From: Eric Sunshine <sunshine@sunshineco.com>
Date: Sun, 20 Mar 2016 01:23:43 GMT
Subject: Re: [PATCH] Fixing segmentation fault when merging FETCH_HEAD
Message-ID: <CAPig+cRVtzLjKTW7FZ-h8thEXkTqBJtbSwpKJs3+wdRHYV1qrQ@mail.gmail.com>
In-Reply-To: <1458411460-26146-1-git-send-email-joseivan@lavid.ufpb.br>

On Sat, Mar 19, 2016 at 2:17 PM, Jose Ivan B. Vilarouca Filho
<joseivan@lavid.ufpb.br> wrote:
> From: "Jose Ivan B. Vilarouca Filho" <joseivan@lavid.ufpb.br>

You can drop this line since it is the same as the From: line in the
email envelope.

> Fixing segmentation fault when merging FETCH_HEAD

Alternate:

    merge: don't dereference NULL pointer

> A segmentaion fault is raised when trying to merge FETCH_HEAD
> formed only by "not-for-merge" refs.
>
> Ex:
>     git init .
>     git remote add origin ...
>     git fetch origin
>     git merge FETCH_HEAD

Can you add a test to ensure that some future change doesn't regress
this fix? The above recipe would make a good basis for the new test.

> Signed-off-by: Jose Ivan B. Vilarouca Filho <joseivan@lavid.ufpb.br>
> ---
> diff --git a/builtin/merge.c b/builtin/merge.c
> @@ -1270,9 +1270,9 @@ int cmd_merge(int argc, const char **argv, const char *prefix)
>                             "an empty head"));
>                 remoteheads = collect_parents(head_commit, &head_subsumed,
>                                               argc, argv, NULL);
> -               remote_head = remoteheads->item;
> -               if (!remote_head)
> +               if ((!remoteheads) || (!remoteheads->item))

Style: drop unnecessary parantheses

    if (!remoteheads || !remoteheads->item)

>                         die(_("%s - not something we can merge"), argv[0]);
> +               remote_head = remoteheads->item;
>                 if (remoteheads->next)
>                         die(_("Can merge only exactly one commit into empty head"));
>                 read_empty(remote_head->object.oid.hash, 0);
> --
> 1.7.10.4