Johan Herland wrote:
> --- a/notes.c
> +++ b/notes.c
> @@ -303,13 +298,17 @@ static void note_tree_insert(struct notes_tree *t, struct int_node *tree,
> GET_PTR_TYPE(*p) == PTR_TYPE_SUBTREE);
> if (is_null_sha1(entry->val_sha1)) { /* skip insertion of empty note */
> free(entry);
> - return;
> + return 0;
> }
> new_node = (struct int_node *) xcalloc(sizeof(struct int_node), 1);
> - note_tree_insert(t, new_node, n + 1, l, GET_PTR_TYPE(*p),
> - combine_notes);
> - *p = SET_PTR_TYPE(new_node, PTR_TYPE_INTERNAL);
> - note_tree_insert(t, new_node, n + 1, entry, type, combine_notes);
> + ret = note_tree_insert(t, new_node, n + 1, l, GET_PTR_TYPE(*p),
> + combine_notes);
> + if (!ret) {
> + *p = SET_PTR_TYPE(new_node, PTR_TYPE_INTERNAL);
> + ret = note_tree_insert(t, new_node, n + 1, entry, type,
> + combine_notes);
> + }
> + return ret;
Micronit: it would probably be clearer to write
if (ret)
return ret;
*p = SET_PTR_TYPE(...
return note_tree_insert(...
to avoid a little nesting.
The patch still looks good to me.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html