Webhelp created with RoboHelp HTML installed on Linux server hangs when viewed with Firefox

[miracle_worker]

[This is a repost & update of a message that I posted on the RoboHelp for HTML forum by mistake]

We have just started using the Adobe Technical Communication Suite to single-source our documentation and have worked out most of the kinks, save for one showstopper in the Webhelp generated with RoboHelp. We need our help system to work with both IE and Firefox. When the help is run locally on a Windows machine, both IE and Firefox work fine.

When the help is installed on our Linux-based server, opening the help causes Firefox to hang unless Firefox is configured to use Unicode encoding. Unfortunately, the Firefox default is Western European. We can't control our customers' environments, and we can't just document this problem because FireFox users won't be able to open the help system to read about the fix.

When investigating this problem, we discovered that RoboHelp is inserting extra characters at the beginning of some of the HTML files. These characters are ignored in IE as well as in Firefox when the help is installed locally, but when the help is installed on a Linux server, they causes Firefox to hang unless Firefox was configured to use Unicode encoding. It seems that on Linux, the characters are being converted to 3 garbage characters, which are visible if you look at the source code with Firefox.

This is a showstopper for us, unless a workaround can be found. My questions are:

1) Is this a known problem?
2) Is there a way to get RoboHelp to force Unicode encoding or to eliminate the characters altogether?
3) Is there a way to force FireFox to use Unicode encoding to display our help system (say, by changing our CSS file)?

Thanks!

[miracle_worker]

Peter Grainge posted the following reply to my original message on the RoboHelp for HTML forum:

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

A bit beyond me but I wonder if Item 22 on this page of my site helps.

http://www.grainge.org/pages/authoring/rh7/using_rh7.htm

The reference to 3 garbage characters is what prompted me to suggest this.

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

I think this may be on the right path. If so, I believe that Adobe should add a switch to RoboHelp to force HTML files to use Unicode encoding.

[HKabaker]

I would like to understand this better. I read Item 22 on Peter's site.

In all of my html output files RH inserted this line:

<meta http-equiv="content-type" content="text/html; charset=utf-8">

As I understand it, the browser uses this flag.

Somewhere -- I think it was during installation of RH 7 -- I had the option to select either the Windows or UTF character set. I selected UTF.

I can't find that setting now, but the meta line appears without any help for me.

Q 1: Where is/was that option?

Q 2: Is there yet another piece to the puzzle, mentioned in this paragraph?

"Set the 3 bytes UTF-8 BOM (byte-order mark) at the very beginning of the file to tell any kind of text file editor that it is UTF-8 encoded file. These 3 bytes are invisible characters so you won't see them in any editor. These are optional bytes and browsers don’t require them to render a UTF-8 encoded HTML file."

Elaborating, the article says some servers (e.g., Novell) need the 3-byte BOM flag or will not interpret the file as UTF-8, despite the META statement.

Q 3: Does this mean we need RH to do something more than inserting the META statement?

Q 4: Is this also essential for getting servers and/or browsers to display certain characters in languages other than English?

Q 5: Is this also relevant to Linux servers, whether for English or other languages?

I found some enlightenment here:
http://en.wikipedia.org/wiki/Byte_Order_Mark

Right now I'm not affected by this. But I'm trying to figure out what to look for in the future.

Thanks for any help.

Harvey