bug#2412: spurious "Sign failed" errors

bug#2412: spurious "Sign failed" errors

[Sven Joachim]

[-- Attachment #2: Type: text/plain, Size: 1290 bytes --]

Gnus v5.13
GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.12.11)
 of 2008-11-30 on turtle, modified by Debian
200 news.motzarella.org InterNetNews NNRP server INN 2.4.5 ready (posting ok).

I'm experiencing painful, but not always reproducible problems with
sending or previewing signed mails.  Often I get an error "Sign failed"
and a backtrace like this:

,----
| Debugger entered--Lisp error: (error "Sign failed")
|   signal(error ("Sign failed"))
|   byte-code("<snipped>" [mml2015-epg-secret-key-id-list error password-cache-remove signal] 4)
|   mml2015-epg-sign((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
|   mml2015-sign((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
|   mml-pgpmime-sign-buffer((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
|   mml-generate-mime-1((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
|   mml-generate-mime()
|   message-encode-message-body()
|   mml-to-mime()
|   mml-preview(nil)
|   call-interactively(mml-preview nil nil)
`----

If this happens during sending, the message actually is signed and sent,
but no copy lands in my archive folder. :-(

I'm using gpg-agent with pinentry-gtk2 for the passphrase.



[-- Attachment #3: User settings --]
[-- Type: application/emacs-lisp, Size: 3572 bytes --]

bug#2412: spurious "Sign failed" errors

[Chong Yidong]

Hi Daiki,

This may be a bug in epg, could you take a look?  This is Bug#2412 in
the Emacs bug tracker.  Thanks.


Sven Joachim <svenjoac@gmx.de> wrote:

> Gnus v5.13
> GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.12.11)
>  of 2008-11-30 on turtle, modified by Debian
> 200 news.motzarella.org InterNetNews NNRP server INN 2.4.5 ready
> (posting ok).

> I'm experiencing painful, but not always reproducible problems with
> sending or previewing signed mails.  Often I get an error "Sign failed"
> and a backtrace like this:

> ,----
> | Debugger entered--Lisp error: (error "Sign failed")
> |   signal(error ("Sign failed"))
> |   byte-code("<snipped>" [mml2015-epg-secret-key-id-list error password-cache-remove signal] 4)
> |   mml2015-epg-sign((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
> |   mml2015-sign((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
> |   mml-pgpmime-sign-buffer((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
> |   mml-generate-mime-1((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))
> |   mml-generate-mime()
> |   message-encode-message-body()
> |   mml-to-mime()
> |   mml-preview(nil)
> |   call-interactively(mml-preview nil nil)
> `----
>
> If this happens during sending, the message actually is signed and sent,
> but no copy lands in my archive folder. :-(
>
> I'm using gpg-agent with pinentry-gtk2 for the passphrase.

bug#2412: spurious "Sign failed" errors

[Daiki Ueno]

Chong Yidong <cyd@stupidchicken.com> writes:

> Hi Daiki,
>
> This may be a bug in epg, could you take a look?  This is Bug#2412 in
> the Emacs bug tracker.  Thanks.

Thanks for noticing this.

> Sven Joachim <svenjoac@gmx.de> wrote:
>> | Debugger entered--Lisp error: (error "Sign failed")
>> |   signal(error ("Sign failed"))
>> |   byte-code("<snipped>" [mml2015-epg-secret-key-id-list error password-cache-remove signal] 4)
>> |   mml2015-epg-sign((part (sign . "pgpmime") (tag-location . 186) (contents . "Nur ein Test.\n")))

It looks like epg is signalling the error.

Sven, could you please try (setq epg-debug t) and show me the contents
of " *epg-debug*" buffer next time you see the error?

Regards,
-- 
Daiki Ueno

bug#2412: spurious "Sign failed" errors

[Chong Yidong]

Hi Sven,

Are you still experiencing these problems?  If so, please do as Daiki
requested:

> could you please try (setq epg-debug t) and show me the contents of
> " *epg-debug*" buffer next time you see the error?

bug#2412: spurious "Sign failed" errors

[Sven Joachim]

On 2009-04-16 06:27 +0200, Chong Yidong wrote:

> Hi Sven,
>
> Are you still experiencing these problems?

Sorry for not replying sooner.  Yes, they still persist. :-(

>  If so, please do as Daiki requested:
>
>> could you please try (setq epg-debug t) and show me the contents of
>> " *epg-debug*" buffer next time you see the error?

