Topic: UTF8 URL problem when used with RewriteRule

[jamie@acclaim]

Hi All

We're trying to allow URLs to be in the format:

http://sitename/article/I_Am…_Sasha_Fierce

NOTE: the ellipse (…) in the above example is a UTF8 char (0xE280A6).

To do this we're using the following RewriteRule command:

RewriteRule ^/article/(.*) /cgi-bin/display.cgi?articlename=$1 [T=application/x-httpd-cgi,E=envarticle:$1,PT,L]

When we getenv("QUERY_STRING") the value of the ellipse is E23FA6 (we expected E280A6, 3F is ? (question mark)). So as an alternate method we tried creating our own variable called envarticle and stored the articlename in there (E=envarticle:$1). But this time when we getenv("envarticle") the result is a latin1 char 0x85.

If we don't use the redirect and enter the url directly:

http://sitename/article/display?action=&articlename=I_Am…_Sasha_Fierce

then genenv("QUERY_STRING") has the correct value "action=&articlename=I_Am%E2%80%A6_Sasha_Fierce" (percent encoded, I'm assuming the browser must have done it).

The question: Is there a config setting in apache that doesn't alter UTF8 chars when used with the RewriteRule?? or any workarounds?? or are we missing something?

We're running apache 2.2 on Windows Web Server 2003, cgi program is written in visual c v6.

Appreciate any help on this one!!!

Thanks

Jamie

[jamie@acclaim]

For anyone interested, we recently revisited the problem and resolved it.

The solution:

We installed Apache 2.2.17 and used the B flag in the RewriteRule to stop apache from modifying the URL.

The B flag was introduced in Apache 2.2.12.

[Steffen]

Thanks for letting us know.


Steffen