sig changes from "--" to "- --"

[DenverD]

with reference to this thread
http://forums.opensuse.org/english/get-help-here/network-internet/448586-nntp-web-problem.html

several nntp users on the openSUSE forum have discussed an unexpected
change in the line preceding the nntp signature from "--" to "- --"
when using several different Linux mail/usenet clients and signing the
message with OpenPGP..

while it doesn't seem to be a problem with the clients (and i highly
doubt it to be an openSUSE problem) i failed in trying to google and
learn if:

1. it is a bug in OpenPGP
2. if a bug, if it has been reported
3. the extra "- " just what the nntp/pgp RFC calls for
4. it is a bug in PGP
5. it is a bug in the interaction between nntp and PGP/OpenPGP
6. ???

any event, if anyone can tell me anything about where to ask or
report, i'd appreciate it..

sorry if i overlooked an easily found answer,

DenverD

[Per Jessen]

Post by DenverD
with reference to this thread

http://forums.opensuse.org/english/get-help-here/network-internet/448586-nntp-web-problem.html

Post by DenverD
several nntp users on the openSUSE forum have discussed an unexpected
change in the line preceding the nntp signature from "--" to "- --"
when using several different Linux mail/usenet clients and signing the
message with OpenPGP..
while it doesn't seem to be a problem with the clients (and i highly
doubt it to be an openSUSE problem) i failed in trying to google and
1. it is a bug in OpenPGP
2. if a bug, if it has been reported
3. the extra "- " just what the nntp/pgp RFC calls for

I can't imagine openpgp nor gnupg needing or wanting to change the
content of the message being signed.

[Patrick Shanahan]

Post by DenverD
with reference to this thread
http://forums.opensuse.org/english/get-help-here/network-internet/448586-nntp-web-problem.html
several nntp users on the openSUSE forum have discussed an unexpected
change in the line preceding the nntp signature from "--" to "- --"
when using several different Linux mail/usenet clients and signing the
message with OpenPGP..

I cannot answer whether it is an openssh bug but it has been noticed and
appears *only* with users of crypto signing. btw, it is changing from
"-- " (dash, dash, space) rather than "--".

[DenverD]

Post by Patrick Shanahan

Post by DenverD
with reference to this thread
http://forums.opensuse.org/english/get-help-here/network-internet/448586-nntp-web-problem.html
several nntp users on the openSUSE forum have discussed an unexpected
change in the line preceding the nntp signature from "--" to "- --"
when using several different Linux mail/usenet clients and signing the
message with OpenPGP..

I cannot answer whether it is an openssh bug but it has been noticed and
appears *only* with users of crypto signing. btw, it is changing from
"-- " (dash, dash, space) rather than "--".

absolutely....of course you are correct it is from "-- " to "- --"
dropping (or moving) the original trailing space and adding a new
leading dash space..

DenverD

[Patrick Shanahan]

Post by DenverD

Post by Patrick Shanahan
I cannot answer whether it is an openssh bug but it has been noticed and
appears *only* with users of crypto signing. btw, it is changing from
"-- " (dash, dash, space) rather than "--".

absolutely....of course you are correct it is from "-- " to "- --"
dropping (or moving) the original trailing space and adding a new
leading dash space..

and appears to *only* affect "in-line" signing, not attachments.

[Per Jessen]

Post by Patrick Shanahan

Post by DenverD

Post by Patrick Shanahan
I cannot answer whether it is an openssh bug but it has been
noticed and
appears *only* with users of crypto signing. btw, it is changing
from "-- " (dash, dash, space) rather than "--".

absolutely....of course you are correct it is from "-- " to "- --"
dropping (or moving) the original trailing space and adding a new
leading dash space..

and appears to *only* affect "in-line" signing, not attachments.

Is the signature still valid?

[Patrick Shanahan]

Post by Per Jessen

Post by Patrick Shanahan
and appears to *only* affect "in-line" signing, not attachments.

Is the signature still valid?

