Re: [Phpwiki-talk] [phpwiki - Help] additional charset

[Reini U. <ru...@x-...>]

Mici Maci schrieb:
> I tried to set the CHARSET in the config.ini to
>  1) utf-8 
>  2) iso-8859-2
> The 1) gone wrong because of a failed sql select.
> [There's a mysql in the background.] the sql query
> string should not contain utf encoded characters.  

of course you have to tell your database to accept utf-8.

ALTER TABLE page DEFAULT CHARACTER SET utf8_general_ci;
ALTER TABLE version DEFAULT CHARACTER SET utf8_general_ci;
ALTER TABLE session DEFAULT CHARACTER SET utf8_general_ci;

what we just do is telling the client talking to
the database which charset we use for the transport.

config.ini: CHARSET=UTF-8
=>
PearDB_mysql.php:
SET NAMES 'utf8';

http://dev.mysql.com/doc/refman/5.0/en/charset-connection.html

> The number 2) was wrong because htmlspecialchars...  So I
> wrote a htmlspecialchars_workaround() function, that
> checks the third argument (which is the requested charset),
> and if it's 8859-2, it makes a few simple string replacement.
> [<, > and & is enough to replace to "&lt;", "&gt;" and
> "&amp;"...]
> It calls the original htmlspecialchars in other cases.
> Now I can use iso-8859-2 in the config.ini, and Wiki
> seems to work.  Of course I have had to append
> _workaround to all occurances os htmlspecialchars
> in the original code.  Brrrr!  Ugly and impossible to
> maintenance but maybe works...  It's under testing
> now.  8-)

Well, why not. "_workaround" might be a bit too long though.

> Why don't use PhpWiki developers their own html
> escaper function??  htmlspecialchars() has more
> disadvatages than advantages, because it keeps
> the usable character set small, so excludes lots
> of people from lots of nations to join the project...

because you are the first with such an exotic eastern charset (czech and 
polish also) which is not supported by such a basic function.
that tells us, that we seriously lack a lot of users.

of course we accept any workaround function.

> Reini Urban <ru...@x-...> írta:
> 
>> Reini Urban schrieb:
>>> Mici Maci schrieb:
>>>> thanks for your answer but it's not so easy.  PhpWiki
>>>> uses htmlspecialchars() to masq strings, and PhpWiki
>>>> passes a third argument to it containing the charset
>>>> name.  htmlspecialchars() does not know about latin2,
>>>> it gives warnings.
>>>>
>>>> lib/XmlElement.php...: 502: Warning: htmlspecialchars():
>>>> charset `iso-8859-2' not supported, assuming iso-8859-1
>>>>
>>>> lib/XmlElement.php...: 170: Warning: htmlspecialchars():
>>>> charset `iso-8859-2' not supported, assuming iso-8859-1
>>>>
>>>> These two lines (502 and 170) are giving the message
>>>> but lots of times in a page request.  What's wrong?  May
>>>> I upgrade or reconfigure php to work, or the
>>>> htmlspecialchars() function needs a workaround in
>>>> PhpWiki?
>>>>
>>>> [phpwiki-1.3.12p3]
>>>> [php-4.4.2-1.1]
>>> Oje!
>>> http://www.mantisbt.org/mantis/bug_view_page.php?bug_id=6624
>>>
>>> So you have to convert to UTF-8 and bite into the bullet.
>>> Note, that UTF-8 is not officially supported and not
> included.
>>> I posted a patch to the sf.net patch area, which should
> get around most 
>>> issues. But since UTF-8 strings are asciiz-safe it is
> not so harmful to 
>>> work without proper UTF-8 support in all string
> functions. (preg_match, 
>>> strlen, ...)
>> BTW: If you do have mbstring loaded in your php it's very
> easy.
>> Simply set
>>    mb_internal_encoding('UTF-8');
>> somewhere ("lib/config.ini") and set
>>
>> mbstring.func_overload in the php.ini or .htaccess to 7
>>
>> .htaccess:
>> php_value mbstring.func_overload 7
>>
>> The hugepatch comes from the fact, that not all servers
> support mbstring 
>> and strlen's are different in UTF-8 land.
>> We do support now, input from UTF-8 land (e.g. MS IE), but
> convert this 
>> input to our charset.
-- 
Reini Urban
http://phpwiki.org/  http://murbreak.at/
http://helsinki.at/  http://spacemovie.mur.at/