Fix fetch-clone in the presense of signals
commit98deeaa82fb2b96395854e0574272ed21cbd81fe
authorLinus Torvalds <torvalds@osdl.org>
Sat, 11 Feb 2006 18:41:22 +0000 (11 10:41 -0800)
committerJunio C Hamano <junkio@cox.net>
Sun, 12 Feb 2006 00:50:03 +0000 (11 16:50 -0800)
tree00461313b09ef04e07561521f5fb307187e7f053
parentc548cf4ee0737a321ffe94f6a97c65baf87281be
Fix fetch-clone in the presense of signals

We shouldn't fail a fetch just because a signal might have interrupted
the read.

Normally, we don't install any signal handlers, so EINTR really shouldn't
happen. That said, really old versions of Linux will interrupt an
interruptible system call even for signals that turn out to be ignored
(SIGWINCH is the classic example - resizing your xterm would cause it).
The same might well be true elsewhere too.

Also, since receive_keep_pack() doesn't control the caller, it can't know
that no signal handlers exist.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
fetch-clone.c