Re: lynx-dev Re: [dev.15] fixup patch: edit TEXTAREA

[Jacob Poon]

On Mon, 1 Feb 1999, Leonid Pauzner wrote:

> 1-Feb-99 12:08 Jacob Poon wrote:
> > Currently, that will mean Windows-based Unicode web pages will be in
> > trouble, because Win32 applications use UCS-2 encoding by default.
> 
> There are two different subjects: web page in Unicode
> (not UCS-2 but actually UTF-8 encoded)
> and lynx' internal representation of the loaded page
> (in "display character set" actually).
> Now you call an external editor: it is assumed that the editor
> works properly with plain text in "display character set",
> it is a reasonable assumption that your editor compatible
> with your current font (try another editor if it is not true).
> 
> What kind of trouble do you mean
> and what "Windows-based Unicode web pages" is for you?

The problem is, since most Windows programs store messages with UCS-2
encodings, if I try to capture the texts and publish them into web pages
(eg: quoting from error logs), I may be unknowingly copying UCS-2 encoded
messages instead of UTF-7/8 data, simply because the effect is invisible. 
When I save those pages from the editors and publish them, Lynx will be
unable to see them. 

In other words, because UCS-2 is handled transparently under Win32, users
may forget to convert such documents into UTF-7/8 encoding, making such
pages unviewable under Lynx.

I have not found a web page using UCS-2 encoding (yet), but I found a text
file encoded with UCS-2 encoding that is not compatible with Lynx
(2.7.2)'s UNICODE UTF-8 Display Character Set.  Under Lynx, it pads each
English letter with the annoying 'U00' word.  It is located at: 
http://www.geocities.com/TimesSquare/Battlefield/3605/srwffaq.txt

However, the text is viewed correctly under Netscape 4.x, except the
annoying double line feed.