Re: [Imap-protocol] authenticate LOGIN question

Tim Showalter <[email protected]> Tue, 31 Oct 2017 19:04:36 -0700
Newsgroups gmane.mail.imap.general
Message-ID <CAByav=hQJDvRjYtTDnU0+B5MfzbpLfhjAvCgxhCvDfjj9jeA3Q@mail.gmail.com>
--===============1000257852182048520==
Content-Type: multipart/alternative; boundary="94eb2c0943ee2bf1dc055ce24bb2"

--94eb2c0943ee2bf1dc055ce24bb2
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

I haven't worked on the Y! IMAP server in several years at this point, and
I can't speak for their current implementation. I know that they have
rewritten a lot of it since I left.

But it is quite possible that it's simply a bug. I don't know which clients
would still support AUTH=3DLOGIN. I would not advise any client to use
AUTH=3DLOGIN, particularly not if PLAIN is available. LOGIN is not a good
mechanism, and is strictly worse than both basic LOGIN and PLAIN. It's just
more round trips for what I recall to be a silly protocol.

Tim


On Tue, Oct 31, 2017 at 3:45 PM, Gene Smith <[email protected]> wrote:

> On 10/31/17 5:02 PM, =C3=81ngel wrote:
>
>> Copying you in CC, Gene as the mailing list thinks this is spam...
>>
>
> Thanks.
>
>> On 2017-10-27 at 01:23 -0400, Gene Smith wrote:
>>
>>> One that I did find that supports it, after providing the uid
>>> and pwd, reports RENEGOTIATING and then openssl crashes.
>>>
>>
>> You have entered in openssl a line beginning with uppercase R
>> This made openssl to renegotiate the TLS connection, which many servers
>> don't support.
>>
>> It is quite similar to that, but openssl s_client does not provide a
>> transparent tunnel for tls connections.
>> This is a common pitfall when testing smtp sending with starttls, as you
>> can't issue "RCPT TO:" (the good news is that the protocol is case
>> insensitive, so "rcpt to:" works).
>>
>> In IMAP the problematic letter will be at an identifier prefix, so the
>> solution is simply to choose a different one.
>>
>> s_client(1) says about this:
>>
>>>   When used interactively (which means neither -quiet nor
>>>   -ign_eof have been given), the session will be renegotiated if the li=
ne
>>>   begins with an R, and if the line begins with a Q or if end of file i=
s
>>>   reached, the connection will be closed down.
>>>
>>
>> Best regards
>>
>
> Thanks for the info on openssl/s_client. I did not know about this
> behavior. You are right. The last thing I entered was the base64 encoded
> password and it begins with R. That explains that. Just to be clear, this
> is a different server that supports authenticate login and not the proble=
m
> yahoo server.
>
> So my question is still why yahoo (and possibly other servers that I
> haven't encountered) won't accept the uid and password (both base64
> encoded) but claim to have authenticate login capability. This is using a
> normal client or openssl and neither uid or password, after base64
> encoding, begin with R.
>
> -gene
>
> _______________________________________________
> Imap-protocol mailing list
> [email protected]
> http://mailman13.u.washington.edu/mailman/listinfo/imap-protocol
>

--94eb2c0943ee2bf1dc055ce24bb2
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">I haven&#39;t worked on the Y! IMAP server in several year=
s at this point, and I can&#39;t speak for their current implementation. I =
know that they have rewritten a lot of it since I left.<div><br></div><div>=
But it is quite possible that it&#39;s simply a bug. I don&#39;t know which=
 clients would still support AUTH=3DLOGIN. I would not advise any client to=
 use AUTH=3DLOGIN, particularly not if PLAIN is available. LOGIN is not a g=
ood mechanism, and is strictly worse than both basic LOGIN and PLAIN. It&#3=
9;s just more round trips for what I recall to be a silly protocol.</div><d=
iv><div><br></div><div>Tim</div><div><br></div></div></div><div class=3D"gm=
ail_extra"><br><div class=3D"gmail_quote">On Tue, Oct 31, 2017 at 3:45 PM, =
Gene Smith <span dir=3D"ltr">&lt;<a href=3D"mailto:[email protected]" targe=
t=3D"_blank">[email protected]</a>&gt;</span> wrote:<br><blockquote class=
=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padd=
ing-left:1ex">On 10/31/17 5:02 PM, =C3=81ngel wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
Copying you in CC, Gene as the mailing list thinks this is spam...<br>
</blockquote>
<br>
Thanks.<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex"><span class=3D"">
On 2017-10-27 at 01:23 -0400, Gene Smith wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
One that I did find that supports it, after providing the uid<br>
and pwd, reports RENEGOTIATING and then openssl crashes.<br>
</blockquote>
<br></span>
You have entered in openssl a line beginning with uppercase R<br>
This made openssl to renegotiate the TLS connection, which many servers<br>
don&#39;t support.<br>
<br>
It is quite similar to that, but openssl s_client does not provide a<br>
transparent tunnel for tls connections.<br>
This is a common pitfall when testing smtp sending with starttls, as you<br=
>
can&#39;t issue &quot;RCPT TO:&quot; (the good news is that the protocol is=
 case<br>
insensitive, so &quot;rcpt to:&quot; works).<br>
<br>
In IMAP the problematic letter will be at an identifier prefix, so the<br>
solution is simply to choose a different one.<br>
<br>
s_client(1) says about this:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
=C2=A0 When used interactively (which means neither -quiet nor<br>
=C2=A0 -ign_eof have been given), the session will be renegotiated if the l=
ine<br>
=C2=A0 begins with an R, and if the line begins with a Q or if end of file =
is<br>
=C2=A0 reached, the connection will be closed down.<br>
</blockquote>
<br>
Best regards<br>
</blockquote>
<br>
Thanks for the info on openssl/s_client. I did not know about this behavior=
. You are right. The last thing I entered was the base64 encoded password a=
nd it begins with R. That explains that. Just to be clear, this is a differ=
ent server that supports authenticate login and not the problem yahoo serve=
r.<br>
<br>
So my question is still why yahoo (and possibly other servers that I haven&=
#39;t encountered) won&#39;t accept the uid and password (both base64 encod=
ed) but claim to have authenticate login capability. This is using a normal=
 client or openssl and neither uid or password, after base64 encoding, begi=
n with R.<span class=3D"HOEnZb"><font color=3D"#888888"><br>
<br>
-gene</font></span><div class=3D"HOEnZb"><div class=3D"h5"><br>
______________________________<wbr>_________________<br>
Imap-protocol mailing list<br>
<a href=3D"mailto:[email protected]" target=3D"_blank">Imap-pr=
[email protected]</a><br>
<a href=3D"http://mailman13.u.washington.edu/mailman/listinfo/imap-protocol=
" rel=3D"noreferrer" target=3D"_blank">http://mailman13.u.washington.<wbr>e=
du/mailman/listinfo/imap-prot<wbr>ocol</a><br>
</div></div></blockquote></div><br></div>

--94eb2c0943ee2bf1dc055ce24bb2--

--===============1000257852182048520==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Imap-protocol mailing list
[email protected]
http://mailman13.u.washington.edu/mailman/listinfo/imap-protocol
--===============1000257852182048520==--