wrong-type-argument exiting nnml group

wrong-type-argument exiting nnml group

[Samuel Padgett]

I'm running Oort Gnus 0.04, updated from CVS yesterday.  When I quit
one of my nnml groups with `q', I get this error:

Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p Sun\,)
  <(Sun\, <200108122203\.f7CM3mq00651@telengard\.sturk\.net>)
  sort((33) <)
  nnml-request-expire-articles((29 31) "mail.misc" "" nil)
  gnus-request-expire-articles((29 31) "nnml:mail.misc")
  gnus-summary-expire-articles()
  run-hooks(gnus-summary-prepare-exit-hook)
  apply(run-hooks gnus-summary-prepare-exit-hook)
  gnus-run-hooks(gnus-summary-prepare-exit-hook)
  gnus-summary-exit()
  call-interactively(gnus-summary-exit)

Any thoughts?

Sam
-- 
May all your PUSHes be POPped.

Re: wrong-type-argument exiting nnml group

[Simon Josefsson]

Samuel Padgett <spadgett1@nc.rr.com> writes:

> I'm running Oort Gnus 0.04, updated from CVS yesterday.  When I quit
> one of my nnml groups with `q', I get this error:
>
> Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p Sun\,)
>   <(Sun\, <200108122203\.f7CM3mq00651@telengard\.sturk\.net>)
>   sort((33) <)
>   nnml-request-expire-articles((29 31) "mail.misc" "" nil)

Maybe the NOV file is corrupt.  Try M-x nnml-generate-nov-databases RET.

(Or manually inspect the .overview file to possibly find why it became
corrupt.)

Re: wrong-type-argument exiting nnml group

[Samuel Padgett]

Simon Josefsson <jas@extundo.com> writes:

> Try M-x nnml-generate-nov-databases RET.

Thanks.  That fixed it.

> (Or manually inspect the .overview file to possibly find why it became
> corrupt.)

Gah.  Information for article 1 was split up over multiple lines
instead of all on one NOV line.  Any idea how that might have
happened?

Hm.  Looking at it after nnml-generate-nov-databases, I see that there
is a ^M after each header field for article 1.  Could this have caused
the problem?  I'm running Debian unstable.  Maybe this is an issue
with my MTA?

Sam
-- 
May all your PUSHes be POPped.

Re: wrong-type-argument exiting nnml group

[Simon Josefsson]

Samuel Padgett <spadgett1@nc.rr.com> writes:

>> (Or manually inspect the .overview file to possibly find why it became
>> corrupt.)
>
> Gah.  Information for article 1 was split up over multiple lines
> instead of all on one NOV line.  Any idea how that might have
> happened?
>
> Hm.  Looking at it after nnml-generate-nov-databases, I see that there
> is a ^M after each header field for article 1.  Could this have caused
> the problem?

Yes.  I do not understand how that happens though, even if the article
in question uses CRLF the NOV parsing routine isn't fooled, at least
for me.

If you edebug `nnml-generate-nov-file', does the *nntpd* buffer
contain ^M's after the call to `nnheader-insert-file-contents'?  It
doesn't for me, and I suspect it is a feature but I'm not sure why it
happens.

Re: wrong-type-argument exiting nnml group

[Samuel Padgett]

Simon Josefsson <jas@extundo.com> writes:

> Samuel Padgett <spadgett1@nc.rr.com> writes:
>
>> Hm.  Looking at it after nnml-generate-nov-databases, I see that there
>> is a ^M after each header field for article 1.  Could this have caused
>> the problem?
>
> Yes.  I do not understand how that happens though, even if the article
> in question uses CRLF the NOV parsing routine isn't fooled, at least
> for me.

I looked at the article more closely and noticed it has mixed line
endings.  The first four lines end with just a LF, whereas the rest
end with CRLF.

> If you edebug `nnml-generate-nov-file', does the *nntpd* buffer
> contain ^M's after the call to `nnheader-insert-file-contents'?  It
> doesn't for me, and I suspect it is a feature but I'm not sure why it
> happens.

Yes, the *nntpd* buffer contains the ^M's.  I'm sure it's because of
the mixed line endings, though.

Sam
-- 
May all your PUSHes be POPped.

