From: Tomi Ollila <tomi.ollila@iki.fi>
To: Jani Nikula <jani@nikula.org>, notmuch@notmuchmail.org
Subject: Re: [PATCH v2 0/3] cli: clean up gmime 2.4 vs. 2.6 conditional build
In-Reply-To: <m2ppyggdh0.fsf@guru.guru-group.fi>
References: <cover.1364651359.git.jani@nikula.org>
	<m2ppyggdh0.fsf@guru.guru-group.fi>
User-Agent: Notmuch/0.15.2+50~g43fdda5 (http://notmuchmail.org) Emacs/24.2.1
	(x86_64-unknown-linux-gnu)
Date: Sun, 31 Mar 2013 12:27:51 +0300
Message-ID: <m2mwtjhqaw.fsf@guru.guru-group.fi>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: list

On Sun, Mar 31 2013, Tomi Ollila <tomi.ollila@iki.fi> wrote:

> On Sat, Mar 30 2013, Jani Nikula <jani@nikula.org> wrote:
>
>> This is v2 of [1]. Added comments per David's request, and while at it,
>> added a third patch to conform the existing conditional build in notmuch
>> show to the same style. The whole series should have no functional
>> changes, and thus v2 should have no functional changes since v1. ;)
>>
>> I have not tested this on gmime 2.4.
>
> I have gmime 2.4(.24) on this system. The code compiles and
> exactly the same set of tests pass/fail as without this patch
> (unfortunately this system cannot cope with just some of the
> crypto patches, so I cannot give absolute guarantee:

Ok, the failure cases below are just those which are changed in
this patch. So I had to dig further...

... actually this
  'PASS   emacs delivery of encrypted message with attachment'
fails to deliver the message.

After the following change:

--- a/test/test-lib.sh
+++ b/test/test-lib.sh
@@ -442,7 +442,8 @@ emacs_deliver_message ()
 
     test_emacs \
        "(let ((message-send-mail-function 'message-smtpmail-send-it)
-              (smtpmail-smtp-server \"localhost\")
+              (smtpmail-smtp-server 'local)
+              ;;(smtpmail-smtp-server \"localhost\")
               (smtpmail-smtp-service \"25025\"))
           (notmuch-hello)
           (notmuch-mua-mail)

The failured cases below PASSes.

(I've stumbled the same problem a few cases before but haven't 
figured out a good patch so far...)

>
> crypto: Testing PGP/MIME signature verification and decryption
>  PASS   emacs delivery of signed message
>  PASS   signature verification
>  PASS   signature verification with full owner trust
>  PASS   signature verification with signer key unavailable
>  PASS   emacs delivery of encrypted message with attachment
>  FAIL   decryption, --format=text
>  FAIL   decryption, --format=json
>  FAIL   decryption, --format=json, --part=4
>  FAIL   decrypt attachment (--part=5 --format=raw)
>  FAIL   decryption failure with missing key
>  PASS   emacs delivery of encrypted + signed message
>  FAIL   decryption + signature verification
>  FAIL   reply to encrypted message
>  FAIL   signature verification with revoked key
>
> anyway, as said, exactly the same set fails without these pathes
>
> The reason for failure is most probably in my system and not something
> w/ notmuch crypto support -- If I run `emacs foo.org.gpg` this just
> hangs waiting me to give passphrase in (invisible?) curses request)
>
> So, bottom line: LGTM.
>
>> BR,
>> Jani.
>
> Tomi
>
>
>> [1] id:a9f12ba474fc51df71e2fd2b7a20a8d101729c6e.1362319765.git.jani@nikula.org
>>
>> Jani Nikula (3):
>>   cli: crypto: abstract gpg context creation for clarity
>>   cli: mime node: abstract decryption and signature verification
>>   cli: conform to same conditional build style as elsewhere in
>>     notmuch-show
>>
>>  crypto.c       |   66 ++++++++++++-----
>>  mime-node.c    |  218 ++++++++++++++++++++++++++++++++++++--------------------
>>  notmuch-show.c |   46 ++++++------
>>  3 files changed, 215 insertions(+), 115 deletions(-)
>>
>> -- 
>> 1.7.10.4