Mailing List Archive: Verifying GnuPG-signed documents with PGP 5.0i

Verifying GnuPG-signed documents with PGP 5.0i

Jun 15, 1999, 10:03 AM

Post #1 of 7 (3074 views)

Greetings,
I have tried to verify GnuPG-signed documents with PGP 5.0i,
(the newest available international version of PGP for unix
computers), with the following results:

- I could import my GnuPG-generated public-key block into PGP
with no problem.

- When I clearsign the document, PGP verifies the signature
with no problem.

- However, when I merely do 'gpg -s' or 'gpg -a -s', then PGP
dies with signal 11 (segmentation violation).

Of course, part of the blame must be placed on PGP 5.0i (even
with incorrect data the program should not segfault), but
since GnuPG claims to be compatible with PGP, it appears to
me that GnuPG needs some more work in this area also.

Here is a message which PGP 5.0i segfaults on

-----BEGIN PGP MESSAGE-----
Version: GnuPG v0.9.7 (GNU/Linux)
Comment: For info see http://www.gnupg.org

owJ4nJvAy8zAJGi/TnTjXVcWZ8Y1fUnsBekFJanFJeapP4VDMlIV8pOyUpNLMsuA
rDSFkozMYgWQrAKIzldIzkhNzgaKJpYoeCoUpeam5ialFgH5qQp5qeUKBYnFxQUZ
RYnFqVyJeSkg9WWpRZlplTANyYl5CqXFqQoB7gEIwxIVijPT8xJLSotS9bg67JlZ
GUAugblQkCn4B8P8fGujSg6hhZorWtu3Ln5/r1altd+GYX6CS4bzPmmRO5MT9Gwz
NaZkFk2ZdhoAeFVRFQ==
=49nM
-----END PGP MESSAGE-----

and here is the public key corresponding to the signature of that
message

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v0.9.7 (GNU/Linux)
Comment: For info see http://www.gnupg.org
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=9Zxu
-----END PGP PUBLIC KEY BLOCK-----

norbert@quill:~/crypto > pgpv pgptest1.asc

Received signal 11.

May blessings from the eternal God surprise and overtake you!
Norbert.

--
New snail mail address: Norbert Bollow, Weidlistr. 18, CH-8624 Gruet.

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

ftobin at bigfoot

Jun 15, 1999, 11:37 PM

Post #2 of 7 (3025 views)

Permalink

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

nb@thinkcoach.com, at 19:03 on Tue, 15 Jun 1999, wrote:

> - However, when I merely do 'gpg -s' or 'gpg -a -s', then PGP
> dies with signal 11 (segmentation violation).

There are two options that have to be passed to GPG in order for it to
create PGP5 readable signatures (at least from my experience).

- --compress-algo 1
- --force-v3-sigs

You can place these in your ~/.gnupg/options file if you so desire
(without the --'s).

- --
Frank Tobin "To learn what is good and what is to be
http://www.bigfoot.com/~ftobin valued, those truths which cannot be
shaken or changed." Myst: The Book of Atrus
FreeBSD: The Power To Serve

PGPenvelope = GPG and PGP5 + Pine PGP: 4F86 3BBB A816 6F0A 340F
http://www.bigfoot.com/~ftobin/resources.html 6003 56FF D10A 260C 4FA3
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.5 (FreeBSD)
Comment: PGPEnvelope - http://www.bigfoot.com/~ftobin/resources.html

iD8DBQE3Z0ZGVv/RCiYMT6MRAv6HAJ0XShvqIQQMCQDTiR5sY++0XKYBoQCfcrwe
UAn0HY6s7h1fCDqXxdqy/SE=
=ujaM
-----END PGP SIGNATURE-----

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

wk at isil

Jun 16, 1999, 12:38 AM

Post #3 of 7 (3038 views)

Permalink

nb@thinkcoach.com writes:

> I have tried to verify GnuPG-signed documents with PGP 5.0i,

5.0i is very old and very buggy.

> (the newest available international version of PGP for unix

Hmmm, it is nearly 2 years old.

> - However, when I merely do 'gpg -s' or 'gpg -a -s', then PGP
> dies with signal 11 (segmentation violation).

Did you keep the --force-v3-sigs in your options file?

> since GnuPG claims to be compatible with PGP, it appears to
> me that GnuPG needs some more work in this area also.

Well, it does not claim to be compatible with PGP 5 or whatever but
to be an OpenPGP implementaion. PGP 5.x is far away to be in
compliance with RFC2440 and 6 has still some problems with OpenPGP.

Your message uses compression algorithm 2 which pgp 5 does not understand,
if you would have encrypted to someone elses key (created by pgp5.0), gpg
would have use algorithm 1 because algo 2 is probably not listed in the
recipients compress algo preferences. But if you don't encrypt, gpg
uses a default algorithm. Compress algorithm yields better compression
rates than 1.

