On Mon, Feb 06, 2006 at 01:00:14PM -0500, Bennett Helm wrote:
> On Feb 5, 2006, at 8:50 AM, Martin Vermeer wrote:
> 
> >On Sat, Feb 04, 2006 at 05:28:16PM -0500, Bennett Helm wrote:
> >>On Feb 4, 2006, at 4:49 PM, Martin Vermeer wrote:
> >>
> >>>Bennett,
> >>>
> >>>did you get around to testing my speedup inside insets patch (bug
> >>>2195)?
> >>
> >>Sorry -- I tried but couldn't apply the patch: there's a conflict
> >>with (if I remember correctly) rowpainter.C, and I hadn't yet had
> >>time to investigate. I'll try to get to it....
> >
> >Try this patch instead.
> 
> OK -- that patch works. I created a document with a large inset, and  
> typing speed is much improved with the patch. (It's not as fast as  
> typing outside an inset, however.)

Yes, some overhead is expected.
 
> There are some minor drawing glitches. Once an inset becomes large  
> enough to take up the whole line, the box around it changes depending  
> on whether the cursor is inside the inset or not. If not, it's a  
> normal box, looking as you'd expect. With the cursor inside, however,  
> the horizontal lines marking the top and bottom of the box stretch  
> from the left edge to the right edge of the document window, and  
> vertical bars on left and right are non-existent. (If the box is  
> colored, as with LyX-note, the color takes up the whole window from  
> left to right.)

Actually this is by design. Without it, you would see *real* glitches...
I think it's a rather elegant solution ;-)

> Presumably related is a drawing glitch that happens with an inset  
> inside an inset (and only with the cursor inside): normally, nested  
> insets appear as you'd expect, but once the line length of the inner  
> inset equals the line length of the outer inset, it again expands to  
> take up the full width. If typing more characters in the inner inset  
> results in the word wrapping to the next line, the line length  
> normally is decreased and the box of the inner inset returns to normal.

Same remark.
 
> A speed issue as well: if there is only a single large inset on  
> screen, typing speed is pretty good. Once there are multiple large  
> insets on screen, typing in any of them is much slower, lagging the  
> keyboard.

I don't remember if this should happen or not... you could try testing
with -dbg PAINTING, which shows how many rows are being painted. 

Does it make a difference if you collapse all insets but the one you're
working in?

- Martin
 

Attachment: pgpspzXxRes8h.pgp
Description: PGP signature

Reply via email to