Inconsistency between `mml2015-use' and `mml2015-verify-function'

Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

[-- Attachment #1: Type: text/plain, Size: 655 bytes --]

I have these lines in my ~/.gnus.el for OpenPGP signature verifying,
however, it doesn't work.

> (require 'epa)
> (require 'epg)
> (require 'epg-config)
> (setq mml2015-use 'epg)

I checked variable `mml2015-use' and it's `epg', however the variable
`mml2015-verify-function' is `mailcrypt-verify'.

Well, this problem have been bothering me for a long time, I guess it's
a configuration issue?  Do I need to change the variable myself?  I
believe it isn't wanted.

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

[-- Attachment #2: Type: application/pgp-signature, Size: 489 bytes --]

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[Daiki Ueno]

XeCycle <xecycle@gmail.com> writes:

> I have these lines in my ~/.gnus.el for OpenPGP signature verifying,
> however, it doesn't work.
>
>> (require 'epa)
>> (require 'epg)
>> (require 'epg-config)
>> (setq mml2015-use 'epg)

Well, how does it not work?  I have the following settings in addition:

~/.gnus:

(setq gnus-buttonized-mime-types '("multipart/signed")
      mm-verify-option 'always)

~/.gnupg/gpg.conf:

keyserver-options auto-key-retrieve
keyserver hkp://keys.gnupg.net

and it successfully verifies your signature.

You could try (setq epg-debug t) and inspect " *epg-debug*" buffer (note
the first letter is a space), if the problem seems to be in EPG itself.

> I checked variable `mml2015-use' and it's `epg', however the variable
> `mml2015-verify-function' is `mailcrypt-verify'.
>
> Well, this problem have been bothering me for a long time, I guess it's
> a configuration issue?  Do I need to change the variable myself?  I
> believe it isn't wanted.

This shouldn't be a problem since the variable is only referred to from
`mml2015-mailcrypt-verify', which wouldn't be called when mml2015-use is
'epg.  I agree that the variable name is a bit misleading though.

Regards,
-- 
Daiki Ueno

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

[-- Attachment #1: Type: text/plain, Size: 6340 bytes --]

Daiki Ueno <ueno@unixuser.org> writes:

> XeCycle <xecycle@gmail.com> writes:
>
>> I have these lines in my ~/.gnus.el for OpenPGP signature verifying,
>> however, it doesn't work.

[...]

> Well, how does it not work?  I have the following settings in addition:
>
> ~/.gnus:
>
> (setq gnus-buttonized-mime-types '("multipart/signed")
>       mm-verify-option 'always)
>
> ~/.gnupg/gpg.conf:
>
> keyserver-options auto-key-retrieve
> keyserver hkp://keys.gnupg.net
>
> and it successfully verifies your signature.

Indeed I have more:

(require 'epa)
(setq mml2015-use 'epg)
(setq epg-debug t)
;; (require 'pgg)
;; (setq mml2015-use 'pgg)
(setq mm-verify-option 'known
      mm-decrypt-option 'known
      gnus-treat-x-pgp-sig t
      gnus-buttonized-mime-types '("multipart/alternative"
				   "multipart/encrypted"
				   "multipart/signed"))

With PGG, it worked fine.  However it'll inform me of PGG being
deprecated in favour of EasyPG, but EasyPG doesn't work for me.

> You could try (setq epg-debug t) and inspect " *epg-debug*" buffer (note
> the first letter is a space), if the problem seems to be in EPG itself.

I did, but it seems useless.

This is the debug trace when I try to display a signed message:

