Re: [PATCH 1/3] write_entry: fix leak when retrying delayed filter

From: Jeff King <peff@peff.net>
Date: Oct 10, 2017, 09:25 UTC
Message-ID: <20171010092543.e4dh2lo4sj3w6w7j@sigill.intra.peff.net>
In-Reply-To: <20171010092319.ugim7ww7t6ks2vqf@ruderich.org>

On Tue, Oct 10, 2017 at 11:23:19AM +0200, Simon Ruderich wrote:

Show 13 quoted lines

> On Tue, Oct 10, 2017 at 09:00:09AM +0900, Junio C Hamano wrote:
> >> --- a/entry.c
> >> +++ b/entry.c
> >> @@ -283,6 +283,7 @@ static int write_entry(struct cache_entry *ce,
> >>  			if (dco && dco->state != CE_NO_DELAY) {
> >>  				/* Do not send the blob in case of a retry. */
> >>  				if (dco->state == CE_RETRY) {
> >> +					free(new);
> >>  					new = NULL;
> >>  					size = 0;
> >>  				}
> 
> FREE_AND_NULL(new)?

Ah, yeah, I forgot we had that now. It would work here, but note that this code ends up going away later in the series.

-Peff