Re: lynx-dev Re: [BUG] Editing in unicode

[Thomas Dickey]

On Tue, Jul 01, 2003 at 08:45:20AM -0700, Ilya Zakharevich wrote:
> On Mon, Jun 30, 2003 at 01:54:53PM -0700, Ilya Zakharevich wrote:
> > While Thomas writes that this is "how it should be expected", I'm not
> > convinced.
> > 
> >   a) Did you note that the display of the edit field is wrong, but the
> >      "actual contents" (e.g., what the form sends) is correct?
> > 
> >   b) While pasting into a text input field shows junk, pasting into a
> >      the input field for `g' works fine (try it on the google site).
> > 
> >   c) How would you explain that ###s appear only on a "cross-page" pasting?
> 
> Now Thomas says:
> 
>     some paths through the program will preserve the bytes you type in, some
>     will check or otherwise process the data (and corrupt it).
> 
> So it is not a problem of interaction with ncurses, as you insisted before?

no (disagree).  The problem is that lynx reads characters from the keyboard
one byte at a time, and (except for function-keys) does not recognize that
some byte-sequences should be combined into a wide-character.
 
> Moreover, I'm discussing *exactly the same* path through program: one
> opens a google search page, and xterm-pastes an entry into the field.
> If the entry was xterm-selected on "this" page, one gets one
> behaviour; it was selected on another one, one gets another.

and you think you're selecting exactly the same set of characters from the
screen - perhaps not.
 
> And to Henry: the page is shown without any problem when a different
> display-charset is present in xterm and lynx.

sure: when using UTF-8, your select/paste behavior changes.

-- 
Thomas E. Dickey <address@hidden> <address@hidden>
http://dickey.his.com
ftp://dickey.his.com

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden