[PATCH] ev: fix epoll_init fd leak

[Christian Parpart]

.... my first post in this thread is about backing the top-poster's patch
for why this patch should be applied, because there might be software that
has 0/1/2 closed at the time it wants to initialize libev AND checking a
return code for <= 0 for errorness instead of < 0 is in my eyes still a
bug. that's it. period.

Secondly, do not conclude things I did not explictely write (closing fds
does not increase performance, I never said saving 3 more fds increases
performance).

Also, you represent something, like you are the maintainer of a very
important piece of software. Do not treat people like shit (indirectly),
like you did the very first time I met you by just reading my host mask,
like you did in almost every other past I read about you, and like you did
the very now. Do not imply anything just by reading email addresses, they
might mean nothing regarding the current topic.
Please stay calm and try being as polite as possible instead of blaming
everywhere around, selfishly, and giving everyone else the feeling that
you're "the" perfect guy in the world. People already forked your project
in the back-rooms for that particular reason, which is not nice, but maybe
the best to get not offendet anymore.

Regards,
Christian.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schmorp.de/pipermail/libev/attachments/20111103/b4dc35e1/attachment.html>