Martin Hauner <martin.hauner@gmx.net> writes:
> I found a patch to svn_client_add on my disk that i did some time ago (I
> didn't post it to the list) after asking about an svn_client api
> inconsistencies: Why do revert/delete take an array as input but add does not?
>
> Statement then was that you moved to multiple input paths but somehow add
> was never changed.
>
> My patch changes svn_client_add to take a const apr_array_header_t *paths
> parameter instead of a single char* path.
>
> Today i started to finish the patch and to adjust it to the current trunk.
> My problem now is that svn_cl__add has changed in a way that makes it difficult
> to change it. svn_cl__add uses svn_cl__try on each item to add to skip a few
> errors without breaking the loop.
>
> I can't move svn_cl__try to svn_client_add but without moving the check to
> svn_client_add the improved behaviour of svn add would be lost.
>
> I think we need to decide
>
> 1. if we forget about the improvement to the api
>
> or
>
> 2. if we want to improve the api and add something equivalent to the
> svn_cl__try check to svn_client_add.
I don't think this improvement to the API is necessary. I think it's
too bad that svn_client_revert() and svn_client_delete() aren't
consistent with svn_client_add3() right now, but it's not a big enough
deal to be worth the code churn to make them consistent. (Especially
since in revert and delete aren't even consistent with each other in
the way they handle errors anyway).
Just my $0.02,
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Jul 8 23:13:00 2005