[nntp.el] patch for the unofficial version of mule based on 19.30

[nntp.el] patch for the unofficial version of mule based on 19.30

[SAKIYAMA Nobuo]

If you are using the current version of September Gnus on the
unofficial version of mule based on 19.29 or 19.30 and not using tm
package, your posting messages to the news would be damaged.

This is a patch resolving this problem. Of course, this code will be
slower than the original.

*** nntp.el.orig	Thu Dec 21 12:09:48 1995
--- nntp.el	Sat Dec 23 19:36:47 1995
***************
*** 1002,1008 ****
  	  (size 100))			;Size of text sent at once.
        (while (/= last (point-max))
  	(process-send-region 
! 	 nntp-server-process last (setq last (min (+ last size) (point-max))))
  	;; Read any output from the server.  May be unnecessary.
  	(accept-process-output)))
      ;; Delete the area we sent.
--- 1002,1013 ----
  	  (size 100))			;Size of text sent at once.
        (while (/= last (point-max))
  	(process-send-region 
! 	 nntp-server-process
! 	 last 
! 	 (setq last (save-excursion
! 		      (goto-char (min (+ last size) (point-max)))
! 		      (or (eobp) (forward-char 1))
! 		      (point))))
  	;; Read any output from the server.  May be unnecessary.
  	(accept-process-output)))
      ;; Delete the area we sent.

--
SAKIYAMA Nobuo        nobuo@is.s.u-tokyo.ac.jp

Re: [nntp.el] patch for the unofficial version of mule based on 19.30

[Hallvard B Furuseth]

> If you are using the current version of September Gnus on the
> unofficial version of mule based on 19.29 or 19.30 and not using tm
> package, your posting messages to the news would be damaged.
> (...)
> Of course, this code will be slower than the original.
> (...)
> ! 	 (setq last (save-excursion
> ! 		      (goto-char (min (+ last size) (point-max)))
> ! 		      (or (eobp) (forward-char 1))
> ! 		      (point))))

How is this different from:

	 (setq last (save-excursion
		      (goto-char (+ last size 1))
		      (point))))

(since goto-char always stays between point-min and point-max), or just

	 (setq last (min (+ last size 1) (point-max))))

?  Except if the result can somehow become smaller than (point-min).
Then you need

	 (setq last (max (min (+ last size 1) (point-max)) (point-min))))


Regards,

Hallvard