Re: [Lynx-dev] Re: Lynx color flaws

[Thomas Dickey]

On Fri, Mar 30, 2007 at 05:37:36PM +0200, Rado S wrote:
> =- Thomas Dickey wrote on Wed 28.Mar'07 at 20:51:35 -0400 -=
> 
> > > What I just noticed:
> > > When the TEXTAREA goes wild, then following INPUT colors are
> > > broken as per the "formfail" case reported above.
> > 
> > I see that the input colors aren't reset as I move the cursor
> > through it.
> 
> "reset"? What did you expect? That after initially being A-color
> they'd revert to their originally correct INPUT-color?
> Or do you mean the wrong TEXTAREA doesn't revert?

The former.  Often repainting will fix things.
 
> Neither does. ;)
> 
> However, when radio-buttons are unselected, it goes "normal"
> color. When then the cursor passes, the initial color is applied
> again (happens for both correct INPUT- and wrong A-color).

ok.  Link-highlighting is in a different part of the code than radio buttons.
(I've done a fair amount of work on link-highlighting).
 
> BTW, the TEXTAREA seems to break always at line 64 here.

64 sounds familiar - I agree there's probably a hardcoded limit.

> For you, too, or at some other line?
> Maybe there is some hardcoded limit? Or not enough memory reserved?
> Maybe lynx wasn't designed to work with > 80x50 terminals.

Actually the only limit it's supposed to have is on line-length (because
of the way lines are stored).  There are also some limits on table cell
size to prevent too-large allocations (though I seem to recall setting
those much higher than 64).
 
> > > > > When you re-check them, then you can reproduce _all_ of them?
> > 
> > ok...
> 
> So all wrong colors are confirmed?

It seemed that way - I may find some detail that looks different on review.
 
> > Colored regions on the screen are at least something that one
> > can see to work on.
> 
> Fine, happy to hear that, and looking forward for the inconsistent
> TR coloring to be fixed, but even more the IMG+INPUT color issues. :)
> 
> > well I'll see what I can do (am swamped til the weekend, but want
> > to make sure I can reproduce things).
> 
> Thanks, confirmation is always welcome to see where the error lies.

yes - it's also good to have a set of files demonstrating the problem...

thanks

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net