MODx 0.9.6 and character encoding

I’ve some problems to see accentuated characters in the french installation page.
The addition of the following line in the intall/index.php file correct the problem.

Quote from: poof65 at Jun 05, 2007, 04:30 PM

I’ve some problems to see accentuated characters in the french installation page. The addition of the following line in the intall/index.php file correct the problem.

header('content-type: text/html; charset=utf-8');

Hmmm, I didn’t experience this problem this is strange normally the french language file for the installer has its own utf-8 header...
Can you tell me what’s in your installer language file ?

Check install/lang/french/french.inc.php

Line 66 should read

[tt]$_lang["encoding"] = ’utf-8’; //charset encoding for html header[/tt]


Quote from: wshawn at Jun 05, 2007, 04:36 PM

The tar file does not have the correct permissions for my server evidently. So a direct pull of the file and tar xzf or unzip (depending on which files) does not work under any circumstances.

I had some issues too, did it this way

[tt]tar -xvzf modx-0.9.6.tar.gz[/tt]

and got all files with
[tt]Owner : qmailq
Group : spamd[/tt]

Let me quote Jason’s explanation

Are you extracting that as root? Root user always preserves tar permissions and ownership when extracting; you can use --no-same-permissions and --no-same-owner flags to prevent the extraction from preserving these things. Non-root users have these by default, reverting to current user and the current umask

I *did* untar as root with ssh, and didn’t apply the above mentionned flags... and that’s why it messed up my permissions...
Hope it helps

The line 66 in my french.inc.php file is like yours.

The $lang[’encoding’] is used here (in header.php)

and is working fine, i have "content="text/html; charset=utf8" in the source code.

The problem is that the browsers i try (firefox and epiphany on linux) doesn’t care of meta headers in html source code but rather the http headers (in particularly the "Content-type" header) to know what encoding is used in the web page. When I look page informations, i can see that the encoding ISO-8859 is used by the browser to read the web page code.

And you, what web browser have you tried ?

Whatever your MODx character encoding is set as will be used to set the header encoding. You can access this using the tag [(modx_charset)] or via the modx config, $modx->config[’modx_charset’]. Your template should then also use this tag to set the Content-Type meta element. I am guessing your database is running in latin1 but you are using a utf8 language file ?

I’ve no problem with my modx site, i am just reporting a problem with the installation script of modx 0.9.6.

Quote from: poof65 at Jun 05, 2007, 05:21 PM

I’ve no problem with my modx site, i am just reporting a problem with the installation script of modx 0.9.6.

Oh, I see; I was misunderstanding what language files were being discussed...is there a bug report entered for this? If not, there needs to be...

@Poof65 : Could this be that the headers sent by your server is set as Latin1 by default (as are still lots of french server) ?
If so using AddCharset in .htaccess could help...

I’ve searched a little in bug listing and i don’t have see this bug.

I’m not registered as bug reporter and i don’t want to do this now.

There is a clean way to correct this in adding

under the line

or directly in lang.php

@davidm > i’ve tested the install script on my local server and on my lamp installation apache does not send encoding headers by default.
But with a little modification of the script there is a way to make it work even on misconfigured servers.

The installer simply needs to be setting the header based on the selected language encoding. It’s a simple bug with a simple fix and is not a problem with his Apache config.

Question: Do you’ve added any encoding conversion while adding UTF-8 support to 0.9.6?

I wanted to update from 0.9.5 to 0.9.6. My 0.9.5 is already running with UTF-8 and now it seems that some parts getting converted again and appear as ISO-8859-1 chars but in UTF-8 encoding.