socket connection: no utf-8 characters?

[Dr__TYPO]

Hello everyone,

I have the following problem:

- I connect to a webserver via socket connection from InDesign in order to read the website.

- the document that I read is a php page

- the content i read is utf-8 encoded

- unfortunately the utf-8 characters are truncated in my result

--> WHY? Any idea is welcome!

Here are the scripts which I'm using:

First the php script on the server:

<?php
header("Content-Type: text/html; charset=utf-8");
echo "Motörhead";
?>

Now the ExtendScript:

Result = "";
Connection = new Socket;

if (Connection.open ("localhost:80")) {
    Connection.write ("GET /test.php HTTP/1.0 \n\n");
    Result = Connection.read(999999);
    Connection.close();
    }

var myDocument = app.documents.add();
var myTextFrame = myDocument.pages.item(0).textFrames.add();
myTextFrame.geometricBounds = ["0mm", "0mm", "100mm", "100mm"];
myTextFrame.contents = Result;

And finally the result in my layout document:

HTTP/1.1 200 OK
Date: Wed, 30 Dec 2009 09:34:32 GMT
Server: Apache/2.2.12 (Win32) DAV/2 mod_ssl/2.2.12 OpenSSL/0.9.8k mod_autoindex_color PHP/5.3.0 mod_perl/2.0.4 Perl/v5.10.0
X-Powered-By: PHP/5.3.0
Content-Length: 9
Connection: close
Content-Type: text/html; charset=utf-8

Motrhead

As you can see, the character "ö" in the last line has been swallowed.

What's wrong here?

Thanks for you help in andvance,

Dr. TYPO

[Dirk Becker]

I do not know PHP, but I think there is something wrong at that side. Apparently setting the HTTP header to utf-8 plus wishful thinking does not change the applied output encoding of PHP. Besides, if you do not deliver HTML you should also use the correct MIME type text/plain.

Motörhead is 9 characters. Your reported content length is also 9 bytes, while the ö Umlaut should be encoded into at least 2 bytes.

Try the snippet below, an alternative is to ensure that your PHP source document is already utf-8 encoded.

<?php
header("Content-Type: text/plain; charset=utf-8");
echo utf8_encode("Motörhead");

?>

http://en.wikipedia.org/wiki/MIME

http://php.net/manual/en/function.utf8-encode.php

http://linux.die.net/man/1/hexdump

Dirk

[Dr__TYPO]

Hello Dirk,

thanks for the hint.

This does the trick:

in the php script the utf8 econded string must be published url encoded, otherwise in would be utf8 interpreted twice from the receiving ExtendScrip

echo urlencode(utf8_encode("Motörhead"));

and then in the ExtendScript decode the content, e.g.

alert (decodeURI(Result));

[Dirk Becker]

The second wrapper should not be required if your web site is supposed to deliver UTF8.

Instead you'd add a conversion step within ExtendScript.

E.g. use one of the JS UTF8 decoders from the web, or write the raw bytes out into a file then adjust the File.encoding variable for re-reading.

File.encoding is described in JavaScript Tools Guide, see the ESTK help menu.

Dirk

[Dirk Becker]

Ah, I always forget that Socket also has its own .encoding variable. Just make sure it is set correctly.

Dirk

[TᴀW]

https://forums.adobe.com/people/Dirk+Becker  wrote

Ah, I always forget that Socket also has its own .encoding variable. Just make sure it is set correctly.

Dirk

OMG! Thank you! Sanity restored...

Ariel