Martin Vermeer wrote:
Hmmm, would the construct
cur.updateFlags(Update::FitCursor | Update::SinglePar);
work here? It would update always only the paragraph the math inset is
in, but the whole screen only if the cursor lands outside it.
- Martin
OK so I tried this... doesn't work.
The problem is that indeed then only the paragraph containing the math
inset is redrawn (as shown by debug output), including the inset -- but
somehow the corner markers are not reset. I get the impression that
without the updateFlags statement, the whole inset is hierarchically
refreshed*), but with that statement, _even including SinglePar_, no
real refresh appears to take place.
I noticed that also, yes. I think the updateFlag is reset somewhere
during the dispatch calling hierarchy. I did use some trick to make sure
that they are not reset for text insets and I think we should do the
same for math inset.
*) Even so, that moving the cursor somewhere in an math array inset will
refresh corner marker "turds" left in other cells of this inset, and at
various nesting depths -- I had for testing modified only "cursor
right", which then started dropping these turds.
What is a turds?
This is mysterious.
Not that much once you take the effort to go through the dispatch
machinery step by step. But this sure takes a great deal of time and is
awfully complicated...
Abdel.