Re: lynx-dev partial display (was: on caching)

[Leonid Pauzner]

> I am working on changing HText_trimHightext() so that it can deal correctly
good!

> with being called several times, and will submit a patch.  But it occured
> to me that there may be a much simper solution to the problems with form
> fields: What happens if you just don't do the
>             highlight(OFF, (nlinks - 1), target);
> in display_page() while being in incremental rendering mode?  Could you

I haven't tried this sort of things because I was not understanding
the logic of display_page()/HText_TrimHightext(),
there are two kind of hightext etc... Just appears that display_page
may be called several times but HText_endAppend only once
(it may be ignored in most cases but made a bad thing for forms input).

To see the forms input problem: try comment out HText_trimHightext()
in HText_pageDisplay(), start lynx with -debug_partial,
set MESSGSECS to something about 5 and look any forms page,
I usually try new 'O'ptions Menu.

The problem from multiple running of HText_trimHightext()
leads to overtrimming currently selected hightext link
in <a href=...>hightext</a> for normal pages,
try accessing lynx-dev month index at www.flora.org
with removed detected_forms_input_partial check in HText_pageDisplay().

Well, it was an empiric but more-or-less works (proper changes wellcome!).

> please try that?  (You know best which flags to check for determine

All necessary flags used in HText_pageDisplay(), they are
display_partial (TRUE for getfile cycle but
became FALSE when we about finish the case NORMAL of mainloop)
and detected_forms_input_partial, you know what this means.

You may probably want to know when HText_pageDisplay called
the very first time, it is NumOfLines_partial == 0
but you should swap two commands in HTFormat.c/HTDisplayPartial() first.

> whether the function is called during partial display.)  Then calling
> HText_trimHightext() may not be necessary at all.

> You may also want to break out of the two major loops in display_page()
> when text->last_line has been reached, because that is the line still
> being filled (only during partial display).

>     Klaus