On Fri, 2012-06-22 at 13:09 +0200, Marc Lehmann wrote: > The kernel should return einval if it doesn't understand that flag > (but this is about libc, not kernel version, right?). Even if the > kernel silently ignores the flag, libev will just set it again as with > epoll_create.
It does, but in the new kernels (which is what i use) that call is essentially being phased out. i.e. if i happen to compile with EPOLL_CLOEXEC I get epoll_create1(EPOLL_CLOEXEC); otherwise i get epoll_create(255) which translates to epoll_create1(0) when glibc call makes it to the kernel. (which of course can be mitigated with fork/close/exec with O_CLOEXEC) > > Well, you can break your stuff as you want, just make sure you don't blame > libev later when it breaks :) I use libev because it does the job well and I observe you do an excellent job worrying about all the intricacies and crazy changes in the libcs and the kernels;-> If i start maintaining personal OCD forks I am defeating the purpose. So it seems to me my best way forward is to just say "use glibc 2.9 and above". cheers, jamal _______________________________________________ libev mailing list [email protected] http://lists.schmorp.de/cgi-bin/mailman/listinfo/libev
