On Mon, 2006-01-23 at 09:34 +0100, Jean-Marc Lasgouttes wrote: > >>>>> "Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes: > > Martin> A patch for this is attached. An unbelievable story really... > Martin> when doing cursor up/down, the new y co-ordinate is just _one > Martin> pixel_ into the next/previous row! Therefore the test for > Martin> being inside a math inset will only trigger for the hull > Martin> inset, not for anything inside there like an array inset... > > So you mean the cursor enters the hull inset, but not the inner ones?
Yes. > I do not like much your use of arbitrary constants like 12... > > Isn't this a case where we should invoke bruteFind with limited to > next/previous row? If we had a properly functioning and acceptably fast bruteFind, yes, perhaps. Although even then I would maintain that the constant 1 in editXY is too conservative. 12 is not an arbitrary constant BTW. It is the width of a top/bottom margin in display math ;-) - Martin
signature.asc
Description: This is a digitally signed message part