Re: wrong-type-argument exiting nnml group

[Simon Josefsson]

Samuel Padgett <spadgett1@nc.rr.com> writes:

> Simon Josefsson <jas@extundo.com> writes:
>
>> Samuel Padgett <spadgett1@nc.rr.com> writes:
>>
>>> Hm.  Looking at it after nnml-generate-nov-databases, I see that there
>>> is a ^M after each header field for article 1.  Could this have caused
>>> the problem?
>>
>> Yes.  I do not understand how that happens though, even if the article
>> in question uses CRLF the NOV parsing routine isn't fooled, at least
>> for me.
>
> I looked at the article more closely and noticed it has mixed line
> endings.  The first four lines end with just a LF, whereas the rest
> end with CRLF.

Ouch.  Nnml articles should have LF line endings.  Can you find out
how this happened?

Anyway, I committed the following to make nnml a little bit more
robust about these kind of errors.  You should be able to M-x
nnml-generate-nov-databases RET and having things work now.

2001-09-15  Simon Josefsson  <jas@extundo.com>

	* nnml.el (nnml-parse-head): Handle CRLF files.
	(nnml-generate-nov-file): Ditto.
	(nnml-retrieve-headers): Ditto.

Index: nnml.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/nnml.el,v
retrieving revision 6.25
diff -u -u -w -u -w -r6.25 nnml.el
--- nnml.el	2001/09/08 20:55:16	6.25
+++ nnml.el	2001/09/15 21:17:19
@@ -157,7 +157,7 @@
 		(setq beg (point))
 		(nnheader-insert-head file)
 		(goto-char beg)
-		(if (search-forward "\n\n" nil t)
+		(if (re-search-forward "\n\r?\n" nil t)
 		    (forward-char -1)
 		  (goto-char (point-max))
 		  (insert "\n\n"))
@@ -713,13 +713,15 @@
       (unless (zerop (buffer-size))
 	(narrow-to-region
 	 (goto-char (point-min))
-	 (if (search-forward "\n\n" nil t) (1- (point)) (point-max))))
+	 (if (re-search-forward "\n\r?\n" nil t) (1- (point)) (point-max))))
       ;; Fold continuation lines.
       (goto-char (point-min))
       (while (re-search-forward "\\(\r?\n[ \t]+\\)+" nil t)
 	(replace-match " " t t))
       ;; Remove any tabs; they are too confusing.
       (subst-char-in-region (point-min) (point-max) ?\t ? )
+      ;; Remove any ^M's; they are too confusing.
+      (subst-char-in-region (point-min) (point-max) ?\r ? )
       (let ((headers (nnheader-parse-head t)))
 	(mail-header-set-chars headers chars)
 	(mail-header-set-number headers number)
@@ -838,7 +840,7 @@
 	  (narrow-to-region
 	   (goto-char (point-min))
 	   (progn
-	     (search-forward "\n\n" nil t)
+	     (re-search-forward "\n\r?\n" nil t)
 	     (setq chars (- (point-max) (point)))
 	     (max 1 (1- (point)))))
 	  (unless (zerop (buffer-size))

Re: wrong-type-argument exiting nnml group

[Samuel Padgett]

Simon Josefsson <jas@extundo.com> writes:

> Ouch.  Nnml articles should have LF line endings.  Can you find out
> how this happened?

Sure, I'll look into it.  I'm confident it doesn't have anything to do
with Gnus, though.  For a while, some months back, I strayed from the
path.  I was using procmail to filter mail into maildir folders that I
read with Mutt.  I noticed some ^M's in mail headers way back then.

Again, I'm running Debian unstable, so it wouldn't surprise me if an
unstable version of procmail or Mutt is to blame.

> Anyway, I committed the following to make nnml a little bit more
> robust about these kind of errors.  You should be able to M-x
> nnml-generate-nov-databases RET and having things work now.
>
> 2001-09-15  Simon Josefsson  <jas@extundo.com>
>
> 	* nnml.el (nnml-parse-head): Handle CRLF files.
> 	(nnml-generate-nov-file): Ditto.
> 	(nnml-retrieve-headers): Ditto.

[...]

Ah, cool.  I see no more ^M's in .overview.  Thanks!

Sam
-- 
May all your PUSHes be POPped.