[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
MacIvory HyperCard window bugs
The transfer between HyperCard and Symbolics routines called from HyperCard
has timing problems and has not been shaken out well yet.
When using a HyperCard script to hide the Card and then call a graphics routine
inside a Frameup window remotely, sometimes the system will transfer over to
the Frameup window (covering the Card), and THEN the "clear" comes through--
resulting in a white card-sized hole in the middle of the Frameup window.
This seems to happen when the Card is over a scroll-bar in a window.
Workaround: Reposition the Card so it isn't over a scroll-bar.
When working inside an autonomous Frameup graphics window, if you "pull-down"
the Apple menu to jump to an Apple application (e.g., the Finder) it does not
kill the mouse right away. As a result, if a presentation was under the
pulled-down Apple menu, that presentation gets highlighted. The highlight
rectangle stays around, even if you go back to the Symbolics, clear the screen,
and display more graphics. You have to click once on an unoccupied section
of the Frameup window for the highlight to go away.
It is possible for a program that works perfectly well under Symbolics
to completely crash the machine when called remotely from HyperCard.
Details are still out; the best guess is that the program is reading
the *terminal-io* stream variable and then attempting to write to that
stream, thus hanging both the Symbolics and the MacIntosh.
In general, the MacIvory appears to be a fragile machine, especially on the
MacIntosh side, and especially when debugging calls from HyperCard to the
Symbolics. The MacIntosh should have an Abort key to kill the currently
running program without having to reboot. I believe the previously-mentioned
crash also freezes the Programmer's Reboot button, so that I had to power-
cycle the machine to get it back. Since my Lisp world takes 15 min to load,
this is no way to do debugging. Both the Symbolics side and the Mac side
should have methods for gracefully killing or forgetting undesired dropped
communication links. This kind of thing seems to happen quite often
during program development.
FYI. Respectfully, John Myers~~