[7569] in Athena Bugs
rt 7.2R: sendbug/comp
daemon@ATHENA.MIT.EDU (David Krikorian)
Tue May 14 23:18:12 1991
Date: Tue, 14 May 91 23:18:01 -0400
From: David Krikorian <dkk@ATHENA.MIT.EDU>
To: bugs@ATHENA.MIT.EDU
Reply-To: dkk@mit.edu
System name: tisiphone
Type and version: RTPC-ROMPC 7.2R (1 update(s) to same version)
Display type: apa16
What were you trying to do?
Mail in a bug report.
What's wrong:
See below.
What should have happened:
See below.
Please describe any relevant documentation references:
See below.
Below:
There has recently been a misunderstanding about whether or not
sendbug honors the EDITOR environment variable. Some expert users
said yes, it does, while other expert users said no, it doesn't. The
answer, I discovered, is that for some people sendbug honors $EDITOR,
and for others it doesn't...
The sendbug program is a Bourne shell script. The code in question is
that which looks at $EDITOR, sometimes:
------------
if [ -r $HOME/.mh_profile ]; then
comp -form $report_file
rm $report_file
exit 0
fi
# not using MH; run the editor, and send, ourselves.
if [ "x$EDITOR" = "x" ]; then
EDITOR=/usr/athena/emacs ; export EDITOR
fi
------------
If .mh_profile exists (ie: if the user has ever run any MH program)
the "comp" program is run, rather than $EDITOR. Comp runs the Editor
set in .mh_profile. The default is emacs. Comp *should* use the
EDITOR environment variable if Editor is not set in .mh_profile, but
it doesn't. As a result, for anyone with a ~/.mh_profile that doesn't
set Editor, sendbug runs the wrong program.
What happens:
sendbug runs comp
comp runs Editor or emacs, in that order
What should have happened:
sendbug runs comp
comp runs Editor, $EDITOR, or emacs, in that order
Documentation references:
tisiphone% man sendbug
No manual entry for sendbug.
from comp(1)
------------
PROFILE COMPONENTS
Editor: To override the default editor
------------
The comp program should try $EDITOR before falling back to a
compiled-in default, or perhaps the sendbug program shouldn't
blindly use comp when it sees ~/.mh_profile.
I'd like to stress that EVEN THOUGH I DON'T USE COMP, the .mh_profile
file left over from 1987 (when I used "inc", "show", "next" and "rmm",
but NEVER comp or repl) prevents sendbug from doing the right thing.
A significant fraction of expert users expect sendbug to use $EDITOR,
find that it doesn't use $EDITOR, and therefore don't use sendbug.
If you want sendbug to be used by the rest of us, you need to either
fix comp, change sendbug, or educate the educated users. (The naive
ones don't need any help...)
And another thing.... I don't know why, but the fill-in fields for
sendbug are lacking something. I can answer the questions literally
and not communicate what the bug is...
------------
When I tried to send in this bug report (after rm'ing .mh_profile so I
could edit it in $EDITOR), still more problems came up. Sendbug
(whatnow, actually) prompted me with:
You already have the standard MH directory "/mit/dkk/Mail".
Do you want to use it for MH?
I gave it "?" and found out that the only possible answers are "yes"
and "no". I looked briefly at the whatnow man page, and couldn't find
any explanation of what the program would do in either case. So I
tried answering "yes"... Sure enough, it re-created ~/.mh_profile,
ensuring that sendbug would break again next time. So, I rm'd
.mh_profile AGAIN, and tried sendbug once more with the same contents.
Once again, I got:
You already have the standard MH directory "/mit/dkk/Mail".
Do you want to use it for MH?
This time, I answered "no". It re-created .mh_profile, of course, and
followed up with:
Do you want a path below your login directory?
What on earth is this supposed to mean!? Yes, I have subdirectories.
So what? How does this have anything to do with submitting a bug
report? So I answered "no" again.
What is the whole path? /
I still have no idea what it's asking for, so I just hit Enter.
[Using existing directory]
install-mh: unable to write /context: Permission denied
spost: [install-mh aborted]
send: message not delivered to anyone
What now?
Just great. I looked over the situation a little more, and concluded
that MH is a prerequisite for sendbug. Anyone who doesn't already
have a .mh_profile, MUST have one created by sendbug or the bug report
will never get submitted. (In which case, why does it even bother to
check for the existence of .mh_profile?) I can probably just set
Editor in ~/.mh_profile to avoid all these problems, but that's
ridiculous. To expect users to modify .mh_profile even though
(actually BECAUSE) they don't use MH is totally absurd.
I certainly hope you don't seriously expect me (and other boycotters)
to use sendbug to mail in bug reports, because I think I'd rather
re-register for Unified Engineering.
