On Tue, Mar 20, 2007 at 09:32:32PM -0400, Stefan Teleman wrote: > #if defined(SOLARIS2) && ((SOLARIS2 + 0) >= 10) > #include <atomic.h> > #if defined(USE_GENERIC_ATOMICS) > #undef USE_GENERIC_ATOMICS > #endif > > If i am in Solaris >= 10 and i have the Solaris atomics, do i want to > still need the generic atomics ?
That's not how it's structured. It's either generic atomics OR a system/architecture specific implementation, there are valid reasons why people may prefer generic atomics at build time. But the reason is more our conventions; we should never over-ride a user-specifiable define. > >Your patch also seems to be a diff against a httpd source tree, and > >includes httpd-specific code. This is [EMAIL PROTECTED], please use it only > >for > >APR code, and it's best to submit patches against trunk , *not* whatever > >httpd happens to ship :-) > > yes it is against 2.2.4. And is it any better than what we have in trunk? I see you're using the _nv calls, I'm not sure if they mean "nevada" or "non-volatile" though ;-) -- Colm MacCárthaigh Public Key: [EMAIL PROTECTED]
