> FYI, with respect to the sequence state crasher, I'm going to revert to my original fix for now (encoding the sequence state as zeros rather than omitting encoding it entirely). Doing the latter requires a more careful investigation than I have the time for right now. So if you could stick with this patch:
Sure, I will try to revert to the original version and then I will apply
this patch. For my purpose, size doesn't really matter (Kerberos does
not have the RADIUS size limitations).
Thank you and best regards,
Alejandro
>
>> diff --git a/mech_eap/util_ordering.c b/mech_eap/util_ordering.c
>> index 71ebfb5..bb7e4d5 100644
>> --- a/mech_eap/util_ordering.c
>> +++ b/mech_eap/util_ordering.c
>> @@ -266,7 +266,10 @@ sequenceExternalize(OM_uint32 *minor,
>> *minor = GSSEAP_WRONG_SIZE;
>> return GSS_S_FAILURE;
>> }
>> - memcpy(*buf, vqueue, sizeof(queue));
>> + if (vqueue != NULL)
>> + memcpy(*buf, vqueue, sizeof(queue));
>> + else
>> + memset(*buf, 0, sizeof(queue));
>> *buf += sizeof(queue);
>> *lenremain -= sizeof(queue);
> That would be great.
>
> -- Luke
>
> On 07/08/2012, at 10:16 PM, Luke Howard <[log in to unmask]> wrote:
>
>> <0001-Don-t-export-sequence-state-for-partial-contexts.patch>
> --
> Luke Howard / [log in to unmask]
> www.padl.com / www.lukehoward.com
|