On Jun 19, 2006, at 12:36 PM, Martin Vermeer wrote:
On Mon, Jun 19, 2006 at 10:44:27AM -0400, Bennett Helm wrote:
On Jun 18, 2006, at 5:24 AM, Martin Vermeer wrote:
...
Here's the patch you've been looking for... it simplifies things
which
is always a good sign.
This is much faster on Mac than the previous patches, but still a bit
slower than unpatched. I suspect the reason is that as I type, the
width of the inset is sometimes changing by small amounts, and so
large parts of the screen then have to be redrawn.
Not when you're in "wide" mode. Then the width of the text area you're
typing in should be strictly constant. Larger-than-single-row redraws
only happen if your paragraph needs to be re-broken.
(I'm not sure what "wide" mode is....)
(This doesn't
happen all the time, and I haven't been able to pin down when it does
happen, though the "jumping" screen effect is quite distracting.)
Could you describe this effect more precisely?
Adding characters to a line causes the line to increase or decrease
in visible width on screen. When in a single (non-nested) inset, this
causes the text in all lines in the inset to shift right or left a
fraction, though the borders of the inset box remain the same.
When there are nested insets, things are a bit different. Typing in a
line causes the width of all lines in the inset to shift right or
left, *and* the right border of the inset box you're typing in also
shifts right or left a bit along with the text width. (Higher-order
inset borders are not affected, however.)
When I say this doesn't happen all the time, what I mean is it
doesn't happen with every key press, it doesn't happen with every
word break, it doesn't seem to happen when LyX needs to wrap a word
onto the next line, etc. But it does happen pretty consistently when
typing in an inset -- many times over the course of typing a sentence.
Nonetheless, it's a nice improvement (and the minor trade-off in
speed for the nicer GUI is worth it, I think).
I am surprised you would see a speed loss, even a small one. In my
understanding there shouldn't be.
As I indicated, my guess is that this is a side effect of the
"jumping" effect.
Bennett
