RE: [PROPOSAL]a new data type 'bytea' for ECPG - Mailing list pgsql-hackers

From Matsumura, Ryo
Subject RE: [PROPOSAL]a new data type 'bytea' for ECPG
Date
Msg-id 03040DFF97E6E54E88D3BFEE5F5480F737A383FB@G01JPEXMBYT04
Whole thread Raw
In response to RE: [PROPOSAL]a new data type 'bytea' for ECPG  ("Tsunakawa, Takayuki" <tsunakawa.takay@jp.fujitsu.com>)
Responses RE: [PROPOSAL]a new data type 'bytea' for ECPG
List pgsql-hackers
> From: Tsunakawa, Takayuki [mailto:tsunakawa.takay@jp.fujitsu.com]
> 
> I think the host variable data type that corresponds to the server-side bytea
> should be bytea.  As the following pages state or imply, it would be better
> to create standard-compliant LOB types someday, and use the keyword BLOB in
> ECPG for that type.  The server-side data types should have the names BLOB,
> CLOB and NCLOB.  Those types should handle data larget than 1 GB and have the
> locator feature defined in the SQL standard.  Maybe we should also advanced
> LOB features like Oracle's SecureFiles LOB and SQL Server's FileTables.

Tsunakawa-san, thanks for your advice.
I understand that C type definition of client-side bytea is not constrained by the standard BLOB.

What should I do next?
For now, I attach a patch that is removed noise(pgindent/typedef.list).

P.S.
The patch does not support ECPG.bytea in sqltype of "struct sqlvar_struct" because of compatibility.

Regards
Ryo Matsumura

Attachment

pgsql-hackers by date:

Previous
From: Melanie Plageman
Date:
Subject: Re: [HACKERS] [PATCH v2] Add and report the new "session_read_only"GUC pseudo-variable.
Next
From: Amit Langote
Date:
Subject: Re: BUG #15212: Default values in partition tables don't work asexpected and allow NOT NULL violation