args-out-of-range in gnus-summary-prepare-threads

args-out-of-range in gnus-summary-prepare-threads

[Ralf Angeli]

[-- Attachment #1: Type: text/plain, Size: 6143 bytes --]

Hi everybody,

when I try to enter the group gmane.linux.debian.devel.emacsen by
typing `1 9 RET' in the Group buffer, the following backtrace results:

Debugger entered--Lisp error: (args-out-of-range -69 141)
  delete-region(141 -69)
  (if (> (current-column) 62) (delete-region (point) (- ... ...)) (insert (make-string ... 32)))
  (progn (if (> ... 62) (delete-region ... ...) (insert ...)))
  (progn (insert (format ":%c%c %-16s |%s %s " gnus-tmp-unread gnus-tmp-replied ... gnus-tmp-thread-tree-header-string gnus-tmp-subject-or-nil)) (progn (if ... ... ...)) (insert (format " |%4s |%s\n" gnus-tmp-lines ...)))
  eval((progn (insert (format ":%c%c %-16s |%s %s " gnus-tmp-unread gnus-tmp-replied ... gnus-tmp-thread-tree-header-string gnus-tmp-subject-or-nil)) (progn (if ... ... ...)) (insert (format " |%4s |%s\n" gnus-tmp-lines ...))))
  gnus-summary-prepare-threads((([1209 "\xd0f6\xe624\xdf35\xe17a\xd248\xdcd8\xe349\xd2e1\xd2eb\xd2de\xd2ac\xd123\xd1b3\xd92f\xd8df\xd0a2\xd1b5\xd92f\xd8df\xe2f9\xd24e\xde7d\xe37e\xde54\xe133\xded0\xd0a6\xe5dc\xe2f6\xd22b\xe45d\xd22b\xd24f\xdf5a\xdaf2\xd247\xd9ce\xe3a7\xd247\xd22d\xd26b\xd0a6\xd123\xd86c\xd24b\xdfcd\xd0a6\xe373\xd24b\xe5fe\xe12d\xd24a\xd22a\xdb62\xd0a6\xddb0\xd24b\xde2b\xe6b3\xd24b\xe37e\xd26b\xe52a\xd0ca\xd861\xdfa9\xdebb\xd0cb\xd24f\xe5ad\xd22b\xd24a\xdfcd\xe038\xd123\xdbca\xdaa4\xd973\xe4fc\xd9ab\xddcf\xd24e\xddfe\xd233\xd23d\xd2c1\xd2e3\xd2f3\xd2b9\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1\xd0a1" "\xd1b6\xe069\xe5fc\xd8df\xd24e\xd9cd\xe340\xe57d\xe5a1\xd0a6\xe42f\xdcd4\xd0a6\xdbd0\xdd51\xe538\xdc4b\xde52\xd2e1\xd2eb\xd2de\xd2ac\xd2b8\xd2f3\xe1b4\xe2f6 <delivery@hosyou-r01.mine.nu>" "Sat, 20 Nov 2004 23:51:59 +0900 (JST)" "<5388864.1100962319219.JavaMail.nobody@hr_01_rev_a>" "" 9985 95 "main.gmane.org gmane.linux.debian.devel.emacsen:1209" nil]) ([1210 "Emacsen startup scripts" "J\xf69r\xf74me Marant <jmarant@nerim.net>" "Sun, 21 Nov 2004 16:41:44 +0100" "<87zn1bfbuv.fsf@marant.org>" "" 3210 22 "main.gmane.org gmane.linux.debian.devel.emacsen:1210" nil] ([1211 "Re: Emacsen startup scripts" "Kai Grossjohann <kai@emptydomain.de>" "Sun, 21 Nov 2004 21:17:29 +0100" "<86ekinyn1i.fsf@ketchup.de.uu.net>" "<87zn1bfbuv.fsf@marant.org>" 3974 22 "main.gmane.org gmane.linux.debian.devel.emacsen:1211" nil] ...)) ([1214 "last upgrade 3/2003" "Dan Jacobson <jidanni@jidanni.org>" "Mon, 22 Nov 2004 05:25:34 +0800" "<877joej3n5.fsf@jidanni.org>" "" 2981 8 "main.gmane.org gmane.linux.debian.devel.emacsen:1214" nil] ([1215 "Re: last upgrade 3/2003" "J\xf69r\xf74me Marant <jmarant@nerim.net>" "Mon, 22 Nov 2004 21:35:43 +0100" "<87zn19zko0.fsf@marant.org>" "<877joej3n5.fsf@jidanni.org>" 3384 17 "main.gmane.org gmane.linux.debian.devel.emacsen:1215" nil])) ([1216 "bug #281129 - emacsen-common" "Michael Prokop <deb-bugreport@michael-prokop.at>" "Sat, 27 Nov 2004 15:03:13 +0100" "<2004-11-27T14-51-24@devnull.michael-prokop.at>" "" 4106 37 "main.gmane.org gmane.linux.debian.devel.emacsen:1216" nil]) ([1217 "language settings in emacs" "ray@nabuli.de" "Mon, 06 Dec 2004 23:10:29 +0100" "<20041206221030.34CBDBC30@atmalok>" "" 5406 61 "main.gmane.org gmane.linux.debian.devel.emacsen:1217" nil]) ([1218 "\xd2c9\xd2ea\xd0bc\xd2e0\xd2a8\xd2af\xd2b9\xd2d7\xd2ec\xd2b9\xd247\xd239" "miya@dream-express.co.jp" "9 Dec 2004 11:45:31 +0900" "<20041209024531.29960.qmail@mail.ohs.estore.co.jp>" "" 3456 23 "main.gmane.org gmane.linux.debian.devel.emacsen:1218" nil]) ([1219 "Error while loading 50... due to debian-pkg-add-load-path-item?" "Peter S Galbraith <p.galbraith@globetrotter.net>" "Wed, 15 Dec 2004 15:20:18 -0500" "<20041215202018.A693FD9EB2@mixing.qc.dfo.ca>" "" 3841 26 "main.gmane.org gmane.linux.debian.devel.emacsen:1219" nil] ([1220 "Bug#270388: Error while loading 50... due to debian-pkg-add-load-path-item?" "ucko@debian.org (Aaron M. Ucko)" "Wed, 15 Dec 2004 15:40:43 -0500" "<udlzn0fwb04.fsf@multics.mit.edu>" "<20041215202018.A693FD9EB2@mixing.qc.dfo.ca>" 5118 18 "main.gmane.org gmane.linux.debian.devel.bugs.rc:43145 gmane.linux.debian.devel.emacsen:1220" nil] ... ...)) ([1224 "Found a bteter soultoin" "\"Milo Bergman\" <snfesaje@dartmail.net>" "Fri, 17 Dec 2004 23:09:00 -0600" "<28269757047606.062izb95202pk__26933.3616052338$1103348064$gmane$org@hotmail.com>" "" 8152 143 "main.gmane.org gmane.spam.detected:424473 gmane.linux.debian.devel.emacsen:1224" nil]) ([1225 "find yourself" "\" Ferris\" <IXLLZFKEZK@dds.nl>" "Fri, 17 Dec 2004 23:08:13 -0600" "<QPMSXUVFJPCJJKHMOVJZQWSOP__35870.1889090755$1103349079$gmane$org@bluemail.ch>" "" 5733 51 "main.gmane.org gmane.spam.detected:424534 gmane.linux.debian.devel.emacsen:1225" nil]) ([1226 "CNEL\xa6c6\x99f6\x9ce3\x9c66\x9bfe\xa771\x9a73\xa9c5\x9ddd    14:2:8:176" "\xab50\x9cfa\x9ca4\xa935\x9ae7\xa377\xa6f8 <sales@cnelc.com>" "Sat, 18 Dec 2004 14:02:16 +0800" "<20041218055851.5A63F2E1CA@murphy.debian.org>" "" 14740 243 "main.gmane.org gmane.linux.debian.devel.emacsen:1226 gmane.spam.detected:424680" nil]) ([1227 "V`iagra is out of fashion, find out why" "\"Bernardo Ramey\" <brameyao@modellbahn-liebscher.de>" "Sat, 18 Dec 2004 10:46:38 +0000" "<aaa301c4e4ee$de33db43$590a2545@cid-bz.de>" "" 3877 19 "main.gmane.org gmane.linux.debian.devel.emacsen:1227 gmane.spam.detected:424895" nil])))
  gnus-summary-prepare()
  gnus-summary-read-group-1("nntp+news.gmane.org:gmane.linux.debian.devel.emacsen" t t nil nil nil)
  gnus-summary-read-group("nntp+news.gmane.org:gmane.linux.debian.devel.emacsen" t t nil nil nil nil)
  gnus-group-read-group(nil t)
  gnus-group-select-group(nil)
  gnus-topic-select-group(nil)
  call-interactively(gnus-topic-select-group)


In my .gnus file I set `gnus-summary-line-format' like this:

