Re: Casts question - Mailing list pgsql-hackers

From Shachar Shemesh
Subject Re: Casts question
Date
Msg-id 40D5EF3D.6040802@shemesh.biz
Whole thread Raw
In response to Re: Casts question  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Casts question  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:

>Shachar Shemesh <psql@shemesh.biz> writes:
>  
>
>>I have defined a datatype called "varcharci", shamelessly yanking the 
>>input, output, recv and send functions from varchar. This means (as far 
>>as I understand things) that this type is binary compatible with varchar.
>>    
>>
>
>Use text, not varchar.
>
>  
>
>>Why is that? Being as it is that no operator = is defined for varcharci, 
>>and that the cast from varchar to varcharci is "as assignment" anyways, 
>>shouldn't postgres be able to do the cast implicitly?
>>    
>>
>
>Yes, it can cast to varchar, but that doesn't help because there are no
>varchar operators ;-).  To resolve the operator, it has to promote both
>sides to text, and you didn't offer a cast to text.
>
>            regards, tom lane
>
>  
>
I don't get it. The cast from varchar to text is using a "no function" 
cast - i.e. - they are binary compatible. And yet, there are two 
seperate functions for receiving from text and from binary 
representation. Why not use the same function?
            Shachar

-- 
Shachar Shemesh
Lingnu Open Source Consulting ltd.
http://www.lingnu.com/



pgsql-hackers by date:

Previous
From: Pierre Emmanuel Gros
Date:
Subject: Re: constraint upon view
Next
From: Tom Lane
Date:
Subject: Re: Casts question