Subj : MultiMail reply corruption (was Re: EDIT text editor)
To   : Craig Healy
From : William McBrine
Date : Fri Nov 15 2002 11:23 pm

-=> Craig Healy wrote to Ken Hrynchuk <=-

 CH> As I've mentioned, I'm more than willing to work with the author to
 CH> try to figure what's happening.

And as I've mentioned, what I need is one of the reply packets that causes
this. There really is very little else that can help me. I appreciate that
you want to help, but unless you have a reply packet for me, I don't see
what else you can do.

At any rate, Jimmy Day did send on a non-reply packet showing the _results_
of his event, which were suggestive. I now suspect a (rare) mismatch
between the length of the actual text of a message in a reply packet, and
the length specified in the header. I'm in the process of rewriting the QWK
code to try harder to avoid this possibility. If this IS what's happenning,
it could account for a number of other -- unfortunately, vague -- bug
reports that I've gotten. However, I still haven't figured out how it
could've happened in the first place; and without seeing one of the reply
packets that caused the problem, I still don't know for sure that this IS
it.

To All:

Please, people, help me to help you. I've gotten so many frustrating bug
reports over the years -- frustrating because there's not enough detail to
figure out what's going on. Now, some people have been very good reporters.
But alas, none of those people seem to be the ones afflicted by THIS bug.

If you see anything like this, SAVE YOUR REPLY PACKET AND SEND IT TO ME.
PLEASE! I promise to keep the contents in confidence. Save it as soon as
you see strange results from the door after upload, and don't try to reopen
it in MultiMail or any other reader until after you've made a copy of it.

If you see anything like this, and you DON'T send the reply packet to me,
then please don't bother reporting it at all, because it doesn't help. And
just in general, if you're reporting ANY bug in MultiMail, I'll probably
want to see the original packet and/or the reply packet. I may not need
them, and I'm not demanding that they accompany every bug report, but you
MUST at least save them in case I do need them.

.... MultiMail: http://multimail.sf.net/
--- MultiMail/Linux v0.44
 * Origin: COMM Port OS/2 juge.com 204.89.247.1 (281) 980-9671 (1:106/2000)

.