Subject: Re: Type scale returned by PQfmod() 65531 for time/timestamp output
 parameter?
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-interfaces@lists.postgresql.org
References: <093e1214-7057-f80c-5c3c-74076b7bb555@4js.com>
 <7537.1518711940@sss.pgh.pa.us>
 <f80e250d-b13f-0852-2a11-03218588092f@4js.com>
 <13939.1518794152@sss.pgh.pa.us>
 <91fabdcd-0eb3-c970-a28c-248a6ff8a360@4js.com>
 <7047.1519056383@sss.pgh.pa.us>
 <2ae51da1-f2b4-a8be-4c4d-c8b96f1f7718@4js.com>
 <17186.1519137546@sss.pgh.pa.us>
 <5a7de250-aceb-7b61-8da6-796a37a28eca@4js.com>
 <21465.1519238190@sss.pgh.pa.us>
From: Sebastien FLAESCH <sf@4js.com>
Organization: Four Js Development Tools
Message-ID: <a6fd2b01-defd-5a26-d9e8-9a31537fa970@4js.com>
Date: Thu, 22 Feb 2018 13:35:12 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.5.2
MIME-Version: 1.0
In-Reply-To: <21465.1519238190@sss.pgh.pa.us>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 02/21/2018 07:36 PM, Tom Lane wrote:
> Sebastien FLAESCH <sf@4js.com> writes:
>> On 02/20/2018 03:39 PM, Tom Lane wrote:
>>> Like I said before, it's datatype-specific and you need to look at the
>>> typmodin/typmodout support functions for each type to see what they do.
> 
>> Are you suggesting me to dig into the PostgreSQL server sources / internals?
> 
> Yup.
> 
>> Any starting point I should look at?
> 
> regression=# select distinct typmodout from pg_type where typmodout != 0;
>        typmodout
> ----------------------
>   intervaltypmodout
>   timestamptypmodout
>   timestamptztypmodout
>   timetypmodout
>   timetztypmodout
>   bpchartypmodout
>   varchartypmodout
>   numerictypmodout
>   bittypmodout
>   varbittypmodout
> (10 rows)
> 
> I think all of those are under src/backend/utils/adt/ in the sources.
> Briefly their charter is to produce the textual representation of a
> typmod value for the data type, or an empty string if there's no typmod
> constraint.  Although in principle code outside the datatype shouldn't
> assume anything at all about the encoding of typmod, there's a widespread
> assumption that all negative values (not just -1) mean "no constraint".
> 
> 			regards, tom lane
> 

OK Thanks!
Seb