Jeff King <peff@xxxxxxxx> writes:
> There's a tiny leak in this one. Here's a fix that can go on top.
>
> -- >8 --
> Subject: [PATCH] fsmonitor: fix leak of warning message
>
> The fsm_settings__get_incompatible_msg() function returns an allocated
> string. So we can't pass its result directly to warning(); we must hold
> on to the pointer and free it to avoid a leak.
>
> The leak here is small and fixed size, but Coverity complained, and
> presumably SANITIZE=leaks would eventually.
>
> Signed-off-by: Jeff King <peff@xxxxxxxx>
> ---
> fsmonitor.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
Is there anybody else who is reading Coverity reports, I wonder.
This one is a clear positive.
Thanks.
> diff --git a/fsmonitor.c b/fsmonitor.c
> index 540736b39f..08af00c738 100644
> --- a/fsmonitor.c
> +++ b/fsmonitor.c
> @@ -309,8 +309,10 @@ void refresh_fsmonitor(struct index_state *istate)
> enum fsmonitor_reason reason = fsm_settings__get_reason(r);
>
> if (!warn_once && reason > FSMONITOR_REASON_OK) {
> + char *msg = fsm_settings__get_incompatible_msg(r, reason);
> warn_once = 1;
> - warning("%s", fsm_settings__get_incompatible_msg(r, reason));
> + warning("%s", msg);
> + free(msg);
> }
>
> if (fsm_mode <= FSMONITOR_MODE_DISABLED ||
