Re: Gnus 5.10.6 problems with PGP/MIME (test cases)

[Katsumi Yamaoka <yamaoka@jpl.org>]

>>>>> In <b258afad-11cd-4abc-99ba-89c99615ef53@well-done.deisui.org>
>>>>>	Daiki Ueno wrote:

>>>>>> In <b4mpsmp3e11.fsf@jpl.org>
>>>>>>	Katsumi Yamaoka <yamaoka@jpl.org> wrote:

>> Thank you for the information.  I roughly understood that
>> `gpg --textmode' generates a *text* packet and recipients should
>> treat it as text because it is *text*.

> Though I have not yet read entire discussion, you are talking about
> non-MIME encryption?

No, I thought that it includes not only non-MIME encryption but
also MIME encryption.

> If so, it might be preferable that --textmode is specified only in
> that case.  If --textmode is specified, gpg _itself_ converts line
> ending.

`pgg-gpg-encrypt-region' always inserts CR in every line ending
of text using `(pgg-as-lbt START END 'CRLF BODY)'.  So, if the
--textmode option is not specified, gpg doesn't treat sending
data as text and recipients will see ^Ms in a decrypted message.

However, only recipients who use Gnus will never see ^Ms because
`pgg-gpg-process-region' uses `raw-text-dos' when reading a
decrypted message from a temp file, and furthermore,
`mml2015-pgg-decrypt' converts CRLF to LF.

> On the GNU/Linux system:

> $ ruby -e 'puts("abc\r\ndef\r\nghi\r\n")' > test.txt
> $ gpg -q --encrypt --textmode -r ueno@unixuser.org test.txt
> $ gpg -q --decrypt test.txt.gpg > test.txt
> You need a passphrase to unlock the secret key for
> user: "Daiki Ueno <ueno@unixuser.org>"
> ...
> $ hd test.txt
> 00000000  61 62 63 0a 64 65 66 0a  67 68 69 0a              |abc.def.ghi.|
> 0000000c

> All CRLF are converted to LF.

Yes, that's right.

> This might cause a problem if you send an
> encrypted and signed data in RFC1847 encapsulation (RFC3156, section
> 6.1).  After decryption, it does no longer contain the signed material
> identical as before encryption.

That's indeed a problem.  Do you know the way to verify it using
Gnus?  I don't know what mml command do that.  In addition, do
you have an idea to solve the ^Ms problem that I mentioned above?

> Please add *me* to the Cc: list if you mention changes obviously
> conflicting with my intension such as below.  Otherwise, I will not be
> aware of them.

I'm sorry to have not informed it to you.  I must do it from the
next time.  I thought informing it to you was not necessary
because I found out someone added --textmode in Gnus of the
Emacs version (i.e., Gnus v5.11) before I did it.  I couldn't
discover who did it in Emacs when, though.