Fix fetch-clone in the presense of signals

Linus Torvalds· Feb 11, 2006, 18:41 UTC · lore

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>
---

diff --git a/fetch-clone.c b/fetch-clone.c
index b67d976..d8216cb 100644
--- a/fetch-clone.c
+++ b/fetch-clone.c
@@ -153,10 +153,13 @@ int receive_keep_pack(int fd[2], const c
 		if (sz == 0)
 			break;
 		if (sz < 0) {
-			error("error reading pack (%s)", strerror(errno));
-			close(ofd);
-			unlink(tmpfile);
-			return -1;
+			if (errno != EINTR && errno != EAGAIN) {
+				error("error reading pack (%s)", strerror(errno));
+				close(ofd);
+				unlink(tmpfile);
+				return -1;
+			}
+			sz = 0;
 		}
 		pos = 0;
 		while (pos < sz) {