> Debugger entered--Lisp error: (wrong-number-of-arguments #[(target old new) "\303\b	
> #\207" [old new target replace-regexp-in-string] 4] 4)
>   mm-replace-in-string("\ntest encrypt.\n" "\n" "\n" t)
>   byte-code("\306\307\211\211\211\211\x19^[\x1d\310\x1d\311\312\313\x1d@#\206\0\314\306#\211\203.\315\x1eA\314\307\306$\211\204R\x0e\x1f\316\317 !\211\x1e\205J\320\312\x1e@G! \x1e@%+\210\321\322\x1e\"\210\323	\324\325\306$\326\b!\327 \322\330\331\217\210\x1f\316\332\333\f\334\"! !\211\x1e\205\206\320\312\x1e@G! \x1e@%+\210\x1e.\207" [signature part signature-file plain context inhibit-redisplay t nil mm-find-raw-part-by-type get-text-property 0 protocol "application/pgp-signature" mm-find-part-by-type gnus-info "Corrupted" put-text-property throw error mm-replace-in-string "\n" "\n" mm-get-part epg-make-context (byte-code "\304\b	\n#\304\207" [context signature part plain epg-verify-string] 4) ((error (byte-code "\b\305\306\x1a\211\205\0\307\310\v@G\n	\v@%+\210\f@\311=\2036\b\312\313\x1a\211\2051\307\310\v@G\n	\v@%+\210\202N\b\312\314\f!\x1a\211\205L\307\310\v@G\n	\v@%+\210\315\304\v\"\207" [mm-security-handle value parameter handle error gnus-info "Failed" put-text-property 0 quit gnus-details "Quit." mml2015-format-error throw] 7))) epg-verify-result-to-string epg-context-result-for verify ctl handle mm-security-handle value parameter] 7)
>   mml2015-epg-verify(((#<buffer  *mm*<6>> ("text/plain") nil nil nil nil nil nil) (#<buffer  *mm*<7>> ("application/pgp-signature") nil nil nil nil nil nil)) (#("multipart/signed" 0 16 (protocol "application/pgp-signature" micalg "pgp-sha1" boundary "=-=-=" buffer #<buffer  *mm*<5>> from "xecycle@gmail.com" start nil)) (protocol . "application/pgp-signature") (micalg . "pgp-sha1") (boundary . "=-=-=")))
>   mml2015-verify(((#<buffer  *mm*<6>> ("text/plain") nil nil nil nil nil nil) (#<buffer  *mm*<7>> ("application/pgp-signature") nil nil nil nil nil nil)) (#("multipart/signed" 0 16 (protocol "application/pgp-signature" micalg "pgp-sha1" boundary "=-=-=" buffer #<buffer  *mm*<5>> from "xecycle@gmail.com" start nil)) (protocol . "application/pgp-signature") (micalg . "pgp-sha1") (boundary . "=-=-=")))
>   mm-possibly-verify-or-decrypt(((#<buffer  *mm*<6>> ("text/plain") nil nil nil nil nil nil) (#<buffer  *mm*<7>> ("application/pgp-signature") nil nil nil nil nil nil)) (#("multipart/signed" 0 16 (protocol "application/pgp-signature" micalg "pgp-sha1" boundary "=-=-=" buffer #<buffer  *mm*<5>> from "xecycle@gmail.com" start nil)) (protocol . "application/pgp-signature") (micalg . "pgp-sha1") (boundary . "=-=-=")) "xecycle@gmail.com")
>   mm-dissect-multipart((#("multipart/signed" 0 16 (protocol "application/pgp-signature" micalg "pgp-sha1" boundary "=-=-=" buffer #<buffer  *mm*<5>> from "xecycle@gmail.com" start nil)) (protocol . "application/pgp-signature") (micalg . "pgp-sha1") (boundary . "=-=-=")) "xecycle@gmail.com")
>   mm-dissect-buffer(nil t)
>   gnus-display-mime()
>   gnus-article-prepare-display()
>   gnus-article-prepare(15 nil)
>   gnus-summary-display-article(15 nil)
>   gnus-summary-select-article(nil nil pseudo)
>   gnus-summary-scroll-up(1)
>   call-interactively(gnus-summary-scroll-up nil nil)

and this is the *epg-debug*:

> /usr/bin/gpg --no-tty --status-fd 1 --yes --command-fd 0 --armor --textmode --output /tmp/epg-output1142m3f --detach-sign
> [GNUPG:] USERID_HINT 8D718EC47795E591 Carl Lei (XeCycle) <XeCycle@GMail.com>
> [GNUPG:] NEED_PASSPHRASE 8D718EC47795E591 8D718EC47795E591 1 0
> [GNUPG:] GET_HIDDEN passphrase.enter
> [GNUPG:] GOT_IT
> [GNUPG:] GOOD_PASSPHRASE
> [GNUPG:] BEGIN_SIGNING
> [GNUPG:] SIG_CREATED D 1 2 01 1314321893 1FB67F1FD45DF681C84527F78D718EC47795E591
> /usr/bin/gpg --no-tty --status-fd 1 --yes --command-fd 0 --armor --textmode --output /tmp/epg-output1142zBm --detach-sign
> [GNUPG:] USERID_HINT 8D718EC47795E591 Carl Lei (XeCycle) <XeCycle@GMail.com>
> [GNUPG:] NEED_PASSPHRASE 8D718EC47795E591 8D718EC47795E591 1 0
> [GNUPG:] GET_HIDDEN passphrase.enter
> [GNUPG:] GOT_IT
> [GNUPG:] GOOD_PASSPHRASE
> [GNUPG:] BEGIN_SIGNING
> [GNUPG:] SIG_CREATED D 1 2 01 1314321902 1FB67F1FD45DF681C84527F78D718EC47795E591

Seems gpg hadn't even been invoked for verifying signatures.  This is
really a painful problem for me...  I just installed Emacs from git
several weeks ago --- emacs-version "24.0.50.1".

