Non Repudiation error in MIC calculation

Hi,

This is for AS2, specifically Signed, then Encrypted message.

Before I encrypt I simply checksum SHA1 the file with the muitipart content:
EDI data on first part and signature on second. Mime Headers are canonical
<cr><lf> at end of each mime header. Signature is binary because my trading
partner wants it like that. Anyway. There is a whole package with
multipart content, boundaries, etc. I attached it here:
[url]http://www.nabble.com/file/p17994998/mictest.txt[/url] mictest.txt

Ok, if you do a SHA1 over it you get
229585b2927684ac1f8dae4290e3e70d6d9cb53f and if the sha1 is run as binary to
then be injected in a base64 encoder (openssl sha1 -binary
mictest.txt|openssl enc -a, you get: IpWFspJ2hKwfja5CkOPnDW2ctT8=

Though, my trading partner with his WPG says that's not the value of the
MIC, he gets "Uiaz1kOChhlSb/f3SJsmJ/O/8SI=" instead.

Because the message is encrypted (asymetric, so one needs a certificate and
private key to open) , the decryption brings out a quite monolithic unit,
headers are canonical and there is a <cr><lf> after the last boundary too,
¿what then could be the error in the MIC calculation?

Thanks for the help
--
View this message in context: [url]http://www.nabble.com/Non-Repudiation-error-in-MIC-calculation-tp17994998p17994998.html[/url]
Sent from the OpenSSL - User mailing list archive at Nabble.com.

______________________________________________________________________
OpenSSL Project [url]http://www.openssl.org[/url]
User Support Mailing List [email]openssl-users@openssl.org[/email]
Automated List Manager [email]majordomo@openssl.org[/email]