(setq gnus-summary-line-format
     ":%U%R %-16,16f |%B %s %-62= |%4L |%&user-date;\n")


The cause of the above error seems to be a wrong calculation of the
second argument of `delete-region' in `gnus-spec-tab'.  With the
following change I can enter the group mentioned above:


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Type: text/x-patch, Size: 423 bytes --]

--- gnus-spec.el	02 Nov 2004 11:45:19 +0100	7.8
+++ gnus-spec.el	21 Dec 2004 10:27:40 +0100	
@@ -293,7 +293,8 @@
 	`(progn
 	   (if (> (current-column) ,column)
 	       (delete-region (point)
-			      (- (point) (- (current-column) ,column)))
+			      (+ (point-at-bol) ,column))
+;; 			      (- (point) (- (current-column) ,column)))
 	     (insert (make-string (max (- ,column (current-column)) 0)
 				  ? ))))))))
 

[-- Attachment #3: Type: text/plain, Size: 280 bytes --]


(One has to delete the setting of `gnus-format-specs' in .newsrc.eld
for it to take effect.)

Unfortunately the long subject which caused the original error is not
shortened in this case.  So either the patch is not fully correct or
there is another problem somewhere.

-- 
Ralf

Re: args-out-of-range in gnus-summary-prepare-threads

[Ralf Angeli]

* Ralf Angeli (2004-12-21) writes:

> Unfortunately the long subject which caused the original error is not
> shortened in this case.  So either the patch is not fully correct or
> there is another problem somewhere.

Some additional information: The subject actually is shortened.  One
can check this by entering the group[1] and typing `M-: (goto-char 63)
RET'.  The interesting thing is now that `M-: (current-column) RET' on
this point reports 115.  This is because the characters in the subject
are two columns wide and this is probably also the problem which makes
the original calculation for `delete-region' fail (unless this is a
problem in Emacs).


Footnotes: 
[1] Assuming the posted patch is applied and the posted summary line
format used, use `1 9 RET' currently to have the problematic article
in the first line.

-- 
Ralf

Re: args-out-of-range in gnus-summary-prepare-threads

[Katsumi Yamaoka]

>>>>> In <cq8ud8$k6p$1@sea.gmane.org> Ralf Angeli wrote:

>> Unfortunately the long subject which caused the original error is not
>> shortened in this case.  So either the patch is not fully correct or
>> there is another problem somewhere.

Oh, I could reproduce the same error using your format spec.

> Some additional information: The subject actually is shortened.  One
> can check this by entering the group[1] and typing `M-: (goto-char 63)
> RET'.  The interesting thing is now that `M-: (current-column) RET' on
> this point reports 115.  This is because the characters in the subject
> are two columns wide and this is probably also the problem which makes
> the original calculation for `delete-region' fail (unless this is a
> problem in Emacs).

> Footnotes: 
> [1] Assuming the posted patch is applied and the posted summary line
> format used, use `1 9 RET' currently to have the problematic article
> in the first line.

Let me have time to investigate how does your format spec mean
and how do those functions behave.

Re: args-out-of-range in gnus-summary-prepare-threads

[Katsumi Yamaoka]

>>>>> In <b9yacs74z2z.fsf@jpl.org> Katsumi Yamaoka wrote:

> Let me have time to investigate how does your format spec mean
> and how do those functions behave.

I've fixed it.  Please test.
The error was caused by many wide characters in summary lines,
they often causes an inconsistency between point and column.
Here's an example which causes an error.

(progn
  (pop-to-buffer (get-buffer-create "*testing*"))
  (erase-buffer)
  (insert-char (make-char 'japanese-jisx0208 34 118) 64)
  (let ((gnus-use-correct-string-widths nil))
    (eval (gnus-spec-tab -62))))

Furthermore, if you change 64 into 62, you will see the nil
value for `gnus-use-correct-string-widths', which is the default
in Emacs, is not helpful these days.

The new `gnus-spec-tab' is fast enough to be able to ignore the
value of `gnus-use-correct-string-widths', I think.

Re: args-out-of-range in gnus-summary-prepare-threads

[Ralf Angeli]

* Katsumi Yamaoka (2004-12-22) writes:

>>>>>> In <b9yacs74z2z.fsf@jpl.org> Katsumi Yamaoka wrote:
>
>> Let me have time to investigate how does your format spec mean
>> and how do those functions behave.
>
> I've fixed it.  Please test.

No error and looks much better now.  Thank you!

The columns are still not aligned here but this is due to the wide
characters not being exactly the width of two "narrow" characters on
my system.  So it basically is a font problem.

-- 
Ralf