Multibyte FAQ item

On Sat, 22 Sep 2001, Bruce Momjian wrote:

Do we need an multibyte/locale/unicode FAQ item? Questions about these
seem to be asked a lot, but the documentation for these items is quite
good.

Well IMHO, this one is very self explaning
<http://www.postgresql.org/idocs/index.php?multibyte.html&gt;

but these aren't...
(from pgsql 7.1.3, I can't grab stuff from the cvs at the moment for some
reason)

jaa @ eowyn 114 $ ./configure --help
--8<--
--enable-multibyte enable multibyte character support
--8<--

and
jaa @ eowyn 114 $ less INSTALL
--8<--
--enable-multibyte

Allows the use of multibyte character encodings. This is
primarily for languages like Japanese, Korean, and Chinese. Read
the Administrator's Guide for details.
--8<--

If you are 8-bit speaking people (ie. iso8859-1), you will just think ok,
thats not concern me. And at the that very same moment you will lose.

So, it will be nice if there is more verbose help text in the INSTALL
file, or even own FOR_8BIT_CHARS_READ_THIS file.

I think that there is enought information after you have realize that if
you would like use 8-bit chars, you have to be multibyte turned on.

If the situation isn't that at the moment, please forgot me.

BR, Jani

P.S. There was a thread about this topic in the jdbc-mailing list starting
from here:

<http://archives.postgresql.org/pgsql-jdbc/2001-05/msg00018.php&gt;

---
Jani Averbach

Jani Averbach writes:

--enable-multibyte

Allows the use of multibyte character encodings. This is
primarily for languages like Japanese, Korean, and Chinese. Read
the Administrator's Guide for details.
--8<--

If you are 8-bit speaking people (ie. iso8859-1), you will just think ok,
thats not concern me. And at the that very same moment you will lose.

This really comes down to the point that multibyte, unicode, and locale
should be the defaults. People generally object to this on performance
grounds, but that reminds me of the saying, "I can make this code
arbitrarily fast if it doesn't have to give the right answer." Nor do I
believe these claims, btw.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Peter Eisentraut <peter_e@gmx.net> writes:

This really comes down to the point that multibyte, unicode, and locale
should be the defaults. People generally object to this on performance
grounds, but that reminds me of the saying, "I can make this code
arbitrarily fast if it doesn't have to give the right answer."

A fair point ...

Nor do I believe these claims, btw.

Has anyone done any measurements of performance with/without these
options? I know I haven't.

regards, tom lane

On Wed, 2001-09-26 at 06:22, Tom Lane wrote:

This really comes down to the point that multibyte, unicode, and locale
should be the defaults. People generally object to this on performance
grounds, but that reminds me of the saying, "I can make this code
arbitrarily fast if it doesn't have to give the right answer."

A fair point ...

Nor do I believe these claims, btw.

Has anyone done any measurements of performance with/without these
options? I know I haven't.

Would be rather silly in my case because all my base are belong to
French locale...

=:-D

Cheers

Tony Grant

i'm new to multibyte support

i'm using unicode

i'm wondering if i have to rethink the size of string fields (char & varchar)
to allow for other languages - like is there some way of figuring out if the
fields should be larger?

any ideas welcome

murray

Bruce Momjian wrote: