On Thu, Dec 04, 2014 at 12:51:04PM +0100, Pavel Březina wrote:
On 12/03/2014 05:58 PM, Jakub Hrozek wrote:
>On Wed, Dec 03, 2014 at 04:33:33PM +0100, Pavel Březina wrote:
>>>- if (groupnames[i] == NULL) {
>>>+ tmpstr = sss_replace_space(groupnames, name,
>>>+
ireq->ifp_ctx->rctx->override_space);
>>
>>Sure, you can allocate tmpstr on groupnames, I think ireq or tmp_ctx would
>>be better choice.
>>
>>>+ if (tmpstr == NULL) {
>>> DEBUG(SSSDBG_MINOR_FAILURE, "Cannot normalize
%s\n", name);
>>> continue;
>>> }
>>> } else {
>>>- groupnames[i] = name;
>>>+ tmpstr = name;
>>> }
>>>+
>>>+ groupnames[i] = sss_get_cased_name(ireq, tmpstr,
domain->case_preserve);
>>
>>Here, groupnames should be used as talloc context instead of ireq.
>>sss_get_cased_names duplicates the string so tmpstr should be freed.
>>
>>>+ if (groupnames[i] == NULL) {
>>>+ DEBUG(SSSDBG_CRIT_FAILURE,
>>>+ "sss_get_cased_name failed, skipping\n");
>>>+ continue;
>>>+ }
>>>+
>>> DEBUG(SSSDBG_TRACE_FUNC, "Adding group %s\n",
groupnames[i]);
>>> }
>
>Both suggestions make sense, please see the attached patch. Thank you
>for the review.
Ack.
* master: 1b4bd7e378deda73a18d86e4b2998bff45883e7b