Re: No longer possible to query catalogs for index capabilities? - Mailing list pgsql-hackers

From Andrew Gierth
Subject Re: No longer possible to query catalogs for index capabilities?
Date
Msg-id 87y445ap58.fsf@news-spur.riddles.org.uk
Whole thread Raw
In response to Re: No longer possible to query catalogs for index capabilities?  (Alvaro Herrera <alvherre@2ndquadrant.com>)
Responses Re: No longer possible to query catalogs for index capabilities?  (Andrew Gierth <andrew@tao11.riddles.org.uk>)
List pgsql-hackers
>>>>> "Alvaro" == Alvaro Herrera <alvherre@2ndquadrant.com> writes:
>> One idea is utils/adt/misc.c.  Or we could make a new file under>> utils/adt/ though I'm not very sure what to name
it. amaccess.c?>> catutils.c?  If there's only ever likely to be one or two functions>> of this ilk, maybe a new file
isoverkill and we should just use>> misc.c.
 
Alvaro> I like the idea of a new file; I have a hunch that it willAlvaro> grow, given that we're expanding in this
area,and perhaps weAlvaro> can find some existing stuff to relocate there in the future.Alvaro> I don't think a small
fileis a problem, anyway.
 
Alvaro> How about amfuncs.c?  Maybe it can live in catalog/ instead ofAlvaro> utils/adt?

Well, the existing patch used access/index/amapi.c for the AM capability
functions. There may be some merit in keeping everything together - I
asked because it didn't seem at first glance that the index column
property function belonged there, but on second thought there's some
overlap in that in future, if indoptions ever acquires any AM-specific
flags, it may be necessary for pg_index_column_has_property to call into
an AM-specific function.

So, here are some options:

1. Put everything in access/index/amapi.c

2. Move either all of access/index/amapi.c, or just the SQL-callable  part of it (amvalidate), to utils/adt/amfuncs.c
andput new stuff in  there
 

3. put pg_index[am]_has_capability in access/index/amapi.c and  pg_index_column_has_property in utils/adt/misc.c

-- 
Andrew (irc:RhodiumToad)



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Set log_line_prefix and application name in test drivers
Next
From: Andrew Borodin
Date:
Subject: Re: Re: GiST optimizing memmoves in gistplacetopage for fixed-size updates [PoC]