Bug 41249 – Missing encoding declarations in the Gentoo Forums Web pages

• Product: Gentoo Infrastructure
• Status: RESOLVED FIXED

[Description Olivier Fisette (RETIRED) 2004-02-11 07:42:46 UTC]

Some forums (such as the french one) use iso-8859-15 characters but the corresponding charset=iso-8859-15 information is absent from the meta section of the HTML files. This can lead to an incorrect display of the non-us-ascii characters on some configurations.

Reproducible: Always
Steps to Reproduce:
1.
2.
3.

Actual Results:  
Here is the current meta-information taken from the french forum index : 
 
<meta http-equiv="Content-Type" content="text/html"> 

Expected Results:  
It should be : 
 
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-15">

[Comment 1 Flammie Pirinen (RETIRED) 2004-06-07 21:51:45 UTC]

I see there would be at least two problems with this approach. 

For once the meta http-equiv is a hack to get proper charset defined if server is not giving any charsets. If something should be configured it will be the server sending the encoding in the actual header data, in that way you'll also get all browsers to obey it.

The more serious problem however is that phpbb relies on absence of the charset definition and browsers' autodetection on multilingual fora. If you set the charset to ISO-8859-15 you'll bork characters 

[Comment 2 Flammie Pirinen (RETIRED) 2004-06-07 21:51:45 UTC]

I see there would be at least two problems with this approach. 

For once the meta http-equiv is a hack to get proper charset defined if server is not giving any charsets. If something should be configured it will be the server sending the encoding in the actual header data, in that way you'll also get all browsers to obey it.

The more serious problem however is that phpbb relies on absence of the charset definition and browsers' autodetection on multilingual fora. If you set the charset to ISO-8859-15 you'll bork characters ½ (1/2 vulgar fraction)  and ' (upwards accent used as apostrophe) for all US-ASCII users. On more serious problem that will bork all CJK-users as well as cyrillic, greek and so on. Should any charset be selected it had to be unicode one, either UTF-8 or UTF-16. I would suggest the former since it is well supported in all modern browsers including even NN 4.8, lynx and IE 6. 

[Comment 3 Christian Hartmann (RETIRED) 2005-09-02 00:37:31 UTC]

Will be fixed in the upcoming conversion to UTF-8. See bug #102537.

[Comment 4 Tom Knight (RETIRED) 2007-09-12 19:09:39 UTC]

Now we've converted the forums to UTF-8 I've added 'charset=UTF-8' to the meta tag.

[Comment 5 Tom Knight (RETIRED) 2007-09-12 19:12:24 UTC]

Fixed.