Filenames and extended characters corrupt

[Geddeth]

We have a Netware 6 server running the company's intranet, but it fails to display filenames with extended characters (from e.g. codepage 865) correctly. I have experienced this on the following occasions:

- When calling PHP 4.2.4's readdir() function to list a directory in which the filenames contain extended characters.
- When using Novell Web Search to index a filesystem in which the filenames contain extended characters, and a search is performed in that index. The summary of the contents of a file is displayed correctly, but the link to it is incorrect.

When viewing those same files thru a mapped network drive on a Windows client, the filenames are displayed fine. I have tried many different things, among these were converting everything to Unicode (HTML, webserver, PHP and Netware), but the browsers displayed things incorrectly. I have read the documentation on code pages and how to change them, but without learning anything new.

Normally, I would not hesitate to use ISO-8859-1 for everything, but I am unsure of what to set for the Netware server in order to make it display and send extended characters properly.

I have tried copying the LC_850.045 and LC_865.045 setups from the NW6 CD (TID2933194), but without any result.

Does anyone have any ideas? I specifically need the Danish vowels to be displayed properly. Right now, they are just shown as squares in the log files and in the browser.

Regards,
/geddeth

 

[michigan]

Hi Geddeth -

Haven't had that issue before, but this may help:

http://support.novell.com/cgi-bin/search/searchtid.cgi?/10011720.htm

In addition, if that link/TID isn't helpful, try search Novell's site for "extended characters" -- there's quite a few TIDs out there that may be helpful for you.

Good Luck.

 

[Geddeth]

I still have not found a solution to this annoying problem. I have attempted to isolate the problem, but without any luck really.

The simplest way for me to reproduce it, is to create a file (that Apache 1.3.27 can access) that is named with extended characters (i.e. Danish vowels), and view it through Apache's directory index from a Windows client's browser.

The listing will then show garbage instead of those characters, and the status bar something similar when I hover the filename.

I have inserted the relevant DOS statements in CONFIG.SYS and AUTOEXEC.BAT (just like the old days) and loaded codepage 850 all the way (Western Europe).

I have also inserted the lines

  LOAD CHARSET Cp850 STAY
  LOAD KEYB DENMARK

in AUTOEXEC.NCF, and can type international characters both at the console and thru a console connection to the server.

In Apache's configuration files ('httpd.conf' and 'adminserv.conf'), I have the following options set:

  AddDefaultCharset     On
  AddLanguage da        .dk
  AddCharset ISO-8859-1 .iso8859-1
  AddCharset CP850      .cp850

However .. when I run the 'LANGUAGE' command from the NW server, I get the following output:

  Current DOS code page is (865).

I'm really at a loss here, so any input is helpful.

Regards,
/geddeth