On 28 May 2016 at 02:53, Keith Packard <[email protected]> wrote: > Emil Velikov <[email protected]> writes: > >> Hi Keith, >> >> Style question: do you/how many others refer having separate functions for >> of ifdeffed code each vs a single with all the ifdeffs. Or in other words >> having func_foo_poll and func_foo_epoll as opposed to having it all in >> func_foo. > > I started having separate versions of each function (with the same > name), but that seemed bad, so I merged the contents of the separate > functions into the single functions you see now. I found it easier to > compare the semantics of the different implementations when they were > right next to each other. > Haven't thought about that. Thanks.
>> +#ifndef _OSPOLL_H_ >>> +#define _OSPOLL_H_ >>> + >>> +#include <poll.h> >>> + >> >> This include should be in the C file with proper ifdef guard, right? > > I'm not sure I understand -- we need <poll.h> on every system as ospoll > uses the POLLIN/POLLOUT defines for its arguments. We could use X server > specific values, but I think the posix ones are common across all of our > supported systems. > There's two things which I should have mentioned: - How does one have poll.h when poll() is missing - Cygwin may have the header, although I'm not sure about other Windows based implementations (the mingw and msvc ones). - Considering that we have our own header with wrapper functions and an opaque struct, it would make sense to have X server specific macros. Even (ideally) if they are ABI compatible with the POSIX ones. -Emil _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
