Yes, svn_sqlite__update() *does* reset the statement before returning
(it isn't something where you can iterate over results; it is done; so
it resets the statement).
Thus, the extra reset should not be there. "Following the pattern"
isn't right here.
On Fri, Jul 16, 2010 at 13:44, Philip Martin <philip.mar...@wandisco.com> wrote:
> Bert Huijben <b...@qqmail.nl> writes:
>
>> Doesn't _update handle this specific reset? If not I think it should.
>
> The compiler was complaing about a missing argument to
> svn_error_createf. I was just following the pattern used by next
> function.
>
>> if (affected_rows != 1)
>> return svn_error_createf(SVN_ERR_WC_PATH_NOT_FOUND,
>> + svn_sqlite__reset(stmt),
>> _("The node '%s' was not found."),
>> svn_dirent_local_style(local_abspath,
>> scratch_pool));
>
> --
> Philip
>
