Re: autogenerating error code lists (was Re: [COMMITTERS] pgsql: Add foreign data wrapper error code values for SQL/MED.) - Mailing list pgsql-hackers

From Jan Urbański
Subject Re: autogenerating error code lists (was Re: [COMMITTERS] pgsql: Add foreign data wrapper error code values for SQL/MED.)
Date
Msg-id 4D179B46.8050005@wulczer.org
Whole thread Raw
In response to Re: autogenerating error code lists (was Re: [COMMITTERS] pgsql: Add foreign data wrapper error code values for SQL/MED.)  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: autogenerating error code lists (was Re: [COMMITTERS] pgsql: Add foreign data wrapper error code values for SQL/MED.)  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On 26/12/10 20:33, Tom Lane wrote:
> Jan Urbański <wulczer@wulczer.org> writes:
>> I then tried to do the same for plerrcodes.h, but found discrepancies.
> 
> You shouldn't assume that errcodes.h necessarily includes all the info
> needed to create the other two files.
> 
> The way I'd envision this working is that we have a master file
> containing the five-letter SQLSTATE code, the ERRCODE macro name,
> the string name to use in plerrcodes.h, and any other required items,
> and then generate all three of the existing files from that.

How about a format like this then:

# Comment
Section: Class 2F - SQL Routine Exception
macro_name sqlstate plpgsqlname is_error

That is: # and blank lines are comments, lines starting with "Section:"
are section names (needed for SGML output), the rest are whitespace
separated strings. is_error is 0 or 1, if it's 0 we don't generate a
plpgsql condition for it.

If we go for that format, we'll end up with one new plpgsql error
(nonstantard_use_of_escape_character) that was omitted from previous
releases.

Jan


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: knngist - 0.8
Next
From: Tom Lane
Date:
Subject: Re: autogenerating error code lists (was Re: [COMMITTERS] pgsql: Add foreign data wrapper error code values for SQL/MED.)