Hi Tomas,
I made some progress and now can dissect CMS and X509 semi-reasonably.
I made some changes to the compiler, please look them over in the CVS log so
they are correct.
Currently I do have two problems I want resolved so I dont have to massage
the generated .c files,
can you have a look at them?
1, the imported types get class/tag == -1. I have to manually massage
the .c files and replace -1 with the proper tag/class to get it working.
No major problem, but can/should say dissect_ber_sequence|set|... be
changed so that if BER_FLAGS_NOOWNTAG is set and
if class==tag==-1 then just ignore checking their values? And just pass
them on to the subdissector helper, no questions asked?
Or do you have a better solution planned?
2, I think I need support for the type ANY in the compiler. This will
require changes/additions to packet-ber.c as well I think.
It might be easiest to implement it as something that is identical to
dissect_ber_octet_string but that it doesnt check the class/tag at all
and just looks at the len. then it, just as the real octet_string one,
creates a subset tvb and passes it on through teh callback, if specified.
To print the ANY fields, just use a FT_BYTES, BASE_NONE hf field and treat
it as an octet string.
apart from that, i managed to generate semicomplete CMS,
InformationFramework and X509AF and tied them into the partial pkinit
dissection in packet-kerberos.c and it works semi-well.
----- Original Message -----
From: "Tomas Kukosa"
Sent: Monday, June 07, 2004 6:43 PM
Subject: [Fwd: Re: [Ethereal-dev] H235 & ASN1 compiler]
Hi Ronnie,
it seems that the mailing list does not work now.
So I send it directly to you so as you can continue with the X509.
Regards,
Tom
