[SAMBA4][PATCH] Recursive registry key delete (from bz #5178)

Jelmer Vernooij jelmer at samba.org
Wed Feb 20 02:14:49 GMT 2008

On Tue, Feb 19, 2008 at 03:46:22PM -0600, Andrew Kroeger wrote:
> > Rather than ifdeffing out some tests, I would rather see the rpc-winreg
> > test be converted to the new torture API and having the individual tests
> > marked as known failing. AFAIK these tests pass against Windows, so it
> > would be a shame to disable them against Windows as well.
> No problem.  I had seen other tests that have blocks ifdeffed out, but
> that probably predates the new torture API.  Do you have any quick
> pointers to use of the new torture API (vs. the old way of doing things)
> so I can make sure I'm following the correct examples?
See torture/ui.h for the full API. The local registry tests are a good
example of the use of the new API.

> > Please don't move functions within the file unless there's a reason for
> > it. It makes it harder to review.
> I moved ldb_del_key() after ldb_del_value() in ldb.c, because I need to
> call ldb_del_value() from ldb_del_key().  As ldb_del_value() is static,
> it isn't defined in any of the header files.  Would it be better to just
> add a declaration of the function earlier in the .c file, or is moving
> functions OK in this case?
Ah, ok - I hadn't noticed that. In that case this is a somewhat gray area :-) 
Generally moving is better because it saves a forward declaration, but it does 
make review harder. I can live with either :-)

> >> P.P.S - I'm not sure as to the patch submission protocol with Git - is
> >> it OK to just point to the patches in my repo, or is it better to keep
> >> sending patches to samba-technical to maintain an audit trail of what I
> >> am actually submitting?
> > Having the patches applied to the mailing list post is nice because it
> > makes it easier to review and comment in-line where necessary.
> Sounds good to me.  I will work your feedback into a new set of patches.

> Still being new to Git, will it create any problems if I reset my
> published repo to remove the original patch versions I pushed?  If I
> understand things correctly, that shouldn't cause any problems unless
> someone has based any new commits off of the history I pushed to my
> topic branch, correct?
Yep, that's correct. Resetting your repo shouldn't cause trouble for
anybody unless, as you say, somebody has already merged them and is
committing on top of them.



More information about the samba-technical mailing list