Change format of FDW options used in \d* commands (was: Re: per-column FDW options, v5) - Mailing list pgsql-hackers

From Shigeru Hanada
Subject Change format of FDW options used in \d* commands (was: Re: per-column FDW options, v5)
Date
Msg-id 4E44FF43.60201@gmail.com
Whole thread Raw
In response to Re: per-column FDW options, v5  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: Change format of FDW options used in \d* commands (was: Re: per-column FDW options, v5)
List pgsql-hackers
(2011/08/12 1:05), Robert Haas wrote:
> On Thu, Aug 11, 2011 at 12:04 PM, Alvaro Herrera
> <alvherre@commandprompt.com>  wrote:
>> Excerpts from Robert Haas's message of jue ago 11 11:50:40 -0400 2011:
>>> 2011/8/9 Shigeru Hanada<shigeru.hanada@gmail.com>:
>>>> I'd like to pick #3, and also change per-column options format.  In
>>>> addition, I'd like to change options format for other FDW objects such
>>>> as wrappers, servers and user mappings for consistency.  Of course, only
>>>> if it's acceptable to break backward compatibility...
>>>
>>> I think it's fine to change the display format.  We haven't had these
>>> features for very long, so users hopefully shouldn't be expecting that
>>> everything is set in stone.  We have made far bigger changes to
>>> backslash commands that have been around for far longer (\df, I'm
>>> looking at you).
>>
>> We've never promised that backslash commands behave identically across
>> releases.  I think they are more for human consumption than machine, so
>> why would we care about changing one of them a bit?
>
> Yeah, I agree.

Thanks for the comments.

Attached patch changes various \d*+ commands to show FDW options in same
format as OPTIONS clause.  IMHO the new format is easier to read.

Example)
  old: {"delimiter=,","quote=\""}
  new: delimiter ',', quote '"'

All regression tests including contrib's installcheck passed.

I'll add this patch to CF app as new item.

Regards,
--
Shigeru Hanada


Attachment

pgsql-hackers by date:

Previous
From: Cédric Villemain
Date:
Subject: Re: index-only scans
Next
From: Simon Riggs
Date:
Subject: Re: our buffer replacement strategy is kind of lame