Thread (2 messages) flat view 2 messages, 2 authors, 2016-06-15

Re: [PATCH decompress BUG] Fix decompress_next_from() wrong argument value

From: Junio C Hamano <hidden>
Date: 2016-06-15 22:44:04

"Marco Costalba" [off-list ref] writes:
quoted hunk
The only two points in the current code in master that I would like to
report to you are a _possible_ missing inflateEnd() before a new
inflateInit(), but I am not confident with that part of code to judge
if is a bug or not, anyway that's the _possible_ diff.
diff --git a/http-push.c b/http-push.c
index 55d0c94..e0a4cc6 100644
--- a/http-push.c
+++ b/http-push.c
@@ -307,6 +307,7 @@ static void start_fetch_loose(struct
transfer_request *request)
 	/* Reset inflate/SHA1 if there was an error reading the previous temp
 	   file; also rewind to the beginning of the local file. */
 	if (prev_read == -1) {
+		inflateEnd(&request->stream);
 		memset(&request->stream, 0, sizeof(request->stream));
 		inflateInit(&request->stream);
 		SHA1_Init(&request->c);
I think this could leak if request->stream already had
something in it, but I do not see anything that is done to it
after it is initialized and the code reaches to this point.

In fact, I suspect that it would make more sense to remove the
earlier memset() that clears request->stream and inflateInit(),
and moving the memset() and inflateInit() we see above out of
that if clause (before checking prev_read).

The same comment applies to the other hunk.

By the way, I was looking at the earlier two series from you
(compress and decompress), and noticed some of them were corrupt
with linewrap.  As I think they are good clean-up patches, I'd
like to apply them as one of the first series post 1.5.4.  As
such, this request is not urgent at all, but please resend with
a clean-up when 'master'/'next' reopens.

Thanks.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help