Re: [PATCH 2/4] Make _asn1_ordering_* return error values and check them.

[Kurt Roeckx]

On Sat, Apr 26, 2014 at 12:45:18PM +0200, Nikos Mavrogiannopoulos wrote:
> On Sat, Apr 26, 2014 at 12:40 PM, Kurt Roeckx <address@hidden> wrote:
> >> > My current understanding of things is I think that second
> >> > "p = p->right" is wrong.  (The tests actually pass if I
> >> > remove it.)  This second one is also not present in the
> >> > _asn1_ordering_set() function.  I see no good reason why
> >> > we need to move one more entry right.
> >>
> >> It looks like a copy paste error, and given that it is not present in
> >> the _asn1_ordering_set() it is most likely one.
> > I already changed my mind about that one, but I'm not really sure.
> > It seems that that asn1_der_coding() also skips it.  When I don't
> > skip it the _asn1_ordering_set_of() actually seems to try and
> > read too much entries, the if (der_len - counter > 0) check
> > fails, it just continues, and probably does some weird things as
> > result.  valgrind is also complaining.
> 
> As it is not related to the failure we can undo that part, but
> shouldn't then it be in the asn1_ordering_set as well?
> Which test causes the valgrind error?

I think I had 26 errors in gnutls, like mini-record-2.  But I'm
not 100% sure against which version that was.


Kurt