On 01/27/2011 09:56 AM, Bert Huijben wrote: >> Modified: subversion/branches/uris-as- >> urls/subversion/libsvn_client/locking_commands.c
[...]
>> @@ -206,9 +206,16 @@ organize_lock_targets(const char **commo
>> {
>> const char *parent, *base;
>> if (url_mode)
>> - svn_url_split(&parent, &base, *common_parent_url,
>> result_pool);
>> + {
>> + svn_url_split(&parent, &base, *common_parent_url,
>> + result_pool);
>> + svn_path_uri_encode(base, result_pool);
>
> I think you tried to store the result in base?
>
> Or should the encode be removed?
Encode can't be removed just yet -- I'll tackle that in a subsequent pass.
But yes, I definitely should have kept the returned value!
>> Modified: subversion/branches/uris-as-
>> urls/subversion/libsvn_subr/dirent_uri.c
[...]
>> @@ -1365,8 +1365,7 @@ svn_url_basename(const char *uri, apr_po
>> else
>> base_name = uri + start;
>>
>> - /* ### TODO: URI decode a non-NULL base_name? */
>> - return base_name;
>> + return svn_path_uri_decode(base_name, pool);
>> }
>
> I wouldn't be surprised if there is a static function in this file which
> would avoid the copying if possible. (If not, that would be a nice
> addition for a future enhancement)
Hrm. Agreed. That would allow us to go back to not requiring a pool for
svn_url_basename().
--
C. Michael Pilato <cmpil...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
signature.asc
Description: OpenPGP digital signature
