Re: Wrong charset convert SOLVED

Jiří Eichler <ejirkae@xxxxxxxxx> · Wed, 01 Jul 2009 19:19:11 +0200

Man you are incredible. Thank you. And I'm idiot because I searched for 

mistake where it wasn't :) I had just to use:

$target_path = utf8_decode($target_path);

Maybe it will be also possible change somehow php.ini, I'm going to look 

at it. I didn't think that it is problem with php, it ran on Linux well. 

And add one line to MediaWiki won't be problem.

Thank you so much for your time.
Jiri Eichler

André Warnier wrote:

Jiří Eichler wrote:
..
I just checked your on-line example.
I used Firefox 3.1, with the "HttpFox" add-on (recommended).
This shows exactly what the browser is sending to the server.

In this case, the form does a POST, in the "multipart/form-data" 

encoding.

I sent a small test file, which I created on my disk under Windows XP 

(German, so basically latin-1, not latin-2 like yours).

I used cut-and-paste from your email, to copy the filename.

The file name is thus the same as your example, but with a .txt 

extension.

In the Windows Explorer (not IE), the file name on my disk looks like :
složka.txt

(I used cut-and-paste again in the Explorer to copy this into this 

email).

I used your (nice) example form to send this file to your server, and 

traced it with HttpFox.

This is actually what the browser is sending, as part of that 

multipart POST :

-----------------------------20037128598723

Content-Disposition: form-data; name="uploadedfile"; 

filename="sloÅ¾ka.txt"

Content-Type: text/plain

test složka

-----------------------------20037128598723--

Important : the browser is NOT sending this filename as a part of the 

URL.  It is sending it in the BODY of the POST request.

It is also not sending it encoded as "slo%C5%BEka". It /is/ sending 

the filename encoded as UTF-8.

That means, that if there is a translation going on here, it is NOT at 

the level of the upload URL.

Now the question is to know how your PHP script really interprets this 

filename.  As UTF-8 ?  How do you know for sure ?

(I am not a PHP specialist, at all)
I mean, precisely :

the PHP script "uploader.php", somehow "gets" the value of the 

parameter "uploadedfile" as a string, representing the filename that 

the browser uploaded.

In which encoding (in PHP) /is/ that string ? does PHP know that this 

is Unicode/UTF-8 ?

Or does /PHP/ (which runs under Apache, which runs under Windows, on a 

Windows system where the default charset is cp-1250) think that this 

string is encoded in cp-1250 ?

And then, when PHP writes this file to the disk, it encodes the 

filename /again/ into Unicode, and thus this time the "Å¾" (which is 

originally 2 bytes representing 1 Unicode character), now becomes 4 

bytes representing the UTF-8 encoding of "Å" and "¾" ....

... and then, PHP generates the index listing.  And, in this index 

page, it generates the "href" as

<a href="slo%c4%b9%c4%beka.txt">

which looks very much like it could be the Unicode/UTF-8 encoding of 

"sloÅ¾ka.txt", but not like the Unicode/UTF-8 encoding of "složka.txt".

---------------------------------------------------------------------

The official User-To-User support forum of the Apache HTTP Server 

Project.

See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
  "   from the digest: users-digest-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
  "   from the digest: users-digest-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx