Paul Querna wrote:
It will always be in the subversion history.If its redone and isn't a leaky abstraction, then sure, we can look at bringing it back, this vote doesn't stop that from happening. This vote is about what we want to do in the short term, and frankly the LDAP stuff has staggered around for years, and the split between mod_ldap and apr-util/ldap has never made sense to me. Either APR needs to wrap the whole thing, or it shouldn't at all, where we are today just causes trouble. It doesn't matter than LDAP is a standard. Lots of things are 'standard' APIs, but different implementations will still manage to mess it up, and I don't see LDAP as being any different in that regard. So, again, this is about what we are doing NOW, it doesn't prevent anyone from coming back to LDAP in the future and doing it 'right', for whatever value of 'right' it might have.
In that case we agree. Two things I'd like to ensure that happen are:- When I get a chance to set aside some time to sort out the LDAP abstraction, I don't want to find people saying afterwards "oh, but I thought we weren't doing LDAP any more"; and
- When end users suddenly see LDAP not-there in apr-2.0, and ask "and now?", we have a clear story to tell them, and not just leave them in the lurch.
(It is for this second reason why these kind of discussions need to be done on-list, because someone is going to suddenly start trawling the mailing list history for the reasons why apr-2.0 isn't supporting an API they rely on, and without any history they will be left in the dark).
Regards, Graham --
smime.p7s
Description: S/MIME Cryptographic Signature
