Horton

Re: Adding the extension name to EData / log_line_prefix

Chapman Flack <jcflack@acm.org>

View thread
From: Chapman Flack <jcflack@acm.org>
To: Tom Lane <tgl@sss.pgh.pa.us>, Andres Freund <andres@anarazel.de>
Cc: Peter Eisentraut <peter@eisentraut.org>, Fabrízio de Royes Mello <fabriziomello@gmail.com>, pgsql-hackers@postgresql.org
Date: 2024-05-15T17:58:54Z
Lists: pgsql-hackers 
On 05/15/24 13:45, Tom Lane wrote:
> if we tell people to write
> 
>    PG_MODULE_MAGIC;
>    #undef TEXTDOMAIN
>    #define TEXTDOMAIN PG_TEXTDOMAIN("hstore")
> 
> then that's 100% backwards compatible and they don't need any
> version-testing ifdef's.

OT for this thread, but related: supposing out-of-core extensions
participate increasingly in NLS, would they really want to use
the PG_TEXTDOMAIN macro?

That munges the supplied domain name with PG's major version and
.so version numbers.

Were such versioning wanted for an out-of-core extension's message
catalogs, wouldn't the extension's own versioning be better suited?

Regards,
-Chap