Bo Peng wrote: >> You patch seems to work but I do not really understand what is >> going on. Could you explain a bit? > > I found this notifyCursorLeaves(old, new) function and see what is > going on. If Abdel does not see any side effect of calling too many > notifyCursorLeaves, your patch has my OK. > > I, starting from the very beginning, dislike the fact that > latex_code() is called every time a cursor leave a mathed. But again, > your patch is simpler and fixes the problem. I may revert yours and > the whole notifyCursorLeaves stuff, and replace them with > inset::markDirty() after 1.5.0.
I don't know Bo, I think that latex_code() is a reasonably efficient way of having a signature of the math inset (to check if it has changed). I would be surprised if this had a real performance impact (typically O(1) operations per user interaction, we do much worse elsewhere in the code) A/