LYNX-DEV Re: RP colors - I have seen more of them

[Klaus Weide]

On Mon, 28 Apr 1997, Rob Partington wrote:

> Klaus Weide wrote:
> 
> > the (n)curses version.  I also didn't see any non-default (i.e. non-black)
> > backgrounds, but that may be because the .lss entries (I used the ones
> > that were posted here by Rob) didn't use them for any of the tags
> > occurring on the few pages I looked at.  (I changed the parsing of the
> > .lss entries to recognize ',' as well as ':' as separator between fg and
> > bg color, since Rob's code and comments seemed inconsistent; maybe I did
> > something wrong there.)
>
> i don't use backgrounds because i couldn't get them working under ncurses
> (hence the comment in the code).  if you're talking about "brightblue,white"
> in LYStyle.c, that's a typo - it should be "brightblue:white".  i've got
> no intention (right now) of changing the format of the lines because i'm
> working on reading real CSS files rather than dodgy LSS files.

I have updated the code, there is now a new version which I call 
2.7.1RP+wb+ct-0.3 in <http://sol.slcc.edu/lynx/klaus/merged/RP/>.
The only change I made to your code was to take out the recognition of ','
as separator.  Amazingly enough, I *do* see background colors now (so my
temporary "improvement" maybe wasn't so good after all..)

This is with a "linux" TERM (console) running lynx on solaris.  The ncurses 
is whatever latest version Tom has installed on SOL.

There are the usual problems with non-default bg colors in text:  it fills
neighboring empty screen regions.  But the different colors for the statusline
(depending on whether it shows an alert, a prompt, etc.) seems to work nicely,
including colors with non-default backgrounds.

> > With colors working, there is still the problem that colors are not
> > "stable" - for example, look at a section (paragraph) that starts within
> 
> i know about that - that's been there since the very first set of code.
> it's not really been on the priority list of things to fix.  i'll have
> another look sometime soon.

But I think it is important to fix...

> > I saw "Refresh screen" ^L toggle the color of some text between magenta
> > and brown in a (apparently) non-deterministic manner.
> 
> example page and stylesheet?

It seems to be related to the next point.  It is the colors that shouldn't
be there anyway, that somehow leak over from a previously viewed page,
which cannot decide what they want.

The stylesheet is the one you provided (and which I have packaged up as 
styles/lynx.lss).

Ok, here's an example: 
using the 2.7.1RP+wb+ct-0.3 code, stty rows 25 cols 80.
I have cd'ed to the Lynx code's top directory.  I start up lynx with 
http://www.shareware.com, then NEXT_LINK three times until I am on that 
link which is shown in ungly gray-on-blue (the style for TT).  
I type 'g' and enter '.'.  I see the file listing (DIRED_SUPPORT and 
LONG_LIST are enabled).  The Title line is magenta, anchors green, rest
of text also green.  
NEXT_LINK down until on CHANGES.new, ACTIVATE, Everything is shown in 
gray-on-blue...
NEXT_PAGE once, now everything is shown in green.
NEXT_PAGE again, now everything is shown in red.
Next page is in green, the one after it, too, then aafter a while it seems
to recover and from now on the text will be in normal white-on-black,
until I view some HTML page again...

I have given all these details because the slightet deviation may change
those colors..

> > the (one) color used on non-HTML pages seems to inherit some (which?)
> > color used on a previously viewed HTML page, and so depends on what one
> > has viewed before.
> 
> ok, that's easily fixed.

Care to tell us what the fix is, if it's easy, so we won't have to wait for
Lynx Styles-0.3? :)

   Klaus

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;