Hello Jean-Marc,
> +++ b/src/Cursor.cpp
> <at> <at> -278,7 +278,7 <at> <at> CursorData::CursorData(DocIterator
const & dit)
> // bv functions are not yet available!
> Cursor::Cursor(BufferView & bv)
> : CursorData(&bv.buffer()), bv_(&bv),
> - x_target_(-1), textTargetOffset_(0)
> + x_target_(-1), textTargetOffset_(0),left_edge_(0)
> {}
>
> Didn't you forget too_wide_row_?
>
> <at> <at> -538,7 +538,33 <at> <at> Row const & Cursor::textRow() const
> {
> CursorSlice const & cs = innerTextSlice();
> ParagraphMetrics const & pm = bv().parMetrics(cs.text(), cs.pit());
> - return pm.getRow(pos(), boundary());
> + bool const bndry = inTexted() ? boundary() : false;
> + return pm.getRow(cs.pos(), bndry);
> +}
> +
> +int Cursor::getLeftEdge() const
> +{
> + return left_edge_;
> +}
> +
> +Row const & Cursor::getToowideRow()
> +{
> + return *too_wide_row_;
> +}
>
> As I already said, I think this name is not good. I would propose
> current_row_ or current_bottpm_row_. THe question is not to know whether
> the row is to wide, but what is the row to which the value of left_edge
> makes reference.
What I thought from the beginning is to save the finally found actual too
wide row to this variable. So until we find a too wide row, we cannot save
something to this. That is why I have not initialized too_wide_row in the
constructor of Cursor. So in that case we have to identify a wide row in
TextMetrics.cpp and save the value from there.
That is what I have tried in this commit:
http://git.lyx.org/?p=gsoc.git;a=commitdiff;h=5ac5e4cd540eb3b9ca9e06dd14aef63d4361b37a
But unexpectedly it stops by giving a runtime error at a point we need to
slide a row. After some debugging I hope the error is within setToowideRow()
method.
As I can see, what you are expecting it save the current row to this
variable and always / just after visiting a wide row; set the left_edge to
0. I will try this approach too.
> +
> +void Cursor::setLeftEdge(int leftEdge)
> +{
> + left_edge_ = leftEdge;
> +}
> +
> +void Cursor::setLeftEdge(int leftEdge) const
> +{
> + const_cast<Cursor *>(this)->setLeftEdge(leftEdge);
> +}
>
> When you have to do something like this, then something is usually
> wrong. I think it would be better to make a cast the cursor to an
> unconstified version in drawParagraphs that to create const methods that
> are not const. This is just cheating.
Yes, I will try that too.
>
> // check if we had something else in mind, if not, this is the
> future
> // target
> - if (x_target_ == -1)
> + if (x_target_ == -1){
> setTargetX(xo);
> + left_edge_=0;
> + }
> else if (inset().asInsetText() && xo - textTargetOffset() !=
> x_target()) {
> // In text mode inside the line (not left or right)
> possibly set a new target_x,
> // but only if we are somewhere else than the previous
> target-offset.
> <at> <at> -1981,8 +2018,10 <at> <at> bool Cursor::upDownInText(bool
up, bool &
> updateNeeded)
> // update the targetX - this is here before the "return false"
> // to set a new target which can be used by InsetTexts above
> // if we cannot move up/down inside this inset anymore
> - if (x_target_ == -1)
> + if (x_target_ == -1){
> setTargetX(xo);
> + left_edge_=0;
> + }
>
> Are you sure that this one is needed? What is the use case?
I was not sure whether the text can go grow beyond the screen limits.
Because of that reason I added this here. According to the new update to the
algorithm (stated above), I think I can remove these two lines. Is not it?
Thank you
Hashini
