[102] in Kerberos-V5-bugs
Decode of "SEQUENCE OF" in "SEQUENCE"
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Wed Jan 9 15:07:14 1991
Date: Wed, 9 Jan 91 15:04:48 EST
From: PAU@IBM.COM
To: krb5-testers@MIT.EDU
In the function KRB5_KDC_REQ2krb_kdc_req, a complaint is made about
the way ASN compiler handle the "etype" field. In KRB5-types.h, the compiler
fails to make it an array. However, I also feel the function
decode_KRB5_KDC_REQ_BODY is handling this field wrong. It looks like the
function is trying to determine the number of "etype" elements from the
cardinal number of KRB5_KDC_REQ_BODY. The C-statements are :
(lines 8931 - 8933 in my KRB5-types.c file)
PE xx_pe = prim2seq(pe) ;
n_element_KRB5_8 = xx_pe -> pe_cardinal > 0 ? xx_pe -> pe_cardinal : 0 ;
If I understand the isode lib correctly, then prim2seq is a macro defined
to be prim2set in psap.h and xx_pe should be equal to pe. Therefore, xx_pe
will not point to the PE of etype.
Could any person tell me if I am right or wrong ?
I checked my posy.c file. It already has the fix from posy-patch.
Regrads, Pau-Chen
