Re: atomic revprop commits at the RA level

Eric Gillespie wrote:
>>Suggest:
>
>
>> svn_error_t *(*change_edit_prop)(void *edit_baton,
>> const char *name,
>> const svn_string_t *value,
>> apr_pool_t *pool);
>
> Why not make it a hash, as you suggested when you suggested
> modifying svn_ra_get_commit_editor? No sense in adding
> unnecessary round trips...

There's nothing in this API that necessitates roundtrips. If the editor
implementor wishes to cache all the edit-prop changes and send them at one
time (maybe at close_edit() time), it can choose to do so. Using a hash,
however, *forces* the editor driver to harvest all those properties into
memory at once. I don't see the value in that.

Plus, the interface as provided is consistent with the change_dir_prop() and
change_file_prop() interfaces, and consistency r00lz.

-- 
C. Michael Pilato <cmpilato@collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand