One quick thing that I've seen. Haven't done a full review:
On Fri, Apr 15, 2011 at 09:57, <rhuij...@apache.org> wrote:
>...
> +++ subversion/trunk/subversion/libsvn_wc/workqueue.c Fri Apr 15 13:57:32 2011
>...
> +run_file_commit(svn_wc__db_t *db,
> + const svn_skel_t *work_item,
> + const char *wri_abspath,
> + svn_cancel_func_t cancel_func,
> + void *cancel_baton,
> + apr_pool_t *scratch_pool)
> +{
> + const svn_skel_t *arg1 = work_item->children->next;
> + const char *local_relpath;
> + const char *local_abspath;
> + svn_boolean_t remove_executable;
> + svn_boolean_t set_read_write;
> + apr_int64_t v;
> +
> + local_relpath = apr_pstrmemdup(scratch_pool, arg1->data, arg1->len);
> + SVN_ERR(svn_wc__db_from_relpath(&local_abspath, db, wri_abspath,
> + local_relpath, scratch_pool,
> scratch_pool));
> +
> + SVN_ERR(svn_skel__parse_int(&v, arg1->next, scratch_pool));
> + set_read_write = (v != 0);
> +
> + SVN_ERR(svn_skel__parse_int(&v, arg1->next, scratch_pool));
This should be arg1->next->next
Maybe this bug isn't properly tested in our test suite?
>...
Cheers,
-g
