Re: git 1.5.1-rc1 doesn't like empty files

From: Junio C Hamano <junkio@cox.net>
Date: Mar 20, 2007, 06:00 UTC
Message-ID: <7v648wbgiy.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <Pine.LNX.4.64.0703192245490.6730@woody.linux-foundation.org>

Linus Torvalds <torvalds@linux-foundation.org> writes:

> We've had this bug before - thinking that we don't need to inflate() 
> anything because we already had it all..
>
> 		Linus

Thanks. I think we _do_ need a big fat warning near the code to avoid the same mistake in the future. Something like this?

diff --git a/sha1_file.c b/sha1_file.c
index 9fe2bd6..d273aff 100644
--- a/sha1_file.c
+++ b/sha1_file.c
@@ -1030,7 +1030,17 @@ static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size
 		n = size;
 	memcpy(buf, (char *) buffer + bytes, n);
 	bytes = n;
-	if (bytes < size) {
+	if (bytes <= size) {
+		/*
+		 * The above condition must be (bytes <= size), not
+		 * (bytes < size).  In other words, even if we expect
+		 * no more output, the input zlib stream may have bytes
+		 * that express "this concludes the stream", and we do
+		 * want to eat that input.  Otherwise we would not be
+		 * able to test that we consumed all the input to reach
+		 * the expected size *and* zlib gave status == Z_STREAM_END
+		 * to signal all went well.
+		 */
 		stream->next_out = buf + bytes;
 		stream->avail_out = size - bytes;
 		while (status == Z_OK)