[tlug] Character encoding on web host's MySQL server

[Dave M G <dave@example.com>]

TLUG,

One of the web hosting services I use has PHP 5.2.5 and MySQL 4.1.16.

I should mention before continuing that the site is for a group I help 
out with, and the issue of changing hosting services is beyond my 
control. So I've got to work with what I've got for now.

Naturally, I want to make a site that displays Japanese text, and my 
preferred encoding is utf-8. And the web hosting service seems to have 
utf-8 supported... mostly.

I just uploaded a new version of the web site, with all new and much 
improved PHP code, and despite all indications that I have set 
everything to use utf-8, the Japanese characters are coming up as all 
question marks.

It is not surprising that I am seeing this problem now, because my 
previous web site version made do with storing Japanese as binary data, 
because it was originally built for a MySQL 3.2 system where utf-8 
support was not very good. Now I want to get everything using more 
current, sensible, and uniform use of utf-8.

Looking in the phpMyAdmin interface for my webhost, I can see that some, 
but not all, of MySQL's system variables are set to use utf-8 for 
encoding and collation:

character set client        utf8
(Global value)            latin1
character set connection     utf8
(Global value)            latin1
character set database        latin1
character set results        utf8
(Global value)            latin1
character set server        latin1
character set system        utf8
character sets dir        /usr/share/mysql/charsets/
collation connection        utf8_unicode_ci
(Global value)            latin1_swedish_ci
collation database        latin1_swedish_ci
collation server        latin1_swedish_ci

On my home machine, where Japanese stored in MySQL works fine, I seem to 
have ensured that everything is set to use utf-8 by adding the following 
lines into my.cnf, under [mysqld]:

# utf8
init-connect='SET NAMES utf8'
character-set-server=utf8
collation-server=utf8_general_ci
skip-character-set-client-handshake

But, on this particular web host, I don't have access to modify my.cnf.

So... my questions:

On my home machine, *all* of the above listed variables are set to use 
utf-8. But, in order to get proper utf-8 functionality, do I need to 
worry about all of them, or are some more critical than others?

I have emailed the web host support with most of the information here, 
requesting that utf-8 support be made more complete. However, in 
anticipation of the possibility that they can't comply because changing 
settings would adversely affect other clients, can I somehow modify my 
local environment enough to ensure proper functionality? If so, how?

Thank you for any advice.

--
Dave M G
http://www.tlug.jp/wiki/User:Dave_M_G