Le 17/07/2015 15:01, Scott Kostyshak a écrit :
We have to understand what is the difference between these two insets.
I see. I was thinking about if from a different way. I was thinking
that there is a method that is not called when entering an inset from
the keyboard. I see though that we need to understand the root of the
problem.
In the meantime, I decided that you had the right idea :) My reasoning
is that moving the cursor always resets cursor font, except when
entering an inset. This at least shall be fixed.
I have followed the code path and now believe that it may be less
intrusive to add the call to setCurrentFont in
Text::checkAndActivateInset just after the call to edit. And if we do
that we should look at the Visible cursor movement versions.
I have to admit that at this point I am in theoretical debugging code. I
have not applied any patch nor launched a debugger. I may have to
actually test your patch and see what are the effects you describe. But
changing Text::checkAndActivateInset should have less (not enough?) impact.
Should I try different layout parameters on the Chunk inset to see
which one changes the behavior?
Let's wait until the problem which is obviously char-forward-related
gets sorted out.
JMarc
