Re: pg_get_triggerdef in pg_dump - Mailing list pgsql-hackers

From Andreas Pflug
Subject Re: pg_get_triggerdef in pg_dump
Date
Msg-id 3EF18DB0.6090408@web.de
Whole thread Raw
In response to Re: pg_get_triggerdef in pg_dump  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:

>>Yes, you can check if they're binary compatible from the pg_cast table....
>>    
>>
>
>But nearly all of the interesting cases require you to understand the
>type's interpretation of typmod, and you can't learn that from a table.
>How many cases are there where blindly looking for a binary-compatible
>cast in pg_cast will really do you much good?  AFAICS you'd have to set
>atttypmod to -1 if you change atttypid without knowing very specifically
>what you are changing from and to.
>
>  
>
AFAICS there's few interpretation about atttypmod necessary because only 
few datatypes binary convertible (castfunc=0) do use atttypmod at all. 
Most special case is varchar<->text, one supporting length, the other 
not; both need atttypmod=-1. bpchar<->varchar both allow typmod in a 
similar fashion.
It's already implemented in pgAdmin3 this way.

Regards,
Andreas



pgsql-hackers by date:

Previous
From: Andreas Pflug
Date:
Subject: Re: allowed user/db variables
Next
From: Andreas Pflug
Date:
Subject: Re: allowed user/db variables