In message <585d1a4b61...@timil.com>
Tim Hill <t...@timil.com> wrote:
> In article <mpro.q8dh4m04f6znk02bd.li...@stevefryatt.org.uk>, Steve Fryatt
> <li...@stevefryatt.org.uk> wrote:
>
> > The issue is that NetSurf's core has to render any frame furniture when a
> > page requests it be drawn, and whilst it defers this to the GUI (IIRC),
> > it's fairly non-trivial for the RISC OS front-end to use the standard
> > desktop furniture.
>
> > It's been a long time, but (again IIRC) I'm fairly sure that I concluded
> > when I looked at this that the only way to get "standard" scrollbars
> > would be to replicate the Wimp's rendering of the component bits within
> > NetSurf's RISC OS front-end -- which, aside from being relatively
> > complex, also took us into areas best described as "sparsely documented"
> > and hence fragile if the OS developers change the way things work.
>
> I just had a quick peek inside the RISC OS Wikipedia page. There is no
> FRAME or IFRAME but it has the extra scrollbar. Its stylesheet defines the
> BODY with 'overflow-y:scroll'. Any browser with window furniture already in
> place should have that set on BODY by default, shouldn't it, and ignore if
> it's set there again? Just a thought.
Interesting. Thanks for investigating the cause, Tim.
I just visited the RISC OS Wikipedia page with NetSurf and did a Full
Save, then looked for occurrences of "overflow-y". There is just one.
I set its value to "visible". Double-clicking the app then rendered
the page with only the normal RISC OS scroll bars, which of course
allowed quick and smooth scrolling in the way one would expect.
So perhaps we should turn this on its head and ask if there is any
situation where this result would not be the best.
David
