Re: 23.0.60; rmail-redecode-body finishing with error "Invalid Search Bound"

[David Golden <david.golden@oceanfree.net>]

On Wednesday 16 January 2008, David Golden wrote:
> On Wednesday 16 January 2008, martin rudalics wrote:
> > With your rmail.redecode I get on my system
>
> Just to check: your system ==> emacs-unicode-2 or cvs head or...?
>
> > at =1=> for `point' the value 223 and for `end' 1146
> > at =2=> for `beg' 183 and for `point' 871
> >
> > hence it doesn't fail for me.  Since it apparently fails at =1=>
> > for you could you please tell me the values for `point' and `end'
> > there?
>
> (point)  :  224
> end: 190
>
> [and beg: 183]
>
> Soo... Indeed end is before point.
>
> in (let ((end (rmail-msgend n)))  in rmail-show-message [line 2800]
> (rmail-msgend n) is returning 190, which seems, well, wrong.
> (n is 1).
>


When rmail-msgend is returning 190 (i.e. when called by 
rmail-show-message called by rmail-redecode-body), rmail-message-vector 
itself seems to have markers at 1,183,190 instead of 1,183,1176 in it.

In rmail-redecode-body, perhaps suspiciously

		    (goto-char x-coding-header)
		    (delete-region (point)
				   (save-excursion
				     (beginning-of-line)
				     (point)))
		    (insert "X-Coding-System: "
			    (symbol-name last-coding-system-used))
===>		    (set-marker x-coding-header nil) 

x-coding-header happens to have a value #<marker at 190 in RMAIL>