I believe so. I no longer bother with crypto signing since six or seven
years ago so this is based on old memory. Carlos ER used to include
inline signing when he used pine comes to mind.

I cannot see that signing verification is of much use except with
contract and/or financial dealings.

[DenverD]

Post by Per Jessen

Post by Patrick Shanahan

Post by DenverD

Post by Patrick Shanahan
I cannot answer whether it is an openssh bug but it has been
noticed and
appears *only* with users of crypto signing. btw, it is changing
from "-- " (dash, dash, space) rather than "--".

absolutely....of course you are correct it is from "-- " to "- --"
dropping (or moving) the original trailing space and adding a new
leading dash space..

and appears to *only* affect "in-line" signing, not attachments.

Is the signature still valid?

yes...well, it appears to be...it is displayed in my client as signed
by [whoever] with not a hint of an error anywhere..

DenverD

[Per Jessen]

Post by Patrick Shanahan

Post by Per Jessen

Post by Patrick Shanahan
and appears to *only* affect "in-line" signing, not attachments.

Is the signature still valid?

I believe so.

DenverD says it appears as "signed", but can anyone confirm that it also
validates? If it's a valid signature it means the email was altered
before it was signed. If it's not, it was altered afterwards.

[Carlos E. R.]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Post by Patrick Shanahan

Post by Per Jessen

Post by Patrick Shanahan
and appears to *only* affect "in-line" signing, not attachments.

Is the signature still valid?

I believe so.

It is part of the signing process to convert dash-dash-space to
dash-space-dash-dash-space (on a line and alone).

I know I have read an explanation of why this is done, but I don't
remember where.

Post by Patrick Shanahan
I no longer bother with crypto signing since six or seven
years ago so this is based on old memory. Carlos ER used to include
inline signing when he used pine comes to mind.

I do >:-)

Post by Patrick Shanahan
I cannot see that signing verification is of much use except with
contract and/or financial dealings.

And PGP signing is not used for any of those: they want a system with a
certification authority (and one they trust). PGP is a kind of renegade
thing (that's not the word I want, but it will do).

Me, I use one because somebody was faking emails in one of the lists
(years ago now), and it is no use to turn to signing after the fact.

- --
Cheers,
Carlos E. R.
(from 11.2 x86_64 "Emerald" at Telcontar)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.12 (GNU/Linux)

iEYEARECAAYFAkzHJ5sACgkQtTMYHG2NR9V1/wCeOa52Y7ovpbimdE0SAb2dEi54
cQ8An1IegJqM7JwFpDIbPOv1wGl15y5Q
=OoL2
-----END PGP SIGNATURE-----

[Per Jessen]

Post by Carlos E. R.
It is part of the signing process to convert dash-dash-space to
dash-space-dash-dash-space (on a line and alone).
I know I have read an explanation of why this is done, but I don't
remember where.

It sounds very dodgy for the contents to be altered by the signing
program.

Post by Carlos E. R.

Post by Patrick Shanahan
I cannot see that signing verification is of much use except with
contract and/or financial dealings.

And PGP signing is not used for any of those: they want a system with
a certification authority (and one they trust). PGP is a kind of
renegade thing (that's not the word I want, but it will do).

Yes and no - it's all about trust, and in the end you've got to trust
someone. There's nothing "renegade" about e.g. gnupg, it's development
was even funded by two Federal German Ministries.

http://de.wikipedia.org/wiki/GNU_Privacy_Guard

[Carlos E. R.]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Post by Per Jessen

Post by Carlos E. R.
It is part of the signing process to convert dash-dash-space to
dash-space-dash-dash-space (on a line and alone).
I know I have read an explanation of why this is done, but I don't
remember where.

It sounds very dodgy for the contents to be altered by the signing
program.

It is part of the standard. Certain letter combinations that are used for
other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and documented,
but I can't remember where.

Post by Per Jessen

Post by Carlos E. R.

