[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Timing bug in C-V, M-V



    Date:  7 JUL 1980 1531-EDT
    From: GLS at MIT-AI (Guy L. Steele, Jr.)
    The C-V (Next Screen) and M-V (Previous Screen) commands have a severe bug.
    If you type them ahead, then they perform relative to the last screen display
    rather than to what the screen would be after the preceding command.
I have fixed C-V and friends to call the correct (new) primitive for first making sure point is
on the screen the way redisplay would first.

    I would surmise that ZWEI is checking for typeahead for immediate eecution
    before updating its idea of what ought to be on the screen.  This idea
    of where the screen-window into the buffer is within the buffer should be
    updated even though none of it has a chance to actually be put up on the
    screen.  That is, ZWEI needs two ideas of what is going on--what IS on
    the screen, and what OUGHT to be (but is not necessarily)--and the second
    should be updated with higher priority than processing typeahead if C-V
    is to work properly without timing screws like this one.
Maintenance of that much information would be much too costly, and mostly useless.  In this
case, all that is important is where the window starts.