On Wed, 27 Jun 2001, Karel Zak wrote:
> Sure, nobody wants to works with something other than is in the
> pg_type.
>
> > problem with your idea is question if you have two functions (for example)
> > foo(timestamp) and foo(int4), you must embed the types into relname, and
> > that's ugly.
>
> Good point. First, you needn't work with types, bacause function oid
> is unique for foo(timestamp) and foo(int4). You can work with function
> oid. But this is not important.
That's not nice. GRANT ALL ON FOO_231234 where 231234 is OID of foo(int4)?
ew.
> The important thing is that in the PostgreSQL is already resolved very
> simular problem. We can define function with same names, unique must
> be function_name + arguments_types. Why not add same thing for tables and
> allows to define as unique table_name + table_type (where table_type
> is 'standard table', 'foo() table' and in future may be some other
> special type of table).
> The parser detect type of table very easy -- 'foo' vs. 'foo()'.
This is a little bit better, but, results in following syntax:
GRANT SELECT ON FOO(int4). I'm not sure if this really makes sense. Its
not a select permission, its an execute permission on a function, and
should be handled when/where execute permission is checked.
Its not hard to implement (just change what parser thinks relation is),
but I'm sure will conflict with _something_.
> IMHO very important is how add new feature and use it together with
> old feature.
>
> > Once its possible to control permission to execute a function via GRANT,
> > it solves the grant problem for function-as-tablesource
>
> The permissions system was an example only. If you add "foo()-tables"
> as something what needs special usage and care you probably found more
> problems. For example, what show command '\d' in the psql client, how
> relation show pg_access ..etc?
\df
Its a function, not a relation. You can do a lot of things to a relation
(such as define rules, triggers, constraints), which do not make any sense
for a function. The function may be used as a table-source, but it does
not make it a table.
If you can give me a better example than permissions system, I'll surely
reconsider, but currently, I see no use for it...
-alex