Post by Patrick Shanahan
I cannot see that signing verification is of much use except with
contract and/or financial dealings.

And PGP signing is not used for any of those: they want a system with
a certification authority (and one they trust). PGP is a kind of
renegade thing (that's not the word I want, but it will do).

Yes and no - it's all about trust, and in the end you've got to trust
someone. There's nothing "renegade" about e.g. gnupg, it's development
was even funded by two Federal German Ministries.

PGP requires that you exchange keys in person, face to face, with the
person you are going to communicate, so that you know that the keys are
really from that person. If you get the key from a repository but nobody
certifies to you that those keys really belong to whom they say, they are
useless as certification of identity. This is why they make "key signing
parties", like the one the held recently at the opensuse conference.


My email is signed, but how do you know that I'm named that way, and that
I'm not possing as somebody else? The only thing I certify with that
signature is that all mails signed with the same key come from the same
person. Not that I'm really Carlos.


The keys that are used for identification rely on a central organization
that verifies who you are (in person) and then they give you a key, or you
make one and they sign it.

- --
Cheers,
Carlos E. R.
(from 11.2 x86_64 "Emerald" at Telcontar)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.12 (GNU/Linux)

iEYEARECAAYFAkzHW0sACgkQtTMYHG2NR9UfyQCcDLhZJgh0Cr+eOYqaWoMmDJ9h
oa0AmwZN8IZgQ7OMxwufpV5KjrPm4Urg
=3/AG
-----END PGP SIGNATURE-----

[DenverD]

Post by Carlos E. R.

Post by Per Jessen

Post by Carlos E. R.
It is part of the signing process to convert dash-dash-space to
dash-space-dash-dash-space (on a line and alone).
I know I have read an explanation of why this is done, but I don't
remember where.

It sounds very dodgy for the contents to be altered by the signing
program.

It is part of the standard. Certain letter combinations that are used
for other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and
documented, but I can't remember where.

Interesting, I didn't know. Does that mean that gpg-aware email agents
should be decoding this too?

yes, when i look at a pgp-signed message in thunderbird (with the
Enigmail addon) i do not see "- --" instead, i see "-- " (well, i
don't see the trailing space, of course..)

DenverD

[Per Jessen]

Post by Carlos E. R.

Post by Per Jessen

Post by Carlos E. R.
It is part of the signing process to convert dash-dash-space to
dash-space-dash-dash-space (on a line and alone).
I know I have read an explanation of why this is done, but I don't
remember where.

It sounds very dodgy for the contents to be altered by the signing
program.

It is part of the standard. Certain letter combinations that are used
for other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and
documented, but I can't remember where.

Interesting, I didn't know. Does that mean that gpg-aware email agents
should be decoding this too?

Post by Carlos E. R.

Post by Per Jessen

Post by Carlos E. R.

Post by Patrick Shanahan
I cannot see that signing verification is of much use except with
contract and/or financial dealings.

And PGP signing is not used for any of those: they want a system
with a certification authority (and one they trust). PGP is a kind
of renegade thing (that's not the word I want, but it will do).

Yes and no - it's all about trust, and in the end you've got to trust
someone. There's nothing "renegade" about e.g. gnupg, it's
development was even funded by two Federal German Ministries.

PGP requires that you exchange keys in person, face to face, with the
person you are going to communicate, so that you know that the keys
are really from that person.

I'm sure I've heard of a scheme in Germany whereby you were able to use
Deutsche Post as an intermediary - Postident I think it is. I don't
know if it still works.

Post by Carlos E. R.
If you get the key from a repository but nobody certifies to you that
those keys really belong to whom they say, they are useless as
certification of identity. This is why they make "key signing
parties", like the one the held recently at the opensuse conference.

Sure - c't has been running their "Crypto-Kampagne" since 1997.

Post by Carlos E. R.
My email is signed, but how do you know that I'm named that way, and
that I'm not possing as somebody else? The only thing I certify with
that signature is that all mails signed with the same key come from
the same person. Not that I'm really Carlos.

