🚀 go-pugleaf

Len Weisberg writes:
> I wrote:
> > I am having problems after deleting my *VM-mail* buffer with 5.23 :
> > This could be by deleting the buffer outside of vm, or by aborting
> > a mail message by using M-x kill-buffer ("^x-k") while in mail mode.
> 
> At Kyle's request, I have tried this under 5.26 and, if anything, it is 
> worse:
> 
> > I have seen two different things happen:
> >     1) Subsequent ^c-^c for sending mail gets a message:
> >         "Selecting deleted buffer"  after successfully sending the mail.
> >        This has happened in X-windows mode as well as terminal mode.
> 

I have noticed this for some time.  I have had debugging turned on in my emacs for quite a while
(for other reasons).  Every time (maybe after the first) that I send mail with ^C-^C, I get the
following traceback (I have also been running VM from the .el files for a while):

Signalling: (error "Selecting deleted buffer")
  set-buffer(#<killed buffer>)
  (save-excursion ...)
  kill-buffer(#<killed buffer>)
  mapcar(kill-buffer (#<killed buffer> #<killed buffer> #<killed buffer> #<killed buffer>))
  (let ...)
  (if ...)
  (save-excursion ...)
  (let ...)
  vm-mail-send-and-exit(nil)
* call-interactively(vm-mail-send-and-exit)

The number of #<killed buffer>'s in the mapcar grows (maybe each time I send another message, I
haven't checked).  I looked at the code, but it wasn't obvious what was wrong (this is VM 5.26).
It has always sent the mail, so I have treated it as fairly low priority.  Let me know if there
is any more useful information I can get, since it is easily reproducible (and seemingly benign
for me).

 -- Rick