void-variable deactivate-mark

void-variable deactivate-mark

[Mats Lidell]

Hi,

Problem with latest from CVS.

gnus-version "Oort Gnus v0.08"
emacs-version "21.5  (beta9) \"brussels sprouts\" (+CVS-20021117)
XEmacs Lucid"

----------------------------------------------------------------------
Signaling: (void-variable deactivate-mark)
  gnus-agent-check-overview-buffer()
  gnus-agent-retrieve-headers((21744 21745 21746 21747 21748) "nnimap+contactor:Mail/redhat" nil)
  gnus-retrieve-headers((21744 21745 21746 21747 21748) "nnimap+contactor:Mail/redhat" nil)
  gnus-cache-retrieve-headers((21744 21745 21746 21747 21748) "nnimap+contactor:Mail/redhat" nil)
  gnus-retrieve-headers((21744 21745 21746 21747 21748) "nnimap+contactor:Mail/redhat" nil)
  gnus-fetch-headers((21744 21745 21746 21747 21748))
  gnus-select-newsgroup("nnimap+contactor:Mail/redhat" nil nil)
  gnus-summary-read-group-1("nnimap+contactor:Mail/redhat" nil nil nil nil nil)
  gnus-summary-read-group("nnimap+contactor:Mail/redhat" nil nil nil nil nil nil)
  gnus-group-read-group(nil nil nil)
  gnus-topic-read-group(nil)
  gnus-mouse-pick-topic(#<buttondown-event button2>)
  call-interactively(gnus-mouse-pick-topic)
----------------------------------------------------------------------

Yours
-- 
%% Mats

Re: void-variable deactivate-mark

[Kai Großjohann]

Mats Lidell <matsl@contactor.se> writes:

> Problem with latest from CVS.
>
> gnus-version "Oort Gnus v0.08"
> emacs-version "21.5  (beta9) \"brussels sprouts\" (+CVS-20021117)
> XEmacs Lucid"

Can you install fsf-compat for some time?  The code in question is a
debugging aid only and will be deleted after a while.

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: void-variable deactivate-mark

[Norbert Koch]

kai.grossjohann@uni-duisburg.de (Kai Großjohann) writes:

> Can you install fsf-compat for some time?  The code in question is a
> debugging aid only and will be deleted after a while.

Uhm, I have fsf-compat installed, but no cigar.  Which library should
deactivate-mark belong to?

This is a 21.4 (patch 10) "Military Intelligence" XEmacs Lucid, but
since you suggest a package, this should be of no importance.

Thanks,
norbert.

Re: void-variable deactivate-mark

[Kai Großjohann]

Norbert Koch <nk@viteno.net> writes:

> Uhm, I have fsf-compat installed, but no cigar.  Which library should
> deactivate-mark belong to?

I've now tried to make the code work on XEmacs, too.  It requires
thingatpt and protects against deactivate-mark being unbound.

Please test it and tell me what you see.

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: void-variable deactivate-mark

[Norbert Koch]

kai.grossjohann@uni-duisburg.de (Kai Großjohann) writes:

> Please test it and tell me what you see.

Looks good!

Thanks,
norbert.

Re: void-variable deactivate-mark

[Katsumi Yamaoka]

>>>>> In <84d6osh97k.fsf@lucy.cs.uni-dortmund.de>
>>>>>	kai.grossjohann@uni-duisburg.de (Kai Großjohann) wrote:

> I've now tried to make the code work on XEmacs, too.  It requires
> thingatpt and protects against deactivate-mark being unbound.

I've wrapped deactivate-mark with symbol-value and added
autoload for number-at-point instead, since XEmacs still
complained about them.
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

wrong number of arguments (Re: void-variable deactivate-mark)

[Katsumi Yamaoka]

While compiling Gnus CVS, the byte-compiler warns as follows:

While compiling gnus-summary-insert-old-articles in file /Work/gnus/lisp/gnus-sum.el:
  ** gnus-range-difference called with 1 argument, but requires 2
Wrote /Work/gnus/lisp/gnus-sum.elc

I don't know how to fix it, though.
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kai Großjohann]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> While compiling Gnus CVS, the byte-compiler warns as follows:
>
> While compiling gnus-summary-insert-old-articles in file /Work/gnus/lisp/gnus-sum.el:
>   ** gnus-range-difference called with 1 argument, but requires 2
> Wrote /Work/gnus/lisp/gnus-sum.elc
>
> I don't know how to fix it, though.

Better now?