Well, it's not about your _identity_ as such, it's about authentication
of the email.

[Per Jessen]

Post by Carlos E. R.
It is part of the standard. Certain letter combinations that are used
for other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and
documented, but I can't remember where.

Interesting, I didn't know. Does that mean that gpg-aware email
agents should be decoding this too?

I can answer that one myself - yes it does, and e.g. knode does such a
decoding too. I guess that also answers DenverDs question:

6) bug in the email client

[DenverD]

Post by Per Jessen

Post by Carlos E. R.
It is part of the standard. Certain letter combinations that are used
for other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and
documented, but I can't remember where.

Interesting, I didn't know. Does that mean that gpg-aware email
agents should be decoding this too?

I can answer that one myself - yes it does, and e.g. knode does such a
6) bug in the email client

seems (per Carlos) there is no bug in the question, only ignorance of
the way it is..

DenverD

[Jim Henderson]

Post by DenverD
seems (per Carlos) there is no bug in the question, only ignorance of
the way it is..

Seems to me that if a newsreader understands "-- " as the marker for a
sig block and that were inside a GPG/PGP signed message, the newsreader
would ignore the actual signature at the end of the message, so changing
it is necessary for the signature to be interpreted by the software so it
can be validated.

Jim

[Carlos E. R.]

Post by Carlos E. R.
It is part of the standard. Certain letter combinations that are used
for other things have to be defanged (is that the word?). The begin
line-dash-dash means something else for pgg, so the signature can not
start that way or it breaks. This change is intentional and
documented, but I can't remember where.

Interesting, I didn't know. Does that mean that gpg-aware email agents
should be decoding this too?

Yep.

I found the reference to this, by Patrick 3 years ago, who got it from the
mutt mail list:

+++···········
<http://lists.opensuse.org/opensuse/2007-06/msg00841.html>

Post by Carlos E. R.

Why is the <dash><dash><space> signature indicator not display
<dash><space><dash><dash><space>.

This is so that no software deletes the mail's signature including the
gpg signature even by accident. I don't know if it's the official
reason but at least it makes sense... :)

It's required by RFC2440 (the OpenPGP standard). See section 7.1
therein.

············++-

And here it is an official reference:

<http://www.ietf.org/rfc/rfc2440.txt>

+++···········

RFC 2440 OpenPGP Message Format November 1998



7.1. Dash-Escaped Text

The cleartext content of the message must also be dash-escaped.

Dash escaped cleartext is the ordinary cleartext where every line
starting with a dash '-' (0x2D) is prefixed by the sequence dash '-'
(0x2D) and space ' ' (0x20). This prevents the parser from
recognizing armor headers of the cleartext itself. The message digest
is computed using the cleartext itself, not the dash escaped form.

As with binary signatures on text documents, a cleartext signature is
calculated on the text using canonical <CR><LF> line endings. The
line ending (i.e. the <CR><LF>) before the '-----BEGIN PGP
SIGNATURE-----' line that terminates the signed text is not
considered part of the signed text.

Also, any trailing whitespace (spaces, and tabs, 0x09) at the end of
any line is ignored when the cleartext signature is calculated.

············++-

Post by Carlos E. R.
PGP requires that you exchange keys in person, face to face, with the
person you are going to communicate, so that you know that the keys
are really from that person.

I'm sure I've heard of a scheme in Germany whereby you were able to use
Deutsche Post as an intermediary - Postident I think it is. I don't
know if it still works.

That is interesting.

I have not seen such meetings here, in Spain. What we have is, that the
same entity that prints paper money (the mint?) emits pkcs certificates.
or signs them. We go to a web page, do something, we print the page, then
go in person to a government office where an official sees the page, our
identification, our face, and then prints another page with which we can
obtain the electronic certificate, which thus identifies us for things
that need official identification, like paying taxes.

- --
Cheers,
Carlos E. R.
(from 11.2 x86_64 "Emerald" at Telcontar)