send-pack: report signal death of pack-objects
authorJeff King <peff@peff.net>
Tue, 7 Mar 2017 13:39:48 +0000 (08:39 -0500)
committerJunio C Hamano <gitster@pobox.com>
Tue, 7 Mar 2017 22:58:36 +0000 (14:58 -0800)
If our pack-objects sub-process dies of a signal, then it
likely didn't have a chance to write anything useful to
stderr. The user may be left scratching their head why the
push failed. Let's detect this situation and write something
to stderr.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
send-pack.c
index e152327394252c53cbaa16102b807e91b8fa7dcd..d2d2a49a0231293e98ab8546945f8d3ef9fa205d 100644 (file)
@@ -72,6 +72,7 @@ static int pack_objects(int fd, struct ref *refs, struct sha1_array *extra, stru
        struct child_process po = CHILD_PROCESS_INIT;
        FILE *po_in;
        int i;
+       int rc;
 
        i = 4;
        if (args->use_thin_pack)
@@ -125,8 +126,20 @@ static int pack_objects(int fd, struct ref *refs, struct sha1_array *extra, stru
                po.out = -1;
        }
 
-       if (finish_command(&po))
+       rc = finish_command(&po);
+       if (rc) {
+               /*
+                * For a normal non-zero exit, we assume pack-objects wrote
+                * something useful to stderr. For death by signal, though,
+                * we should mention it to the user. The exception is SIGPIPE
+                * (141), because that's a normal occurence if the remote end
+                * hangs up (and we'll report that by trying to read the unpack
+                * status).
+                */
+               if (rc > 128 && rc != 141)
+                       error("pack-objects died of signal %d", rc - 128);
                return -1;
+       }
        return 0;
 }