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Katsumi Yamaoka]

>>>>> In <844ra4h67b.fsf@lucy.cs.uni-dortmund.de>
>>>>>	kai.grossjohann@uni-duisburg.de (Kai Großjohann) wrote:

>>   ** gnus-range-difference called with 1 argument, but requires 2

> Better now?

Good enough.  Thanks.
By the way, I noticed the function gnus-agent-check-overview-buffer
uses two save-excursion.  Is it needed?

P.S. I'm using Develock to see *.el files.  The new gnus-agent.el
     is much colorful. :-p  If you are interested in it, get it
     from:  ftp://ftp.jpl.org/pub/elisp/develock.el.gz
-- 
Katsumi Yamaoka <yamaoka@jpl.org>

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kai Großjohann]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> P.S. I'm using Develock to see *.el files.  The new gnus-agent.el
>      is much colorful. :-p  If you are interested in it, get it
>      from:  ftp://ftp.jpl.org/pub/elisp/develock.el.gz

I see.  Yes, I know that the source code should be reformatted, but I
decided to wait a bit before I do that.

It's more important to get it *working* first 8-)

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kai Großjohann]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> By the way, I noticed the function gnus-agent-check-overview-buffer
> uses two save-excursion.  Is it needed?

I think so.  The first save-excursion makes sure to go back to the
original buffer (if BUFFER is non-nil), the second save-excusion
saves point in the given BUFFER.

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Reiner Steib]

On Tue, Nov 26 2002, Kai Großjohann wrote:

> Katsumi Yamaoka <yamaoka@jpl.org> writes:
>
>> While compiling Gnus CVS, the byte-compiler warns as follows:
>>
>> While compiling gnus-summary-insert-old-articles in file [...]gnus-sum.el:
>>   ** gnus-range-difference called with 1 argument, but requires 2
>> Wrote /Work/gnus/lisp/gnus-sum.elc
[...]
> Better now?

Sorry, not for me. When I call `/ o' I get the following backtrace:

,----[ `/ o' with revision 6.243 of gnus-sum.el ]
| Debugger entered--Lisp error: (wrong-type-argument
| number-or-marker-p (121011 . 121196))
|   gnus-sorted-difference(((121011 . 121196)) (121197 121198 121199))
|   gnus-summary-insert-articles(((121011 . 121196)))
|   gnus-summary-insert-old-articles(nil)
| * call-interactively(gnus-summary-insert-old-articles)
`----

When I replace `gnus-summary-insert-articles' and
`gnus-summary-insert-old-articles' with the versions from revision
6.238 (the one before the changes in the agent code) it works fine.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kevin Greiner]

[-- Attachment #1: Type: text/plain, Size: 1460 bytes --]

Reiner,
I'd appreciate it if you would try the following (Right now, you're the 
only one who is reliably producing this particular problem).  First, 
reinstall the 6.43 revision of gnus-sum.el.  Then apply this patch.
diff -r6.243 gnus-sum.el
11200c11200,11201
<      (all nil)
---
 >      (all 
 >       (setq older (gnus-uncompress-range older)))

Kevin

Reiner Steib wrote:

>On Tue, Nov 26 2002, Kai Großjohann wrote:
>
>  
>
>>Katsumi Yamaoka <yamaoka@jpl.org> writes:
>>
>>    
>>
>>>While compiling Gnus CVS, the byte-compiler warns as follows:
>>>
>>>While compiling gnus-summary-insert-old-articles in file [...]gnus-sum.el:
>>>  ** gnus-range-difference called with 1 argument, but requires 2
>>>Wrote /Work/gnus/lisp/gnus-sum.elc
>>>      
>>>
>[...]
>  
>
>>Better now?
>>    
>>
>
>Sorry, not for me. When I call `/ o' I get the following backtrace:
>
>,----[ `/ o' with revision 6.243 of gnus-sum.el ]
>| Debugger entered--Lisp error: (wrong-type-argument
>| number-or-marker-p (121011 . 121196))
>|   gnus-sorted-difference(((121011 . 121196)) (121197 121198 121199))
>|   gnus-summary-insert-articles(((121011 . 121196)))
>|   gnus-summary-insert-old-articles(nil)
>| * call-interactively(gnus-summary-insert-old-articles)
>`----
>
>When I replace `gnus-summary-insert-articles' and
>`gnus-summary-insert-old-articles' with the versions from revision
>6.238 (the one before the changes in the agent code) it works fine.
>
>Bye, Reiner.
>  
>


[-- Attachment #2: Type: text/html, Size: 2187 bytes --]

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kai Großjohann]

Reiner Steib <4uce.02.r.steib@gmx.net> writes:

> On Tue, Nov 26 2002, Kai Großjohann wrote:
>
>> Katsumi Yamaoka <yamaoka@jpl.org> writes:
>>
>>> While compiling Gnus CVS, the byte-compiler warns as follows:
>>>
>>> While compiling gnus-summary-insert-old-articles in file [...]gnus-sum.el:
>>>   ** gnus-range-difference called with 1 argument, but requires 2
>>> Wrote /Work/gnus/lisp/gnus-sum.elc
> [...]
>> Better now?
>
> Sorry, not for me. When I call `/ o' I get the following backtrace:

Hm.  But now gnus-summary-insert-old-articles uses gnus-range-difference
from Kevin!  Hm.

I just tried `10 / o' and that worked well.  And `/ o 10 RET' also
works.  Let me see whether my workspace is up to date.  [time passes,
which is its job]  Oh, I still have uncommitted changes.  But those
are just changes in docstrings.

I'm confused.

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Reiner Steib]

On Tue, Nov 26 2002, Kai Großjohann wrote:

> Reiner Steib <4uce.02.r.steib@gmx.net> writes:
[...]
>> Sorry, not for me. When I call `/ o' I get the following backtrace:
>
> Hm.  But now gnus-summary-insert-old-articles uses gnus-range-difference
> from Kevin!  Hm.

gnus-summary-insert-old-articles calls gnus-summary-insert-articles
and there is ...

,----[ gnus-sum.el ]
| (defun gnus-summary-insert-articles (articles)
|   (gnus-message 9 "[1] articles: %s" articles) (sit-for 1) ;; Added for
| 							   ;; debugging!
|   (when (setq articles
| 	      (gnus-sorted-difference articles
| 				      (mapcar (lambda (h)
| 						(mail-header-number h))
| 					      gnus-newsgroup-headers)))
|     (gnus-message 9 "[2] articles: %s" articles) (sit-for 1) ;; Added for
| 							     ;; debugging!
`----

,----[ *Messages* ]
| [1] articles: ((9240 . 9565) (9567 . 9573) (9575 . 9595) (9597 .
| 9600) 9602 (9606 . 9607))
| setq: Wrong type argument: number-or-marker-p, (9240 . 9565)
`----

I don't grok that code, but it seems there's some mismatch in the
arguments (`articles' argument passed to gnus-summary-insert-article
from gnus-summary-insert-old-articles).

> I just tried `10 / o' and that worked well.

That's fine for me too. `111111 / o' too. Could you (or someone else)
try `/ o' without any prefix argument?

> And `/ o 10 RET' also works.

I don't understand this (maybe I'm too tired): `10 RET' after `/ o'?
Gives me the backtrace immediately after `/ o'

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Reiner Steib]

On Tue, Nov 26 2002, Kevin Greiner wrote:

> I'd appreciate it if you would try the following (Right now, you're
> the only one who is reliably producing this particular
> problem). 

Strange. I can reproduce it with a minimal Emacs/Gnus configuration.

> First, reinstall the 6.43 revision of gnus-sum.el. Then apply this
> patch.

Done. Same backtrace as before (of course with different numbers):

,----
| Debugger entered--Lisp error: (wrong-type-argument
| number-or-marker-p (121011 . 121200))
|   gnus-sorted-difference(((121011 . 121200)) (121201 121202 121203
| 121204 121205 121206 121207 121208 121209))
|   gnus-summary-insert-articles(((121011 . 121200)))
|   gnus-summary-insert-old-articles(nil)
| * call-interactively(gnus-summary-insert-old-articles)
`----

Kevin, could you use plain text instead of html when posting patches?
The text alternative was unusable:

> diff -r6.243 gnus-sum.el
> 11200c11200,11201
> <      (all nil)
> ---
>  >      (all >       (setq older (gnus-uncompress-range older)))

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Nevin Kapur]

Reiner Steib <4uce.02.r.steib@gmx.net> writes:

> On Tue, Nov 26 2002, Kai Großjohann wrote:
>
> That's fine for me too. `111111 / o' too. Could you (or someone else)
> try `/ o' without any prefix argument?

I get the same error as you, again only without a prefix argument.