>> I checked variable `mml2015-use' and it's `epg', however the variable
>> `mml2015-verify-function' is `mailcrypt-verify'.
>>
>> Well, this problem have been bothering me for a long time, I guess it's
>> a configuration issue?  Do I need to change the variable myself?  I
>> believe it isn't wanted.
>
> This shouldn't be a problem since the variable is only referred to from
> `mml2015-mailcrypt-verify', which wouldn't be called when mml2015-use is
> 'epg.  I agree that the variable name is a bit misleading though.

Thank you, so this may not be the major problem.

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

[-- Attachment #2: Type: application/pgp-signature, Size: 489 bytes --]

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[Daiki Ueno]

XeCycle <xecycle@gmail.com> writes:

> This is the debug trace when I try to display a signed message:
>
>> Debugger entered--Lisp error: (wrong-number-of-arguments #[(target
>> old new) "\303\b
>> #\207" [old new target replace-regexp-in-string] 4] 4)
>>   mm-replace-in-string("\ntest encrypt.\n" "\n" "\n" t)

So I guess the culprit is a third-party replace-regexp-in-string defined
somewhere.  Could you check where it is defined, with C-h f
replace-regexp-in-string?

Regards,
-- 
Daiki Ueno

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

[-- Attachment #1: Type: text/plain, Size: 995 bytes --]

Daiki Ueno <ueno@unixuser.org> writes:

> XeCycle <xecycle@gmail.com> writes:
>
>> This is the debug trace when I try to display a signed message:
>>
>>> Debugger entered--Lisp error: (wrong-number-of-arguments #[(target
>>> old new) "\303\b
>>> #\207" [old new target replace-regexp-in-string] 4] 4)
>>>   mm-replace-in-string("\ntest encrypt.\n" "\n" "\n" t)
>
> So I guess the culprit is a third-party replace-regexp-in-string defined
> somewhere.  Could you check where it is defined, with C-h f
> replace-regexp-in-string?

It showed these:

> replace-regexp-in-string is a compiled Lisp function in
> `subr.el'.

> (replace-regexp-in-string REGEXP REP STRING &optional FIXEDCASE
> LITERAL SUBEXP START)

I checked the file `subr.el', it is the one that comes with Emacs.
What's up with it?

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

[-- Attachment #2: Type: application/pgp-signature, Size: 489 bytes --]

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[Daiki Ueno]

XeCycle <xecycle@gmail.com> writes:

>>>> Debugger entered--Lisp error: (wrong-number-of-arguments #[(target
>>>> old new) "\303\b
>>>> #\207" [old new target replace-regexp-in-string] 4] 4)
>>>>   mm-replace-in-string("\ntest encrypt.\n" "\n" "\n" t)
>>
>> So I guess the culprit is a third-party replace-regexp-in-string defined
>> somewhere.  Could you check where it is defined, with C-h f
>> replace-regexp-in-string?
>
> It showed these:
>
>> replace-regexp-in-string is a compiled Lisp function in
>> `subr.el'.
>
>> (replace-regexp-in-string REGEXP REP STRING &optional FIXEDCASE
>> LITERAL SUBEXP START)

Well, then how about mm-replace-in-string?  Is mm-util.elc correctly
located under /usr/local/share/emacs/24.0.50/lisp/gnus/ or such?

I mean, mm-replace-in-string seems not working for you.  Could you try:
(mm-replace-in-string "\ntest encrypt.\n" "\n" "\n" t)

Regards,
-- 
Daiki Ueno

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

[-- Attachment #1: Type: text/plain, Size: 787 bytes --]

Daiki Ueno <ueno@unixuser.org> writes:

[...]

> Well, then how about mm-replace-in-string?  Is mm-util.elc correctly
> located under /usr/local/share/emacs/24.0.50/lisp/gnus/ or such?

Seems this is the issue:

mm-replace-in-string is an alias for `replace-in-string' in
`mm-util.el'.

