NT / pgnus draft problem

NT / pgnus draft problem

[Alfred J Correira]

Are you an NTemacs user and do you ever use drafts (I'm making an assumption
here but I'd be very surprised if this behavior isn't Windows specific)?  I do
from time to time and somewhere between v0.79 and v0.84 drafts broke for me in
the following way: I save a message as a draft, and then revisit it later (via
"D e"). When I do so, I am placed in a buffer with the draft email, but the
buffer has broken EOL conventions in force (^Ms at the end of every line); as a
result, I get an error when Gnus attempts to look for two EOL characters in a
row and can't find them.

I have verified this behavior under NTEmacs 20.3.8.2 / pgnus v0.84 and someone
else did under NTEmacs 20.3 / pgnus v0.83.  If you are running something older,
let me know if it does / doesn't work correctly for you ... of course, if you
already know what is going wrong, that's even better :-)

-- 
Work: Alfred.Correira@GlobeSet.Com  PH: 512.427.7715
Home: Correira@Flash.Net           FAX: 512.427.5101
PGP: 55 42 BB 2E 17 E9 0D 9C 0B 51 66 58 58 96 1E 45
---
Progress is made by lazy men looking for easier ways to do things.
-- Robert Heinlein

Re: NT / pgnus draft problem

[Jack Vinson]

>>>>> "AJC" == Alfred J Correira <alfred@GlobeSet.com> writes:

[snip description]

AJC> I have verified this behavior under NTEmacs 20.3.8.2 / pgnus v0.84 and
AJC> someone else did under NTEmacs 20.3 / pgnus v0.83.  If you are running
AJC> something older, let me know if it does / doesn't work correctly for
AJC> you ... of course, if you already know what is going wrong, that's
AJC> even better :-)

I am the second verifier.  I just looked into the problem a little more.
While I can't fix the problem, its source is in nndraft-request-article.
There is a code chunk that does

		 (let ((nnmail-file-coding-system
			(if (file-newer-than-file-p file auto)
			    'binary
			  message-draft-coding-system)))
		   (nnmail-find-file newest))

The normal/default nnmail-file-coding-system seems to be 'binary on my
machine, but the message-draft-coding-system is 'emacs-mule.  Might this be
the source of the difficulties?

-- 
Jack Vinson <jvinson@chevax.ecs.umass.edu>    http://www.cis.upenn.edu/~vinson/
Zippy: Jesus is my POSTMASTER GENERAL..

Re: NT / pgnus draft problem

[Shenghuo ZHU]

[-- Attachment #1: Type: text/plain, Size: 1345 bytes --]

>>>>> "JV" == Jack Vinson <jvinson@chevax.ecs.umass.edu> writes:

>>>>> "AJC" == Alfred J Correira <alfred@GlobeSet.com> writes:
JV> [snip description]

AJC> I have verified this behavior under NTEmacs 20.3.8.2 / pgnus v0.84 and
AJC> someone else did under NTEmacs 20.3 / pgnus v0.83.  If you are running
AJC> something older, let me know if it does / doesn't work correctly for
AJC> you ... of course, if you already know what is going wrong, that's
AJC> even better :-)

JV> I am the second verifier.  I just looked into the problem a little
JV> more.  While I can't fix the problem, its source is in
JV> nndraft-request-article.  There is a code chunk that does

JV> 		 (let ((nnmail-file-coding-system
JV> 			(if (file-newer-than-file-p file auto)
JV> 			    'binary
JV> 			  message-draft-coding-system)))
JV> 		   (nnmail-find-file newest))

JV> The normal/default nnmail-file-coding-system seems to be 'binary
JV> on my machine, but the message-draft-coding-system is 'emacs-mule.
JV> Might this be the source of the difficulties?

I hope someone verify it again after 
        (setq message-draft-coding-system 'emacs-mule-dos)

(Sorry, I haven't verified it in NT Emacs.)

If it works, here is the patch.

Wed May  5 01:15:08 1999  Shenghuo ZHU  <zsh@cs.rochester.edu>

	* message.el (message-draft-coding-system): Use emacs-mule-dos.