You can override this with the option --compress-algo 1 or don't use
compression at all (-z 0)

--
Werner Koch at guug.de www.gnupg.org keyid 621CC013

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

nb at thinkcoach

Jun 16, 1999, 2:10 AM

Post #4 of 7 (3038 views)

Permalink

I wrote:

> > I have tried to verify GnuPG-signed documents with PGP 5.0i,
> > (the newest available international version of PGP for unix

Werner Koch <wk@isil.d.shuttle.de> replied:

> 5.0i is very old and very buggy.

You are so right. I'm very glad that GnuPG is available now!!!

> > - However, when I merely do 'gpg -s' or 'gpg -a -s', then PGP
> > dies with signal 11 (segmentation violation).
>
> Did you keep the --force-v3-sigs in your options file?

Yes.

> Your message uses compression algorithm 2 which pgp 5 does not understand,
> if you would have encrypted to someone elses key (created by pgp5.0), gpg
> would have use algorithm 1 because algo 2 is probably not listed in the
> recipients compress algo preferences. But if you don't encrypt, gpg
> uses a default algorithm. Compress algorithm yields better compression
> rates than 1.

Many thanks for this explanation!

However I'm wondering why don't you make algo 1 the default algorithm,
for better compatibility with PGP?

> You can override this with the option --compress-algo 1 or don't use
> compression at all (-z 0)

I have now put

compress-algo 1

in my ~/.gnupg/options file; however I'm wondering whether this might
have the undesired side-effect of using algo 1 also when encrypting
to a recipient who is known to use software that can handle algo 2 also.

May blessings from the eternal God surprise and overtake you!
Norbert.

--
New snail mail address: Norbert Bollow, Weidlistr. 18, CH-8624 Gruet.

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

wk at isil

Jun 16, 1999, 5:38 AM

Post #5 of 7 (3028 views)

Permalink

nb@thinkcoach.com writes:

> However I'm wondering why don't you make algo 1 the default algorithm,
> for better compatibility with PGP?

Maybe I can do this if a file gets only signed. The problem with
algorithm 1 is, that it is not documented and there is only a
undocumented feature in zlib to allow this old algorithm. Algorithm 2
is well defined in RFC195{0,1}

--
Werner Koch at guug.de www.gnupg.org keyid 621CC013

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

ftobin at bigfoot

Jun 16, 1999, 5:53 AM

Post #6 of 7 (3032 views)

Permalink

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

nb@thinkcoach.com, at 11:10 on Wed, 16 Jun 1999, wrote:

> in my ~/.gnupg/options file; however I'm wondering whether this might
> have the undesired side-effect of using algo 1 also when encrypting
> to a recipient who is known to use software that can handle algo 2 also.

You don't have it both ways :) You can't be guaranteed what software your
recipient is using; therefore, you have to decide one way or the other;
be safe and use algo 1, which has worse performance, or take a risk the
person won't be able to verify the signature correctly, and use compress
algorithm 2.

- --
Frank Tobin "To learn what is good and what is to be
http://www.bigfoot.com/~ftobin valued, those truths which cannot be
shaken or changed." Myst: The Book of Atrus
FreeBSD: The Power To Serve

PGPenvelope = GPG and PGP5 + Pine PGP: 4F86 3BBB A816 6F0A 340F
http://www.bigfoot.com/~ftobin/resources.html 6003 56FF D10A 260C 4FA3
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.5 (FreeBSD)
Comment: PGPEnvelope - http://www.bigfoot.com/~ftobin/resources.html

iD8DBQE3Z55dVv/RCiYMT6MRArxOAJ0cB+8JJdasmJNKO1AuR2jeuX0i4gCdH/Aw
3gsESMtwMGZ2P6e18PTPfRo=
=453B
-----END PGP SIGNATURE-----

Re: Verifying GnuPG-signed documents with PGP 5.0i [ In reply to ]

nb at thinkcoach

Jun 16, 1999, 6:32 AM

Post #7 of 7 (3027 views)

Permalink

Frank Tobin <ftobin@bigfoot.com> wrote:

> nb@thinkcoach.com, at 11:10 on Wed, 16 Jun 1999, wrote:
>
> > in my ~/.gnupg/options file; however I'm wondering whether this might
> > have the undesired side-effect of using algo 1 also when encrypting
> > to a recipient who is known to use software that can handle algo 2 also.
>
> You don't have it both ways :) You can't be guaranteed what software your
> recipient is using; therefore, you have to decide one way or the other;
> be safe and use algo 1, which has worse performance, or take a risk the
> person won't be able to verify the signature correctly, and use compress
> algorithm 2.

What I'm looking for is an option (or improved defaults) that would
allow me to use algo 1 when signing only, but which would not affect
the strategy for determining the compression algo when encrypting.

BTW, it's not just a matter of some people not being able to verify
signatures.... if their software is not able to uncompress the
data I'm giving them, they can't get at the signed data at all.

May blessings from the eternal God surprise and overtake you!
Norbert.

--
New snail mail address: Norbert Bollow, Weidlistr. 18, CH-8624 Gruet.