Re: more ALTER .. DEPENDS ON EXTENSION fixes

On 2020-Mar-11, Tom Lane wrote:

> BTW, I did not like the syntax too much. "NO DEPENDS ON EXTENSION"
> doesn't seem like good English. "NOT DEPENDS ON EXTENSION" is hardly
> any better. The real problem with both is that an ALTER action should
> be, well, an action. A grammar stickler would say that it should be
> "ALTER thing DROP DEPENDENCY ON EXTENSION ext", but perhaps we could
> get away with "ALTER thing DROP DEPENDS ON EXTENSION ext" to avoid
> adding a new keyword. By that logic the original command should have
> been "ALTER thing ADD DEPENDS ON EXTENSION ext", but I suppose it's
> too late for that.

The problem with DROP DEPENDS is alter_table_cmd, which already defines
"DROP opt_column ColId", so there's a reduce/reduce conflict for the
ALTER INDEX and ALTER MATERIALIZED VIEW forms because "depends" could be
a column name. (It works fine for ALTER FUNCTION/ROUTINE/PROCEDURE/TRIGGER
because there's no command that tries to define a conflicting DROP form
for these.)

It works if I change DEPENDS to be type_func_name_keyword (currently
unreserved_keyword), but I bet we won't like that.

(DEPENDENCY is not a keyword of any kind, so DROP DEPENDENCY require us
making it one of high reservedness, which I suspect we don't like
either).

It would also work to use a different keyword in the DROP position;
maybe REMOVE. But that's not a keyword currently.

How about ALTER .. REVOKE DEPENDS or DELETE DEPENDS? Bison is okay
with either of those forms.

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services