.  But is it rational
to worry about supporting 386s with binary distributions in 2008?  dunno
about the SPARC market.  This wasn't a concern for Event because it is
experimental.

I will take a look at the APR atomics and see if the operations that Event's
fdqueue is using are less supported than the atomics used in worker.

Greg

Reply via email to