[Samba4][Patch] Implement idmap for winbind (try 2)
kai at samba.org
Tue Feb 19 16:38:26 GMT 2008
On Tuesday 19 February 2008 17:25:59 Stefan (metze) Metzmacher wrote:
> Kai Blin schrieb:
> 1.) idmap_ctx should be long term hanging on struct wbsrv_service
Check. I'll look into that.
> 2.) the usage of tmp_ctx is useless as 'state' is already a context
> with the correct scope.
> 3.) you should first try to do a ldb_search without a transaction
> before the code you currently have:
> - if we find the mapping we do a search and are done
> - if we don't find a mapping we should start a transaction
> and redo the seach to make sure no one else has added it
> in between, then we should create a new mapping.
Ah, I didn't understand it when you tried to explain this in IRC. That makes
> 4.) the if (new_uid != low) logic breaks if
> there's the highwater mark set and the value match the low
> bound, you would end up with two values for the attribute...
You were the one who said I shouldn't set up the database like that in the
first place so the code could just always go that path.
> 5.) please use hwm++; instead of ++hwm;
Sure, all the same on it's own.
> 6.) if you have a line break within a function call, indent the 2nd line
> to the opening '(' and not just one \t.
> 7.) Doesn't simo proposed that we allocate uids and gids from the same
> range. So that a number is never given away as uid and gid.
The way I understood simo, this was about using the same range for both uids
and gids, but two pools of numbers. What would be the benefit of only using
WorldForge developer http://www.worldforge.org/
Wine developer http://wiki.winehq.org/KaiBlin
Samba team member http://www.samba.org/samba/team/
Will code for cotton.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://lists.samba.org/archive/samba-technical/attachments/20080219/1daeb25d/attachment.bin
More information about the samba-technical