[PATCH] -> Re: [PATCH] Problems accessing spam group (again)

[PATCH] -> Re: [PATCH] Problems accessing spam group (again)

[Nelson Ferreira]

As I suspected, I was not fixing the right problem. It just so happens
that somehow when gnus-dup-unsupress-article gnus-data-find returns
nil for the article.

Not sure why, and still maybe not fixing the right problem, this patch now really does the trick for me.

diff -u -r7.3 gnus-dup.el
--- gnus-dup.el	20 May 2004 08:02:39 -0000	7.3
+++ gnus-dup.el	8 Jun 2004 12:10:20 -0000
@@ -154,9 +154,10 @@
 
 (defun gnus-dup-unsuppress-article (article)
   "Stop suppression of ARTICLE."
-  (let* ((header (gnus-data-header (gnus-data-find article)))
+  (let* ((art    (gnus-data-find article))
+	 (header (when art (gnus-data-header art)))
 	 (id     (when header (mail-header-id header))))
-    (when id
+    (when (and id gnus-dup-hashtb)
       (setq gnus-dup-list-dirty t)
       (setq gnus-dup-list (delete id gnus-dup-list))
       (unintern id gnus-dup-hashtb))))


"njsf" == Nelson Ferreira <nelson.ferreira@verizon.net> writes:

    njsf> Intermitently I've been getting the following error:
    njsf> Signaling: (wrong-type-argument arrayp nil)
    njsf>   gnus-dup-unsuppress-article(35829)
    njsf>   gnus-mark-article-as-unread(35829 ?$)
    njsf>   gnus-summary-mark-article(35829 ?$)
    njsf>   spam-mark-junk-as-spam-routine() spam-summary-prepare()
    njsf>   run-hooks(spam-summary-prepare) apply(run-hooks
    njsf>   spam-summary-prepare)
    njsf>   gnus-run-hooks(gnus-summary-prepare-hook)
    njsf>   gnus-summary-prepare()
    njsf>   gnus-summary-read-group-1("nnml:mail.spam" nil t nil nil
    njsf>   nil) gnus-summary-read-group("nnml:mail.spam" nil t nil
    njsf>   nil nil nil) gnus-group-read-group(nil t)
    njsf>   gnus-group-select-group(nil) gnus-topic-select-group(nil)
    njsf>   call-interactively(gnus-topic-select-group)

    njsf> I traced it to gnus-dup-unsuppress-article being called when
    njsf> gnus-dup-hashtb is not an array.

    njsf> I fixed it with the following (trivial) patch, although not
    njsf> sure if it is the solution to the real problem:


[...]


    njsf> -- Nelson Ferreira

-- 
Nelson Ferreira

Re: [PATCH] -> Re: [PATCH] Problems accessing spam group (again)

[Ted Zlatanov]

On Tue, 08 Jun 2004, nelson.ferreira@verizon.net wrote:

> As I suspected, I was not fixing the right problem. It just so happens
> that somehow when gnus-dup-unsupress-article gnus-data-find returns
> nil for the article.
> 
> Not sure why, and still maybe not fixing the right problem, this patch now
> 	really does the trick for me.

I don't understand the reasons for this problem, but I've observed it
too.  Maybe someone who understands summary generation well can tell
us if your patch is what's needed.

Ted

Re: [PATCH] -> Re: [PATCH] Problems accessing spam group (again)

[Nelson Ferreira]

"Ted" == Ted Zlatanov <tzz@lifelogs.com> writes:

    Ted> On Tue, 08 Jun 2004, nelson.ferreira@verizon.net wrote:
    >> As I suspected, I was not fixing the right problem. It just so
    >> happens that somehow when gnus-dup-unsupress-article
    >> gnus-data-find returns nil for the article.
    >>
    >> Not sure why, and still maybe not fixing the right problem,
    >> this patch now really does the trick for me.

    Ted> I don't understand the reasons for this problem, but I've
    Ted> observed it too.  Maybe someone who understands summary
    Ted> generation well can tell us if your patch is what's needed.

I think it might have something to do with scoring plus expiry-wait ....

Maybe someone more familiar with the code can verify that when the
score for an article reaches below the expunge and the parameters
below would make the article unavailable before dup-unsupress is
called ?

((auto-expire . t)
 (total-expire . t)
 (spam-contents gnus-group-spam-classification-spam)
 (spam-process-destination)
 (ham-process-destination . "nnml:mail.misc")
 (expiry-wait . immediate)
 (total-expiry . t))




    Ted> Ted

-- 
Nelson Ferreira