[-- Attachment #2: 84-zsh4.diff --]
[-- Type: application/octet-stream, Size: 487 bytes --]

--- message.el	1999/05/05 05:08:26	1.1
+++ message.el	1999/05/05 05:08:56
@@ -862,7 +862,9 @@
 (defvar message-draft-coding-system
   (cond
    ((not (fboundp 'coding-system-p)) nil)
-   ((coding-system-p 'emacs-mule) 'emacs-mule)
+   ((coding-system-p 'emacs-mule) 
+    (if (string-match "nt" system-configuration)
+	'emacs-mule-dos 'emacs-mule))
    ((memq 'escape-quoted (mm-get-coding-system-list)) 'escape-quoted)
    ((coding-system-p 'no-conversion) 'no-conversion)
    (t nil))

[-- Attachment #3: Type: text/plain, Size: 20 bytes --]



-- 
Shenghuo ZHU

Re: NT / pgnus draft problem

[Alfred J Correira]

Shenghuo ZHU <zsh@cs.rochester.edu> writes:

> >>>>> "JV" == Jack Vinson <jvinson@chevax.ecs.umass.edu> writes:
> 
> >>>>> "AJC" == Alfred J Correira <alfred@GlobeSet.com> writes:
> JV> [snip description]
> 
> AJC> I have verified this behavior under NTEmacs 20.3.8.2 / pgnus v0.84 and
> AJC> someone else did under NTEmacs 20.3 / pgnus v0.83.  If you are running
> AJC> something older, let me know if it does / doesn't work correctly for
> AJC> you ... of course, if you already know what is going wrong, that's
> AJC> even better :-)
> 
> JV> I am the second verifier.  I just looked into the problem a little
> JV> more.  While I can't fix the problem, its source is in
> JV> nndraft-request-article.  There is a code chunk that does
> 
> JV> 		 (let ((nnmail-file-coding-system
> JV> 			(if (file-newer-than-file-p file auto)
> JV> 			    'binary
> JV> 			  message-draft-coding-system)))
> JV> 		   (nnmail-find-file newest))
> 
> JV> The normal/default nnmail-file-coding-system seems to be 'binary
> JV> on my machine, but the message-draft-coding-system is 'emacs-mule.
> JV> Might this be the source of the difficulties?
> 
> I hope someone verify it again after 
>         (setq message-draft-coding-system 'emacs-mule-dos)
> 
> (Sorry, I haven't verified it in NT Emacs.)

I just tried it and it didn't work for me ...

> If it works, here is the patch.
> 
> Wed May  5 01:15:08 1999  Shenghuo ZHU  <zsh@cs.rochester.edu>
> 
> 	* message.el (message-draft-coding-system): Use emacs-mule-dos.

-- 
Work: Alfred.Correira@GlobeSet.Com  PH: 512.427.7715
Home: Correira@Flash.Net           FAX: 512.427.5101
PGP: 55 42 BB 2E 17 E9 0D 9C 0B 51 66 58 58 96 1E 45
---
With searching comes loss
And the presence of absence:
"My Novel" not found.
-- Haiku: Last Listed Source: Cynthia Johnson, Minnesota Public Radio

Re: NT / pgnus draft problem

[Matt Armstrong]

alfred@GlobeSet.com (Alfred J Correira) writes:

> Are you an NTemacs user and do you ever use drafts (I'm making an
> assumption here but I'd be very surprised if this behavior isn't
> Windows specific)?  I do from time to time and somewhere between
> v0.79 and v0.84 drafts broke for me in the following way: I save a
> message as a draft, and then revisit it later (via "D e"). When I do
> so, I am placed in a buffer with the draft email, but the buffer has
> broken EOL conventions in force (^Ms at the end of every line); as a
> result, I get an error when Gnus attempts to look for two EOL
> characters in a row and can't find them.
>
> I have verified this behavior under NTEmacs 20.3.8.2 / pgnus v0.84
> and someone else did under NTEmacs 20.3 / pgnus v0.83.  If you are
> running something older, let me know if it does / doesn't work
> correctly for you ... of course, if you already know what is going
> wrong, that's even better :-)

I'm using NTEmacs at work with pgnus 0.76 (or so) and drafts work.
So, this problem was probably introduced in recent versions.

