🚀 go-pugleaf

Skip Montanaro wrote:
> I am using VM 8.0.12 more-or-less successfully with Aquamacs 2.0
> preview 2 and 1.7.  I encounter repeated problems with errors like
> this:
> 
>     Wrong type argument: listp, =\?iso-2022-jp\?Q\?

Dear Skip, There were clear bugs mentioned in the *Messages* buffer.
Would you like to try out the version 8.1 I announced in the vm.info
newsgroup, and see if it solves the problems?

Thanks,
Uday

The problem that has occurred is that VM wrote a badly-formed 
X-VM-v5-Data header.  Here is the header again, formatted a bit more 
clearly:

([nil nil nil nil nil nil nil nil nil]
  ["15279" "Sunday" "16" "August" "2009" "12:10:44" "-0500"
   "Carl M. Zapf\fe" "cmzapffe@comcast.net" nil "266"
   =?iso-2022-jp?Q?#("Fra_lein?= Zuffie's Hom\ecoming Party" 0 33
		    (vm-coding iso-latin-1 vm-charset "ISO-8859-1"
			       vm-string t \charset iso-8859-1)
		    33 34 (vm-coding iso-latin-1 vm-charset
				     "ISO-8859-1" vm-str\ing t))
   "^From:" nil nil "8" nil
   =?iso-2022-jp?Q?#("Fra_lein?= Zuffie's Homecom\ing Party"
		    0 8 (vm-coding iso-latin-1 vm-charset "ISO-8859-1"
				   vm-string t chars \et iso-8859-1)
		    9 17 (vm-coding iso-latin-1 vm-charset
				    "ISO-8859-1" vm-string t \charset
				    iso-8859-1)
		    18 28 (vm-coding iso-latin-1 vm-charset
				     "ISO-8859-1" vm-str\ing t charset
				     iso-8859-1)
		    29 33 (vm-coding iso-latin-1 vm-charset
				     "ISO-8859-1" \vm-string t charset
				     iso-8859-1)
		    33 34 (vm-coding iso-latin-1 vm-charset
				     "ISO-88\59-1" vm-string t))
   (number " " mark
	  "     Carl M. Zapffe    Aug 16  266/15279 "
	  thread-indent
	  "\"=?ISO-8859-1?Q?Fra=FClein_Zuffie's_Homecoming_Party?=\"\n")
   \nil nil nil nil nil nil nil]

There are three copies of the subject string here: "Fraulein Zuffie's 
Homecoming Party" (written in ASCII).  The last copy is good.  That is 
how the string is supposed to appear.  The first two copies are some 
badly corrupted versions of the string.  We have to figure out where 
they are coming from.  Some change that was incorporated between 
revisions 610 and 623 is causing this.  That is a guess.  I am guessing 
that revision 610 was all right because it was merged into the devo 
version, which I use, and I haven't noticed the problem.  Revision 623 
was released as v8.0.12.

Here is a pointer to the revisions:

   http://bazaar.launchpad.net/~hack-robf/viewmail/vm-8.0.x/changes

Cheers,
Uday

Uday S Reddy wrote:

> Here is a pointer to the revisions:
> 
>   http://bazaar.launchpad.net/~hack-robf/viewmail/vm-8.0.x/changes

Oh, you would notice that the revision 624 claims to have fixed the 
problem and the revision 625 also has a clean up routine to 
automatically clean badly formed v5-Data headers.  It is possible that 
these fixes work.  But I can't be entirely sure because I didn't find a 
mapvector procedure (which was used in revision 624).

If you would like to try out the revision 625 and see if it works 
correctly, you can use this mapvector procedure:

(defun mapvector (proc vec)
   (let ((new-vec (make-vector (length vec) nil))
	(i 0)
	(n (length vec)))
     (while (< i n)
       (aset new-vec i (apply proc (aref vec i) nil))
       (setq i (1+ i)))
     new-vec))

Cheers,
Uday