Looks innocent enough to me:

,----
| gpg --no-tty --status-fd 1 --yes --use-agent --command-fd 0 --armor --textmode --output /tmp/epg-output2077H-b --detach-sign
| [GNUPG:] USERID_HINT ECB679C787E9AABE Sven Joachim (Debian User) <svenjoac@gmx.de>
| [GNUPG:] NEED_PASSPHRASE ECB679C787E9AABE ECB679C787E9AABE 17 0
| [GNUPG:] GOOD_PASSPHRASE
| [GNUPG:] BEGIN_SIGNING
| [GNUPG:] SIG_CREATED D 17 2 01 1239906974 9AACBC0C69283380715D3788ECB679C787E9AABE
| gpg --no-tty --status-fd 1 --yes --use-agent --command-fd 0 --armor --textmode --output /tmp/epg-output2077UIi --detach-sign
| [GNUPG:] USERID_HINT ECB679C787E9AABE Sven Joachim (Debian User) <svenjoac@gmx.de>
| [GNUPG:] NEED_PASSPHRASE ECB679C787E9AABE ECB679C787E9AABE 17 0
| [GNUPG:] GOOD_PASSPHRASE
| [GNUPG:] BEGIN_SIGNING
| [GNUPG:] SIG_CREATED D 17 2 01 1239906984 9AACBC0C69283380715D3788ECB679C787E9AABE
`----

IIRC, the first try did show the preview of the test mail, the second
one gave the "Sign failed" error.

Sven

bug#2412: spurious "Sign failed" errors

[Daiki Ueno]

[-- Attachment #1: Type: text/plain, Size: 508 bytes --]

Hi,

Sven Joachim <svenjoac@gmx.de> writes:

>>> could you please try (setq epg-debug t) and show me the contents of
>>> " *epg-debug*" buffer next time you see the error?

> Looks innocent enough to me:

> | [GNUPG:] SIG_CREATED D 17 2 01 1239906974 9AACBC0C69283380715D3788ECB679C787E9AABE
> | [GNUPG:] SIG_CREATED D 17 2 01 1239906984 9AACBC0C69283380715D3788ECB679C787E9AABE

Thanks.  It looks innocent for me too, but now I can suspect a timing
problem.  Could you please try the patch attached below?


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: epg.el.diff --]
[-- Type: text/x-diff, Size: 533 bytes --]

diff --git a/lisp/epg.el b/lisp/epg.el
index e96258f..3c80ef9 100644
--- a/lisp/epg.el
+++ b/lisp/epg.el
@@ -1175,7 +1175,8 @@ This function is for internal use only."
 (defun epg-wait-for-completion (context)
   "Wait until the `epg-gpg-program' process completes."
   (while (eq (process-status (epg-context-process context)) 'run)
-    (accept-process-output (epg-context-process context) 1)))
+    (accept-process-output (epg-context-process context) 1))
+  (sleep-for 0.1))
 
 (defun epg-reset (context)
   "Reset the CONTEXT."

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


Regards,
-- 
Daiki Ueno

bug#2412: spurious "Sign failed" errors

[Sven Joachim]

On 2009-05-03 07:43 +0200, Daiki Ueno wrote:

> Sven Joachim <svenjoac@gmx.de> writes:
>
>>>> could you please try (setq epg-debug t) and show me the contents of
>>>> " *epg-debug*" buffer next time you see the error?
>
>> Looks innocent enough to me:
>
>> | [GNUPG:] SIG_CREATED D 17 2 01 1239906974 9AACBC0C69283380715D3788ECB679C787E9AABE
>> | [GNUPG:] SIG_CREATED D 17 2 01 1239906984 9AACBC0C69283380715D3788ECB679C787E9AABE
>
> Thanks.  It looks innocent for me too, but now I can suspect a timing
> problem.  Could you please try the patch attached below?
>
>
> diff --git a/lisp/epg.el b/lisp/epg.el
> index e96258f..3c80ef9 100644
> --- a/lisp/epg.el
> +++ b/lisp/epg.el
> @@ -1175,7 +1175,8 @@ This function is for internal use only."
>  (defun epg-wait-for-completion (context)
>    "Wait until the `epg-gpg-program' process completes."
>    (while (eq (process-status (epg-context-process context)) 'run)
> -    (accept-process-output (epg-context-process context) 1)))
> +    (accept-process-output (epg-context-process context) 1))
> +  (sleep-for 0.1))
>  
>  (defun epg-reset (context)
>    "Reset the CONTEXT."

That seems to do the trick.  At least, I had no problem in ~15 attempts
to preview/send signed messages, whereas without your patch the "Sign
failed" error would happen about every second time.

Cheers,
       Sven