(mm-replace-in-string TARGET OLD NEW)

However, mm-util.elc is under /usr/share/emacs/24.0.50/lisp/gnus/.

> I mean, mm-replace-in-string seems not working for you.  Could you try:
> (mm-replace-in-string "\ntest encrypt.\n" "\n" "\n" t)

So this is annoying --- seems the "&optional literal" is ignored.

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

[-- Attachment #2: Type: application/pgp-signature, Size: 489 bytes --]

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

Daiki Ueno <ueno@unixuser.org> writes:

[...]

> Well, then how about mm-replace-in-string?  Is mm-util.elc correctly
> located under /usr/local/share/emacs/24.0.50/lisp/gnus/ or such?
>
> I mean, mm-replace-in-string seems not working for you.  Could you try:
> (mm-replace-in-string "\ntest encrypt.\n" "\n" "\n" t)

Oops...  I found out where this problem come from.

mm-replace-in-string is an alias for `replace-in-string' in
`mm-util.el'.

(mm-replace-in-string TARGET OLD NEW)

------------------------

replace-in-string is a compiled Lisp function in
`color-theme.el'.

(replace-in-string TARGET OLD NEW)

So...  F**ked up.  Anyway to force mm-replace-in-string use the one from
gnus?

Thank you.

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[Daiki Ueno]

[-- Attachment #1: Type: text/plain, Size: 563 bytes --]

XeCycle <xecycle@gmail.com> writes:

> replace-in-string is a compiled Lisp function in
> `color-theme.el'.
>
> (replace-in-string TARGET OLD NEW)
>
> So...  F**ked up.  Anyway to force mm-replace-in-string use the one from
> gnus?

It looks your color-theme.el is too old.  There is a change not defining
replace-in-string, which is available in 6.6.0 release:

http://bzr.savannah.gnu.org/lh/color-theme/trunk/revision/55

Anyway, I'll install the attached fix in Gnus side as well since mml2015
usage of mm-replace-in-string does not really need the 4th arg.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: mml2015-replace-in-string.diff --]
[-- Type: text/x-diff, Size: 493 bytes --]

=== modified file 'lisp/gnus/mml2015.el'
--- lisp/gnus/mml2015.el	2011-06-30 23:27:10 +0000
+++ lisp/gnus/mml2015.el	2011-08-26 23:27:42 +0000
@@ -923,7 +923,7 @@
 	(mm-set-handle-multipart-parameter
 	 mm-security-handle 'gnus-info "Corrupted")
 	(throw 'error handle))
-      (setq part (mm-replace-in-string part "\n" "\r\n" t)
+      (setq part (mm-replace-in-string part "\n" "\r\n")
 	    signature (mm-get-part signature)
 	    context (epg-make-context))
       (condition-case error


[-- Attachment #3: Type: text/plain, Size: 25 bytes --]


Regards,
-- 
Daiki Ueno

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[XeCycle]

Daiki Ueno <ueno@unixuser.org> writes:

[...]

> It looks your color-theme.el is too old.  There is a change not defining
> replace-in-string, which is available in 6.6.0 release:
>
> http://bzr.savannah.gnu.org/lh/color-theme/trunk/revision/55

Thank you for pointing this out.  But they didn't seem to fix it in the
release, so I just modified it myself.

> Anyway, I'll install the attached fix in Gnus side as well since mml2015
> usage of mm-replace-in-string does not really need the 4th arg.
>
>
> === modified file 'lisp/gnus/mml2015.el'
> --- lisp/gnus/mml2015.el	2011-06-30 23:27:10 +0000
> +++ lisp/gnus/mml2015.el	2011-08-26 23:27:42 +0000
> @@ -923,7 +923,7 @@
>  	(mm-set-handle-multipart-parameter
>  	 mm-security-handle 'gnus-info "Corrupted")
>  	(throw 'error handle))
> -      (setq part (mm-replace-in-string part "\n" "\r\n" t)
> +      (setq part (mm-replace-in-string part "\n" "\r\n")
>  	    signature (mm-get-part signature)
>  	    context (epg-make-context))
>        (condition-case error

After all it was solved, I don't want to change this.  If there's
another application using `replace-in-string' which needs the 4th arg,
this change will be useless.

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591

Re: Inconsistency between `mml2015-use' and `mml2015-verify-function'

[Lars Magne Ingebrigtsen]

XeCycle <xecycle@gmail.com> writes:

> So...  F**ked up.  Anyway to force mm-replace-in-string use the one from
> gnus?

Alter your load path the first thing in your .emacs.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/