Entering spam-group, using all CPU, XEmacs 21.4p16 to blame?

Entering spam-group, using all CPU, XEmacs 21.4p16 to blame?

[Adam Sjøgren]

[-- Attachment #1: Type: text/plain, Size: 4939 bytes --]

  Hi.


Recently XEmacs 21.4p16 entered Debian unstable. I think that my problem
started with that, but I don't know - here we go:

When I enter my spam-group, XEmacs starts using a lot of CPU, but
nothing seems to happen. I let it sit for 5 minutes, without it
(seemingly) getting anywhere.

Here's the backtrace I get, if I kill xemacs after starting it from
the commandline, try to enter the spam-group, and then wait for 5
minutes:

 asjo@topper:~$ xemacs

 Lisp backtrace follows:

   string-match("=\\?\\([^][- ()<>@,;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\\(B\\|Q\\)\\?\\([!->@-~ ]*\\
 )\\?=" "=?utf-8?q?select?=")
   # bind (word)
   rfc2047-parse-and-decode("=?utf-8?q?select?=")
   # (unwind-protect ...)
   # (unwind-protect ...)
   # bind (case-fold-search b e start end)
   # (unwind-protect ...)
   # (unwind-protect ...)
   # bind (temp-buffer m string)
   mail-decode-encoded-word-string("=?utf-8?q?Meliorate your sperm?= =?utf-8?q? cell quantity and ?= 
 =?utf-8?q?select?=")
   byte-code("..." [x gnus-decode-encoded-word-function string eol search-forward "      " move mm-su
 bst-char-in-string ?\r ?- ?\n t] 8)
   # (condition-case ... . ((error x)))
   # bind (x)
   # (unwind-protect ...)
   # bind (eol buffer header references in-reply-to number dependencies force-new)
   byte-code("..." [id number eol id-dep references header read point-at-eol nil ((widen)) (byte-code "[...elided...]" [x gnus-decode-encoded-word-function string eol search-forward
  "      " move mm-subst-char-in-string ?\r ?- ?\n t] 8) ((error x)) (byte-code "[...elided...]" [x gnus-decode-encoded-word-function string eol search-forward "      " move mm-su
 bst-char-in-string ?\r ?- ?\n t] 8) ((error x)) search-forward "        " move string-match "^<[^>]+
 >$" format "fake+none+%s+%d" "fake+none+%s+%s" int-to-string ?\t 0 (byte-code "[...elided...]" [read] 2) ((err
 or)) 1 (byte-code "[...elided...]" [read] 2) ((error)) looking-at "Xref: " (?\n nil) "^\\([^ :]+\\): " intern 
 match-string vector "" 9 In-Reply-To 5 gnus-extract-message-id-from-in-reply-to 4 boundp t nnmail-me
 ssage-id 8 zerop gnus-split-references "\\(<[^<]+>\\)[  ]*\\'" intern-soft "none" ids sequence ref-d
 ep dependencies ref n force-new in-reply-to string num out parent-id allp ref-header replaced gnus-n
 ewsgroup-name nnheader-fake-message-id extra xref lines chars date from subject gnus-alter-header-fu
 nction headers cur buffer x gnus-summary-ignore-duplicates] 16)
   # (condition-case ... . ((error (byte-code "[...elided...]" ... 6))))
   # (unwind-protect ...)
   # bind (mail-parse-charset mail-parse-ignored-charsets cur dependencies allp number headers header
  also-fetch-heads group dependencies force-new sequence)
   gnus-get-newsgroup-headers-xover((104569 104570 104571 104572 104573) nil nil "nnml:spam" t)
   # bind (name articles)
   gnus-fetch-headers((104569 104570 104571 104572 104573))
   # bind (cached fetched-articles articles info gnus-summary-ignore-duplicates entry select-articles
  read-all group)
   gnus-select-newsgroup("nnml:spam" nil nil)
   # bind (quit-config new-group select-articles no-display kill-buffer no-article show-all group)
   gnus-summary-read-group-1("nnml:spam" nil nil nil nil nil)
   # bind (gnus-auto-select-next result select-articles backward no-display kill-buffer no-article sh
 ow-all group)
   gnus-summary-read-group("nnml:spam" nil nil nil nil nil nil)
   # bind (no-display group number active marked entry select-articles group no-article all)
   gnus-group-read-group(nil nil nil)
   # bind (group no-article all)
   gnus-topic-read-group(nil)
   # bind (command-debug-status)
   call-interactively(gnus-topic-read-group)
   # (condition-case ... . error)
   # (catch top-level ...)
 Terminated
 asjo@topper:~$ 


Attached is the email it seems to be chewing on for reference.


