[LDB] Simplify ldb_wait()

[Andrew Bartlett]

On Mon, 2007-11-12 at 20:56 +1100, Andrew Bartlett wrote:
> On Mon, 2007-11-12 at 03:05 -0500, simo wrote:
> > On Mon, 2007-11-12 at 18:55 +1100, Andrew Bartlett wrote:
> > > On Mon, 2007-11-12 at 02:50 -0500, simo wrote:
> > > > On Mon, 2007-11-12 at 08:32 +1100, Andrew Bartlett wrote:
> > > > > On Sun, 2007-11-11 at 12:33 -0500, simo wrote:
> > > > > > Why have you changed the ldb_wait() prototype from getting an handle to
> > > > > > get a full request?
> > > > > > It seem that the new ldab_wait() never uses req at all except as in
> > > > > > req->handle.
> > > > > 
> > > > > Because nobody passed ldb_wait anything by req->handle.  It also could
> > > > > allow the handle to be replaced, which would allow modules to simplify
> > > > > their tail (and avoid the full blocking that currently occurs, even when
> > > > > the module is just dealing with it's final operation). 
> > > > 
> > > > I think I decided to pass the handle and not the request intentionally,
> > > > right now I can't remember why. I think manipulation of the request from
> > > > modules may have been one reason, ie *avoid* making it possible to
> > > > change the handle you are waiting for. but memory is not good I might be
> > > > wrong.
> > > > 
> > > > > > It seem to me an unnecessary ABI change for ldb.
> > > > > 
> > > > > Are we seriously keeping an ABI on ldb yet?
> > > > 
> > > > No but I want to keep API (not ABI) changes to a minimum, and avoid
> > > > unnecessary changes.
> > > 
> > > So is this a yes or no? 
> > 
> > We committed not to change the API if not absolutely necessary.
> > There are other projects using ldb now (OpenChange is one of the most
> > prominent), and changing API is not welcome unless we need to do it for
> > a good reason.
> 
> This isn't a public API - it is the relationship between LDB and it's
> modules.  But if that is your position, so be it.

Sorry, I re-read the patch, and this statement isn't correct.

This isn't in the public sync API, which we have largely stuck to for
quite some time now.  Have we decided to freeze the async API at this
point (as I think it is still far from ideal)? 

Andrew Bartlett

-- 
Andrew Bartlett
http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Red Hat Inc.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.samba.org/archive/samba-technical/attachments/20071112/bea231fb/attachment.bin