Re: [PATCH] grep: handle corrupt index files early

From: Junio C Hamano <gitster@pobox.com>
Date: May 15, 2018, 23:58 UTC
Message-ID: <xmqq603o8o9w.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <20180515010425.149200-1-sbeller@google.com>

Stefan Beller <sbeller@google.com> writes:

> Any other caller of 'repo_read_index' dies upon a negative return of
> it, so grep should, too.

Makes sense. When the function returns a failure, there is no sensible fallback action anyway, so dying here is alright.

Will queue; thanks.

Show 24 quoted lines

>
> Signed-off-by: Stefan Beller <sbeller@google.com>
> ---
>
> Found while reviewing the series
> https://public-inbox.org/git/20180514105823.8378-1-ao2@ao2.it/
>
>  builtin/grep.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/builtin/grep.c b/builtin/grep.c
> index 6e7bc76785a..69f0743619f 100644
> --- a/builtin/grep.c
> +++ b/builtin/grep.c
> @@ -488,7 +488,8 @@ static int grep_cache(struct grep_opt *opt, struct repository *repo,
>  		strbuf_addstr(&name, repo->submodule_prefix);
>  	}
>  
> -	repo_read_index(repo);
> +	if (repo_read_index(repo) < 0)
> +		die("index file corrupt");
>  
>  	for (nr = 0; nr < repo->index->cache_nr; nr++) {
>  		const struct cache_entry *ce = repo->index->cache[nr];