[Reportbug-maint] Bug#543687: reportbug: Changing editor interactively (to vi) got me a vi session, but didn't edit the report.

[Edward Welbourne]

Package: reportbug
Version: 4.6
Severity: normal


I have
export EDITOR=emacsclient
set in my environment.

Today I installed emacs23 and immediately hit a bug (can't use rmail),
so fired up reportbug.  When it came time for me to edit the bug
e-mail, I just got error messages in reportbug's virtual console;
emacsclient wasn't answering :-(

So I opted (I think the command letter was "c") to change editor
command; I used vi, which was duly started up, but on an empty buffer,
instead of on the familiar buffer I expected to see, with package
details in it.  I described the problem, saved the result and exited;
reportbug said I hadn't edited the report, so I tried again; I saw the
same file I'd been editing previously.  I told reportbug to show me
the message it was going to send - it displayed something whose last
few lines (my console is inside screen and doesn't respond usefully to
anything that I expect to cause scroll-back; nor did reportbug think
to pipe the file through a pager) looked like the normal package
description that I *expected* to be editing, but hadn't seen.

So I duly submitted the bug, over-riding reportbug's desire to abort
because I wasn't changing text.  Sure enough, when I got the
auto-responder mail, I found the familiar package description
information with no sign of the text I'd edited.

I've now exited emacs23 and started up emacs22, so that I can run
reportbug on itself - and sure enough, it invoked emacsclient; but I
found myself with a buffer called "-dir" in the working directory in
which I ran reportbug; and that directory turns out to contain a file
called "-c" containing what I edited using vi.  A quick grep of the
ouptut from ps revealed

emacsclient +6 /tmp/reportbug-reportbug-20090826-847-7U03sn

so I opened the file that named - only to discover emacs already had a
buffer open on it, that I merely hadn't yet seen (another bug for me
to report on emacs).  So now I'm editing that - who knows, it might
even work.  This may be a bug in emacsclient; but the fact that vi had
similar problems suggests there *is* a bug in reportbug, whether
that's provoking the emacsclient problem or merely independent and
confusing.

... ooooh ! I have some other suspicious-looking buffers open:
    -file                    0  Fundamental	  /tmp/-file
    -position                0  Fundamental	  /tmp/-position
    screen.rxvt              0  Fundamental	  /tmp/screen.rxvt
 %  4                        0  Fundamental	  /dev/pts/4
 %  -tty                     0  Fundamental	  /dev/pts/-tty
 %  -current-frame           0  Fundamental	  /dev/pts/-current-frame
Again, this is probably emacsclient's fault, but may be material to
investigating the reportbug issue ...

-- Package-specific info:
** Environment settings:
EDITOR="emacsclient"
VISUAL="emacsclient"
NAME="Edward Welbourne"
INTERFACE="text"

** /disk/home/eddy/.reportbugrc:
reportbug_version "3.44"
mode standard
ui text

-- System Information:
Debian Release: squeeze/sid
  APT prefers testing
  APT policy: (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)

Kernel: Linux 2.6.26-2-686 (SMP w/2 CPU cores)
Locale: LANG=en_GB.ISO-8859-15, LC_CTYPE=en_GB.ISO-8859-15 (charmap=ISO-8859-15)
Shell: /bin/sh linked to /bin/bash

Versions of packages reportbug depends on:
ii  apt                           0.7.22.2   Advanced front-end for dpkg
ii  python                        2.5.4-2    An interactive high-level object-o
ii  python-reportbug              4.6        Python modules for interacting wit

reportbug recommends no packages.

Versions of packages reportbug suggests:
pn  debconf-utils                 <none>     (no description available)
ii  debsums                       2.0.46     verification of installed package 
ii  dlocate                       1.02       fast alternative to dpkg -L and dp
ii  exim4-daemon-light [mail-tran 4.69-11    lightweight Exim MTA (v4) daemon
ii  file                          5.03-1     Determines file type using "magic"
ii  gnupg                         1.4.9-4    GNU privacy guard - a free PGP rep
pn  python-gnome2-extras          <none>     (no description available)
ii  python-gtk2                   2.14.1-3   Python bindings for the GTK+ widge
pn  python-urwid                  <none>     (no description available)
ii  python-vte                    1:0.20.5-1 Python bindings for the VTE widget
ii  xdg-utils                     1.0.2-6.1  desktop integration utilities from

-- no debconf information