26/06/2013 20:14, Hashini Senaratne:
As this also did not work for me I did not try the following.
Following two are the most recent commits I have done and you can also see
the output after the final commit.

http://git.lyx.org/?p=gsoc.git;a=commitdiff;h=b673a5b0a650d19383ec836c693e72af3c432d71
http://git.lyx.org/?p=gsoc.git;a=commitdiff;h=refs/heads/scroll/testing

What I can see in latest commit is that the drawParagraph debug message does not trigger when moving cursor in text. So actually no redrawing takes place.

There is some code to avoid redraws in BufferView::draw.  If you add
  d->update_strategy_ = FullScreenUpdate;
just before the switch, then your debug code should start to work, but of course this is not what we want to do.

However, we could imagine at this point to compute the cursor position and see whether it is visible with the current offset offset of the current row. In this case, we update the offset and set Update strategy to SingleParUpdate if it was NoScreenUpdate.

Does this make sense?

I think that I can stop worrying about that because there is no scrolling
for text sections.

Note that on somewhat unrelated rewrite of the metrics/display code I am trying to do, I removed the "functionality" that splits a word larger than the screen. While this probably does not happen in practice (unless it is useful in Chinese??), this may change the situation. But I may undo this particular change.

I have a problem regarding the outcome.
The current partial solution for Math has been implemented in such a way;
when the cursor moves back from the right most edge in a too wide inset, the
cursor remains at the right edge and the row slides right. Also if we select
a left cell next to the cursor is in, that also make sliding to right. Are
we going to follow the same pattern for this implementation?
I think that makes editing a too wide row, hard.

This is why I proposed that the row (or the cursor) remembers the current offset and only changes it if is is not suitable. See one of my previous mails.

JMarc

Reply via email to