Am 22.05.2012 12:16, schrieb Jim Meyering:
> From: Jim Meyering <[email protected]>
>
> Without this, envlist_to_environ may silently fail to copy all
> strings into the destination buffer, and both callers would leak
> any env strings allocated after a failing strdup, because the
> freeing code stops at the first NULL pointer.
>
> Signed-off-by: Jim Meyering <[email protected]>
> ---
> envlist.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/envlist.c b/envlist.c
> index e44889b..df5c723 100644
> --- a/envlist.c
> +++ b/envlist.c
> @@ -234,8 +234,16 @@ envlist_to_environ(const envlist_t *envlist, size_t
> *count)
> return (NULL);
>
> for (entry = envlist->el_entries.lh_first; entry != NULL;
> - entry = entry->ev_link.le_next) {
> - *(penv++) = strdup(entry->ev_var);
> + entry = entry->ev_link.le_next, penv++) {
Scratch my comment on 1/2, there's an added penv++ that I overlooked.
Not changing the indentation twice would still be nice.
> + *penv = strdup(entry->ev_var);
> + if (*penv == NULL) {
> + char **e = env;
> + while (e <= penv) {
> + free(*e++);
> + }
> + free(env);
> + return NULL;
> + }
> }
> *penv = NULL; /* NULL terminate the list */
>
This leak fix looks good then.
For anyone wondering like me, the "env" here is not the usual
CPUArchState *env but a local char **env.
Andreas
--
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