Re: NT / pgnus draft problem

[Shenghuo ZHU]

[-- Attachment #1: Type: text/plain, Size: 559 bytes --]

>>>>> "AJC" == Alfred J Correira <alfred@GlobeSet.com> writes:
>> 
>> I hope someone verify it again after 
>> (setq message-draft-coding-system 'emacs-mule-dos)
>> 

AJC> I just tried it and it didn't work for me ...


Use this, 
 (setq message-draft-coding-system 'emacs-mule-unix)

I've verified it in NT Emacs 20.3.6.1. 

Note : it does not work for the previous drafts, i.e. drafts already
contains ^Ms.

-- 
Shenghuo ZHU


Wed May  5 12:15:08 1999  Shenghuo ZHU  <zsh@cs.rochester.edu>

	* message.el (message-draft-coding-system): Use emacs-mule-dos.


[-- Attachment #2: 84-zsh4.diff --]
[-- Type: application/octet-stream, Size: 488 bytes --]

--- message.el	1999/05/05 05:08:26	1.1
+++ message.el	1999/05/05 16:18:14
@@ -862,7 +862,9 @@
 (defvar message-draft-coding-system
   (cond
    ((not (fboundp 'coding-system-p)) nil)
-   ((coding-system-p 'emacs-mule) 'emacs-mule)
+   ((coding-system-p 'emacs-mule) 
+    (if (string-match "nt" system-configuration)
+	'emacs-mule-unix 'emacs-mule))
    ((memq 'escape-quoted (mm-get-coding-system-list)) 'escape-quoted)
    ((coding-system-p 'no-conversion) 'no-conversion)
    (t nil))

Re: NT / pgnus draft problem

[Alfred J Correira]

Shenghuo ZHU <zsh@cs.rochester.edu> writes:

> >>>>> "AJC" == Alfred J Correira <alfred@GlobeSet.com> writes:
> >>
> >> I hope someone verify it again after
> >> (setq message-draft-coding-system 'emacs-mule-dos)
> >>
>
> AJC> I just tried it and it didn't work for me ...
>
>
> Use this,
>  (setq message-draft-coding-system 'emacs-mule-unix)
>
> I've verified it in NT Emacs 20.3.6.1.
>
> Note : it does not work for the previous drafts, i.e. drafts already
> contains ^Ms.

Yes, I knew that :-)

Anyway, I tried the above setting (emacs-mule-unix) with this reply ...

... and it worked!

I also tried this setting from a private reply:

(setq-default buffer-file-coding-system 'undecided-unix)

... and it also worked.

Any opinions about which is "more correct" in this context or do we let Lars
decide? :-) 

--
Work: Alfred.Correira@GlobeSet.Com  PH: 512.427.7715
Home: Correira@Flash.Net           FAX: 512.427.5101
PGP: 55 42 BB 2E 17 E9 0D 9C 0B 51 66 58 58 96 1E 45
---
Pat Williams, Orlando Magic general manager, on his team's 7-27 record:
We can't win at home. We can't win on the road. As general manager, I just
can't figure out where else to play.

Re: NT / pgnus draft problem

[Christophe CUQ]

On 04 May 1999 13:35:10 -0500, Alfred J Correira wrote :

> Are you an NTemacs user and do you ever use drafts (I'm making an assumption
> here but I'd be very surprised if this behavior isn't Windows specific)?  I do
> from time to time and somewhere between v0.79 and v0.84 drafts broke for me in
> the following way: I save a message as a draft, and then revisit it later (via
> "D e"). When I do so, I am placed in a buffer with the draft email, but the
> buffer has broken EOL conventions in force (^Ms at the end of every line); as a
> result, I get an error when Gnus attempts to look for two EOL characters in a
> row and can't find them.

I had the same in my Draft group and in my Gcc: group.

The problem is in the buffer-file-coding-system in Ntemacs which is
'dos' by default.

To fix this problem I put the following in my .emacs (found in
gnus.emacs.vm.info) :

(setq-default buffer-file-coding-system 'undecided-unix)

And it runs well.

Hth

-- 
CHC    PGP 0xC387F232     http://www.cuq.org/