Here's the start of two other traces (same symptoms):

 Lisp backtrace follows:

   string-match("=\\?\\([^][- ()<>@,;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\\(B\\|Q\
 \)\\?\\([!->@-~ ]*\\)\\?=" "=?utf-8?q? cell quantity and ?=")
   # bind (word)
   rfc2047-parse-and-decode("=?utf-8?q? cell quantity and ?=")
   # (unwind-protect ...)


 Lisp backtrace follows:

   string-match("=\\?\\([^][- ()<>@,;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\\(B\\|Q\
 \)\\?\\([!->@-~ ]*\\)\\?=" "=?utf-8?q?Theodora Cay?=")
   # bind (word)
   rfc2047-parse-and-decode("=?utf-8?q?Theodora Cay?=")
   # (unwind-protect ...)


Any ideas?


 XEmacs 21.4 (patch 16) "Corporate Culture" [Lucid]
 (i386-debian-linux, Mule) of Sun Dec 12 2004 on penell

 No Gnus v0.3 (cvs from 2004-11-22, I re-bytecompiled after installing
               the new XEmacs packages).


 The box is a P4 2.4 GHz with RAM enough, so I don't think it is
 because I'm being too impatient.


  Best regards,

-- 
 "Snurra min jord igen"                                       Adam Sjøgren
                                                         asjo@koldfront.dk


[-- Attachment #2: 104569 --]
[-- Type: message/rfc822, Size: 2581 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 214 bytes --]

These pills is a in a scientific way formalised herbaceous
food alloy to reinforce birthrate by amending spermatozoon class,
number and motion.

http://namgmounuc.mpromeds.info/?ccdeqicamxtvuychzsmnkenpesemwshm

[-- Attachment #2.1.2: Type: text/html, Size: 491 bytes --]

Re: Entering spam-group, using all CPU, XEmacs 21.4p16 to blame?

[Reiner Steib]

On Sun, Dec 19 2004, Adam Sjøgren wrote:

> Recently XEmacs 21.4p16 entered Debian unstable. I think that my problem
> started with that, but I don't know - here we go:
[...]
> Here's the backtrace I get, if I kill xemacs after starting it from
> the commandline, 

Doesn't XEmacs have `toggle-debug-on-quit'?

> try to enter the spam-group, and then wait for 5 minutes:[...]

Here and also in you other message <news:87u0qibbw2.fsf@koldfront.dk>,
XEmacs hangs in `rfc2047-parse-and-decode'.

Can you reproduce the inf-loop by putting ...

(require 'rfc2047)
(rfc2047-parse-and-decode "=?utf-8?q? cell quantity and ?=")
(string-match rfc2047-encoded-word-regexp "=?utf-8?q? cell quantity and ?=")

... into the scratch buffer and hitting `C-j' after each line?

For me, it looks like an XEmacs bug.  Probably you should use `M-x
report-xemacs-bug RET' and report it.

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

Re: Entering spam-group, using all CPU, XEmacs 21.4p16 to blame?

[Adam Sjøgren]

On Sun, 19 Dec 2004 18:51:11 +0100, Reiner wrote:

> Doesn't XEmacs have `toggle-debug-on-quit'?

It does, but I don't know anything about debugging elisp, so I gave
the information I had.

> Can you reproduce the inf-loop by putting ...

> (require 'rfc2047) (rfc2047-parse-and-decode "=?utf-8?q? cell
> quantity and ?=") (string-match rfc2047-encoded-word-regexp
> "=?utf-8?q? cell quantity and ?=")

> ... into the scratch buffer and hitting `C-j' after each line?

No.

That gives me:

 (require 'rfc2047)
 rfc2047

 (rfc2047-parse-and-decode "=?utf-8?q? cell quantity and ?=")
 " cell quantity and "

 (string-match rfc2047-encoded-word-regexp "=?utf-8?q? cell quantity and ?=")
 0

Which I guess means that it works as advertised (sometimes).

> For me, it looks like an XEmacs bug. Probably you should use `M-x
> report-xemacs-bug RET' and report it.

Thanks.


  Best regards,

    Adam

-- 
 "Snurra min jord igen"                                       Adam Sjøgren
                                                         asjo@koldfront.dk