On 06/04/2012 10:44 PM, Greg Stein wrote:
> On Mon, Jun 4, 2012 at 4:05 PM,<stefan2_at_apache.org> wrote:
>> ...
>> +++ subversion/trunk/subversion/libsvn_ra_svn/marshal.c Mon Jun 4 20:05:23 2012
>> ...
>> @@ -1163,24 +1160,18 @@ svn_error_t *svn_ra_svn_write_cmd_respon
>> va_list ap;
>> svn_error_t *err;
>>
>> - SVN_ERR(svn_ra_svn_start_list(conn, pool));
>> - SVN_ERR(svn_ra_svn_write_word(conn, pool, "success"));
>> + SVN_ERR(writebuf_write_short_string(conn, pool, "( success ", 10));
> This kinda breaks the encapsulation, but it seems okay since it's the
> same file after all :-)
Yeah .. well, it sure won't win any beauty contest ...
>> va_start(ap, fmt);
>> err = vwrite_tuple(conn, pool, fmt, ap);
>> va_end(ap);
>> - if (err)
>> - return err;
>> - SVN_ERR(svn_ra_svn_end_list(conn, pool));
>> - return SVN_NO_ERROR;
>> + return err ? err : svn_ra_svn_end_list(conn, pool);
> On something like this, it is nice to put an svn_error_trace() in there.
Fixed in r1348471.
Thanks for the review!
-- Stefan^2.
Received on 2012-06-09 20:51:05 CEST