Re: uudecode barfs on XEmacs

[Reiner Steib <reinersteib+gmane@imap.cc>]

On Sat, May 20 2006, Michael Sperber wrote:

> Error while decoding: (invalid-state Bogus value for
> `default-process-coding-system' nil)
>
> Now, I can't say I know what I'm doing (or that this works on GNU
> Emacs), but this fixed it for me:
>
> 2006-05-20  Michael Sperber  <sperber@deinprogramm.de>
>
> 	* uudecode.el (uudecode-decode-region-external): Use '(binary
> 	. binary) for `default-process-coding-system' instead of nil to
> 	keep `call-process-region' from barfing.

I seem to recall that this (or maybe something similar) has been
suggested previously.  Yes, here[1] it is.  Alas, Stephen neither was
sure that it's the right thing to do.

But I think making it conditional on (featurep 'xemacs) should be
better than the current state, so I added this (in the v5-10 branch;
will be synced to the trunk soon):

--8<---------------cut here---------------start------------->8---
--- uudecode.el	08 Feb 2006 13:34:07 +0100	6.6.2.6
+++ uudecode.el	22 May 2006 18:53:08 +0200	
@@ -100,7 +100,11 @@
 			      (make-temp-name "uu")
 			      uudecode-temporary-file-directory))))
 	(let ((cdir default-directory)
-	      default-process-coding-system)
+	      (default-process-coding-system
+		(if (featurep 'xemacs)
+		    ;; In XEmacs, `nil' is not a valid coding system.
+		    (binary . binary)
+		  nil)))
 	  (unwind-protect
 	      (with-temp-buffer
 		(insert "begin 600 " (file-name-nondirectory tempfile) "\n")
--8<---------------cut here---------------end--------------->8---

Bye, Reiner.

Footnotes: 
[1]  
,----[ http://article.gmane.org/gmane.emacs.gnus.general/60028 ]
| sigh* I hate doing this---I don't have very good reason to suppose
| this will work for anyone but me---but I don't have time to do it
| right.  Here's a patch against current XEmacs package CVS, but it
| should be pretty close for this ancient protocol.
`----
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/