Re: [PATCH] Added TRANSFORM FOR for COMMENT tab completion

>>> +   else if (Matches("COMMENT", "ON", "PROCEDURAL"))
>>> +       COMPLETE_WITH("LANGUAGE");
>>> +   else if (Matches("COMMENT", "ON", "PROCEDURAL", "LANGUAGE"))
>>> +       COMPLETE_WITH_QUERY(Query_for_list_of_languages);
>>> I don't think that there is much point in being this picky either 
>>> with
>>> the usage of PROCEDURAL, as we already complete a similar and simpler
>>> grammar with LANGUAGE.  I would just remove this part of the patch.
>> In my opinion, it is written in the documentation, so tab-completion 
>> of "PROCEDURAL"is good.
>> How about a completion with "LANGUAGE" and "PROCEDURAL LANGUAGE", like 
>> "PASSWORD" and "ENCRYPTED PASSWORD" in CREATE ROLE?

I kept LANGUAGE and PROCEDURAL LANGUAGE just like PASSWORD and ENCRYPTED 
PASSWORD.


>>> +   else if (Matches("COMMENT", "ON", "OPERATOR"))
>>> +       COMPLETE_WITH("CLASS", "FAMILY");
>>> Isn't this one wrong?  Operators can have comments, and we'd miss
>>> them.  This is mentioned upthread, but it seems to me that we'd 
>>> better
>>> drop this part of the patch if the operator naming part cannot be
>>> solved easily.
>> As you said, it may be misleading.
>> I agree to drop it.

Hearing all the opinions given, I decided not to support OPERATOR CLASS 
or FAMILY in COMMENT.
Therefore, I drooped Query_for_list_of_operator_class_index_methods as 
well.


> +static const SchemaQuery Query_for_list_of_text_search_configurations 
> = {
> 
> We already have Query_for_list_of_ts_configurations in tab-complete.c.
> Do we really need both queries? Or we can drop either of them?

Thank you for pointing out!
I didn't notice that there already exists 
Query_for_list_of_ts_configurations,
so I changed TEXT SEARCH completion with using Query_for_list_of_ts_XXX.

I made the changes to the points above and updated the patch.

-- 
Best wishes,

Ken Kato
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION