Re: character encoding in StartupMessage

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Christopher Kings-Lynne <chriskl(at)familyhealth(dot)com(dot)au>
Cc: John DeSoi <desoi(at)pgedit(dot)com>, PG Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: character encoding in StartupMessage
Date: 2006-02-28 06:38:03
Message-ID: 10937.1141108683@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Christopher Kings-Lynne <chriskl(at)familyhealth(dot)com(dot)au> writes:
>> I could not find anything in the Frontend/Backend protocol docs about
>> character encoding in the StartupMessage. Assuming it is legal for a
>> database or user name to have unicode characters, how is this handled
>> when nothing yet has been said about the client encoding?

> A similar badness is that if you issue CREATE DATABASE from a UTF8
> database, the dbname will be stored as UTF8. Then, if you go to a
> LATIN1 database and create another it will be stored as LATIN1.

Yeah, this has been discussed before. Database and user names both
have this affliction.

I don't see any very nice solution at the moment. Once we get support
for per-column locales, it might be possible to declare that the shared
catalogs are always in UTF8 encoding and get the necessary
conversions to happen automatically.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Christopher Kings-Lynne 2006-02-28 06:42:22 ipcclean in 8.1 broken?
Previous Message Tom Lane 2006-02-28 06:32:19 Re: Vacuum dead tuples that are "between" transactions