Signaling: (wrong-type-argument number-char-or-marker-p (1 . 47902))
  =((1 . 47902) 47903)
  (cond ((= ... ...) (setq list1 ... list2 ...)) ((< ... ...) (setq out ...) (setq list1 ...)) (t (setq list2 ...)))
  (while (and list1 list2) (cond (... ...) (... ... ...) (t ...)))
  (let (out) (while (and list1 list2) (cond ... ... ...)) (nconc (nreverse out) list1))
  gnus-sorted-difference(((1 . 47902)) (47903 47904 47905 47906 47907 47908 47909 47910 47911 47912))
  gnus-summary-insert-articles(((1 . 47902)))
  gnus-summary-insert-old-articles(nil)
  call-interactively(gnus-summary-insert-old-articles)

>> And `/ o 10 RET' also works.
>
> I don't understand this (maybe I'm too tired): `10 RET' after `/ o'?
> Gives me the backtrace immediately after `/ o'

I think it depends on your value of gnus-large-newsgroup.  If the
number old articles is greater than that then Gnus will prompt you for
a number.  Hitting RET then produces the error.  If the number of old
articles is smaller than g-l-n, then I immediately see the error.

-Nevin

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Reiner Steib]

On Wed, Nov 27 2002, Nevin Kapur wrote:

> Reiner Steib <4uce.02.r.steib@gmx.net> writes:
>
>> On Tue, Nov 26 2002, Kai Großjohann wrote:
>>
>> That's fine for me too. `111111 / o' too. Could you (or someone else)
>> try `/ o' without any prefix argument?
>
> I get the same error as you, again only without a prefix argument.

Thank you for the confirmation.

>>> And `/ o 10 RET' also works.
>>
>> I don't understand this (maybe I'm too tired): `10 RET' after `/ o'?
>> Gives me the backtrace immediately after `/ o'
>
> I think it depends on your value of gnus-large-newsgroup.  If the
> number old articles is greater than that then Gnus will prompt you for
> a number.  Hitting RET then produces the error.  If the number of old
> articles is smaller than g-l-n, then I immediately see the error.

I've set gnus-large-newsgroup to nil and my leafnode (local news
server) is configured to expire already after some days, therefore, I
completely forgot about this.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Reiner Steib]

Kevin Greiner's second patch (sent to me by PM) fixed the
problem. Could somebody please commit it?

--8<---------------cut here---------------start------------->8---
--- gnus-sum.el	2002/11/27 01:57:25	6.245
+++ gnus-sum.el	2002/11/27 18:03:34
@@ -11216,9 +11216,10 @@
                       (setq all (1- all)
                             max (1- max))))))
             (setq older (gnus-uncompress-range older))))
-	 (all nil)
+	 (all
+	  (setq older (gnus-uncompress-range older)))
 	 (t
-	  (if (and (numberp gnus-large-newsgroup)
+	  (when (and (numberp gnus-large-newsgroup)
 		   (> len gnus-large-newsgroup))
 	      (let* ((cursor-in-echo-area nil)
 		     (initial (gnus-parameter-large-newsgroup-initial
@@ -11248,8 +11249,8 @@
                                         (> all 0))
                               (push max older)
                               (setq all (1- all)
-                                    max (1- max))))))
-                    (setq older (gnus-uncompress-range older))))))))
+				    max (1- max))))))))))
+	  (setq older (gnus-uncompress-range older))))
 	(if (not older)
 	    (message "No old news.")
 	  (gnus-summary-insert-articles older)
--8<---------------cut here---------------end--------------->8---

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[Kevin Greiner]

Reiner Steib wrote:

>Kevin Greiner's second patch (sent to me by PM) fixed the
>problem. Could somebody please commit it?
>  
>
Committed. gnus-sum.el revision 2.246

Kevin

>  
>

Re: wrong number of arguments (Re: void-variable deactivate-mark)

[David S Goldberg]

>>>>> On Wed, 27 Nov 2002 00:06:10 +0100, Reiner Steib <4uce.02.r.steib@gmx.net> said:

> On Tue, Nov 26 2002, Kevin Greiner wrote:
>> I'd appreciate it if you would try the following (Right now, you're
>> the only one who is reliably producing this particular
>> problem). 

Actually I was reproducing that same error pretty reliably too but was
focused more on my inability to get gnus-agent-generate to properly
handle my inbox folder.  Since my last report, I've also been offline
for a week so no news was not necessarily good news :-).  Now that I'm
back I'll try a new snapshot and see if I'm still having troubles.

Thanks,
-- 
Dave Goldberg
david.goldberg6@verizon.net

Re: void-variable deactivate-mark

[Lars Magne Ingebrigtsen]

Mats Lidell <matsl@contactor.se> writes:

> Signaling: (void-variable deactivate-mark)
>   gnus-agent-check-overview-buffer()

I've now removed the reference to that variable.  Whether the mark is
active or not should be pretty irrelevant...

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen