[www-features] Re: [IMC-Tech] www-features mir admin playing up!

[Max]

Hi,

> mod_security-message: Access denied with code 403. Error parsing
> multipart parameters: Error normalizing parameter value: Invalid character detected
> [226]
> 
> if you look at the ascii chart for character 226, you see it is weird:

Looking at an ascii chart doesn't make much sense if we use utf-8.

> The boat of  <E2><80><9C>women on waves<E2><80><9D> 

utf-8 is an encoding used to be able to show the whole (16-bit) unicode range,
without taking too much space up. The way this is achieved is roughly as
follows:

- ascii characters (7 bit, so from 0-127) are represented by a single byte
- another set of characters takes up 2 bytes
- another set of characters takes up 3 bytes
- another set of characters takes up 4 bytes

So, the <E2><80><9C> combination is a 3 byte combination. It may be that in this
case, the character in question is a rare (but valid) quote symbol or so. But
there will be similar combinations if languages such as greek or japanese are
used.

In short: the way mod_security looks at characters makes little sense for us. I
hope this feature is disabled entirely.

Greetings,
Max