On Fri, Oct 07, 2005 at 04:04:10PM +0200, Jean-Marc Lasgouttes wrote: > >>>>> "Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes: > > Martin> On Fri, 2005-10-07 at 15:30 +0200, Jean-Marc Lasgouttes wrote: > >> >>>>> "Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes: > >> > Martin> You're in good company. But your patch works... for page down. > Martin> Apparently it should be extended for cursorPrevious ;-) > >> Yes, I mentioned that in my message. Should I apply it as-is or > >> try to use metrics() instead of update()? Do we care? > > Martin> Which is best for understandability? > > It is more a matter of speed, the drawing phase is unwanted. I guess > we do not really care for now. Actually, dEPM could do it in a finer > way by removing a paragraph and updating the cached values for all > paragraphs after it (it is just a vertical translation). > > Here is the combined patch I came up with. > > JMarc >
OK, at least part of this patch was tested by me and found functional. Saw no side effects (and couldn't really imagine any). So, I would say, put it in. - Martin
pgpyKNI2hB5B5.pgp
Description: PGP signature
