Terms of service encoding problem

SlashedZero

Hi folks, I'm reporting this issue in a separated discussion as Max_B suggested.

I've just found another encoding problem which was not reported yet, as far as I know. The terms of service don't display international characters correctly in latin-1. However this time they aren't replaced by a "?", but by other characters combinations. For example, "é" are replaced by "é".

This is another issue, the term of service link open a new window to display termofservices.php. This file does not include header.php nor any meta element so the encoding is left to the browser, wich default most often to latin1, while the text is utf8 encoded.

akc

Just upgraded to 1.1.2, and happened to see the localized TOS just now.

Problem I see:

xml:lang is incorrectly stated as "en-ca" rather than according to the language file (zh-tw in my case), leading to gibberish (and requiring manual browser adjustment).

[ed]The gibberish is due to the browser risking a guess.[/ed]

jakob_r

I don't think xml:lang will change the character encoding. For that you need to specify the charset. See this w3c article on charsets.

If you write your foreign characters as html-entities, e.g.

o = ö

then it should work regardless.

Stash

Surely the xml:lang should be controlled by the language pack?

SlashedZero

I ran into this problem again after upgrading to 1.1.2.

My solution was quick and dirty: I replaced all non-latin characters in the terms of service definition by their HTML code.

Max_B

The best solution IMO is to add the following:// PROPERLY ENCODE THE CONTENT (core patch) header ('content-type: text/html; charset='.$Configuration['CHARSET']);in the /termofservice.php file, just after the language inclusion lines.

I know, this is a core patch you'll need to followup until it's added to the core…

akc

There are at least 2 problems here: (1) the incorrect xml:lang meta tag (apparent solution: rely on the language pack's lang definition); (2) the incorrect charset encoding (apparent solution: rely on the language pack's lang definition). Regarding using HTML entities: these are okay for some European languages with only a few "funny" characters, but for other languages (say, Vietnamese) a massive dose of entities makes docs difficult if not impossible to read (for developers or casual editors, that is).