Toxic Text Crashes MC 2.5
David Epstein
dfepstein at gmail.com
Thu May 19 09:31:01 EDT 2005
One topic of great interest that I have not seen much discussed on
this list or the Revolution list is this:
The conditions under which MC "unexpectedly quits", and how to detect
and prevent those conditions in advance.
Here is a case.
Field 1 is a background field with different content on each card, and
on most cards all works as expected. But on one card the particular
content of that field is causing crashes. The contents are displayed
properly, but any attempt to "get the htmlText of field 1" causes an
unexpected quit of MC 2.5, both in Mac OSX 10.2 and in Windows XP. I
was able to copy a chunk of the text, but then had a crash when I
tried to delete the copy from the field where I had pasted it. Trying
to "get the rtfText of field 1" also causes a crash. I made a copy of
the whole stack, and the problem card of the copied stack showed the
same symptoms.
In Windows XP, the dialog reporting the crash says:
The instruction at "0x0044d8ef" referenced memory at "0x00000000".
The memory could not be "read".
Now field 1 of the problem card has various suspect elements: some
chunks styled "link" with assigned linkText; some chunks displayed in
color; some runs of hard spaces. But whatever is causing a flaw in
the htmlText is not preventing MC from displaying it. Selecting and
backspacing to try to delete part of the field (my attempt to narrow
down the source of the problem) causes a crash, and it's not clear
whether that shows a problem with the part I selected.
I could of course simply delete the card that is behaving this way,
but I would like to make some diagnosis of the problem so that I can
be confident that my various editing scripts that rely on getting the
htmlText will not crash my program in the future. Since I can't
inspect the htmlText in question, I can't really know what's wrong
with it.
Anyone have any related experiences, insights, or suggestions of how
to proceed? Thanks very much.
David Epstein
More information about the metacard
mailing list