Re[2]: [PHP-DEV] Bug #6173: urlencode() doesn't respect locale From: Michael Bravo (mbravo <email protected>)
Date: 08/15/00

Hello Joey,

Tuesday, August 15, 2000, 8:33:54 PM, you wrote:

JS> Hmm...
JS> I'm not sure that this is a bug.

JS> RFC 1738 says:
JS> URLs are written only with the graphic printable
JS> characters of the US-ASCII coded character set.

JS> I realize that urlencode() is not RFC 1738 directly,
JS> but still...

You see, the problem is twofold. First - it's highly practical. I've
encountered it in phpMyAdmin. Problem being, that it uses redirection
and GET method to transmit string values between frames and screens.
Therefore, significant string values get passed through the URL.
Because of several codepage per language situation, about half of all
Apache-based webservers in Russia (and not only in Russia), use
so-called Russian apache which translates content between codepages
based on server and client charset settings. I won't get any deeper so
as not to confuse people not really interested, but if I'm not able to
leave characters in <a href> unencoded, I get significant problems.
Second - I do not consider myself a guru, but looking at the actual
implementation of php3_urlencode(), I feel it should be respecting
locale settings, but it doesn't.

I would think that in case no actual immediate error is found in the
implementation, creating a separate urlencode function for nonUS
people would be very nice (locale_urlencode() or whatever).

For the time being, I can manage by replacing a call to urlencode()
with a single-line regex call. 

-- 
Best regards,
 Michael                            mailto:mbravo <email protected>

-- 
PHP Development Mailing List <http://www.php.net/>
To unsubscribe, e-mail: php-dev-unsubscribe <email protected>
For additional commands, e-mail: php-dev-help <email protected>
To contact the list administrators, e-mail: php-list-admin <email protected>