[9116] in Athena Bugs
Re: [Roger A. Roach : Multiple Bug Lists]
daemon@ATHENA.MIT.EDU (Mark Curby)
Mon Mar 23 17:57:30 1992
Date: Mon, 23 Mar 92 17:56:58 EST
From: mlc@MIT.EDU (Mark Curby)
To: techmail@MIT.EDU, bugs@Athena.MIT.EDU, dcnsm@MIT.EDU,
Greetings -
Re: [9099] in bugs
-- 17mar92 rar bug report to bugs@athena.mit.edu
and
Re: Problem with forwarded enclosed (binhex) mail
-- rar 18mar92 to: bug-techmail
and
Re: [Roger A. Roach <rar@MIT.EDU>: Multiple Bug Lists]
-- cec 19mar92 to dcnsm, mar, mlc
and
Re: [Roger A. Roach <rar@MIT.EDU>: Multiple Bug Lists]
-- mar 20mar92 to cec
etc.
The problems:
1) It appears that a binhex enclosure in a TechMail mail message can no
longer be extracted if the mail message was forwarded by mh or xmh.
2) There are some weaknesses in our bug reporting and tracking systems,
particularly in cases where the problem involves interactions between
multiple systems managed by different groups.
Problem 1:
My understanding of the problem is that some mail clients (e.g. mh and
xmh) modify the contents of mail messages when forwarding them, rather
than simply encasing the message in
------- Forwarded Message
------- End of Forwarded Message
without additional modification. This is done with the best of
intentions, compliant with the proposed encapsulation standard RFC-934.
From my cursory reading of RFC-934 it seems to recommend that lines
beginning with "-" be converted to lines beginning with "- -". The text
of RFC-934 implies that this is done to make handling of digests
(multiple mail messages in one mail message) work right.
Unfortunately, as is often the case, the fix to one problem may break
something else. In this case it appears that if a mail message
containing a binhexed file (e.g. a message w/ an enclosure created by
TechMail 2) is received and forwarded by xmh or mh, and if furthermore
said binhexed file contains at some point a line beginning with "-",
then this will be converted to "- -" in the forwarded message. If the
forwarded message is then received by TechMail 2, and the user attempts
to extract the enclosure, TechMail will recognize that the binhex file
does not conform to the binhex specs and give an error message.
We live in an imperfect world.
I recommend that the following be added to the TechMail wish list to be
addressed as resources permit after TechMail-S 2.0 is released.
- investigate the problem to confirm that the above understanding of the
problem is accurate, and nothing else is also going on. (If you know of
flaws in the above understanding please send comments and corrections to
techmail@mit.edu).
- if there is a easy fix in TechMail which does not appear likely to
break anything else then we should add this to a future version of
TechMail. An obvious candidate is to have the de-binhexing code look
for "- -" at the beginning of a line and convert it to "-" before
proceeding with the decoding. An alternative would be, as Roger
suggests, to modify the mh and xmh forwarding commands to allow the user
some way to turn off the "-" to "- -" conversion.
Problem 2:
I think that our present bug tracking and management systems could be
improved. I believe that this would require the commitment of
significant initial and ongoing resources, on the order of that required
for the the help desk trouble tracking system, if not more. I leave as
an exercise for DCNSM where this fits in the overall priorities.
- Mark
------------
Mark Curby
MIT Network Services MIT Room E40-342G
(617) 253-7725 77 Massachusetts Ave.
mlc@mit.edu Cambridge, MA 02139
