On 22 May 2012 16:57, Florian Pflug <fgp@phlo.org> wrote:
> On May22, 2012, at 16:09 , Tom Lane wrote:
>> Thom Brown <thom@linux.com> writes:
>>> Conflicts would occur where localrolename matches an existing local
>>> role name within the same database, or a global role name, but not a
>>> local role name within another database. The problem with this,
>>> however, is that creating global roles would need conflict checks
>>> against local roles in every database, unless a manifest of all local
>>> roles were registered globally.
>>
>> Yeah. The same type of issue arises for the roles' OIDs. You'd really
>> want local and global roles to have nonconflicting OIDs, else it's
>> necessary to carry around an indication of which type each role is;
>> which would be more or less a show-stopper in terms of the number of
>> catalogs and internal APIs affected. But I don't currently see any
>> nice way to guarantee that if each database has a private table of
>> local roles.
>
> Maybe we could simply make all global role's OIDs even, and all local ones
> odd, or something like that.
Wouldn't that instantly make all previous versions of database
clusters un-upgradable?
--
Thom