Re: [kitten] New Version Notification for draft-kaduk-kitten-gss-loop-01.txt (fwd)

[[e-mail] (Martin Rex)]

Martin Rex wrote:
> Martin Rex wrote:
> > > 
> > > (*) I also noticed that ~all of Appendix A of RFC 2744 (the sample 
> > > gssapi.h header) has incorrect signatures for the listed functions. 
> > > Basically, any pointer argument which is not in the final position, is not 
> > > listed as a pointer type.  It seems likely that a script to remove 
> > > argument names was buggy.  I'm not sure that I'm up to submitting a full 
> > > erratum to note all the changes, though.
> > 
> > I'm not aware of any problem with that header file (aside from the
> > invalid use of const for custom pointer types).
> 
> Ooops, you're correct.
> 
> Admittedly, I never tried to extract and use the header from rfc2744,
> and just looking at it again, I agree, it seems to have a few errors
> (lacking "*") in many of the output/modify function parameters
> --seemingly except for the last function parameter, where the "*"
> is usually present.


The "*" on the output parameters in the Appendix A sample header file
seem to have disappeared during I-D revision 05->06 of the C-bindings
draft, they were present during -00 through -05:

  http://tools.ietf.org/rfcdiff?url2=rfc2744.txt&url1=draft-ietf-cat-gssv2-cbind-05.txt#diff0410

I put together my gssapi_2.h header in 1995 (-00) and didn't notice
the corruption in Appendix A starting in cbind-06.


A "fix" of that size appears somewhat huge for the errata process.
Filing an errata that describes the nature of the problem--maybe including
a suitable URL to rfcdiff like the one above for visualizing the problem,
might be preferable.


The formatting of the C-Bindings I-D changed considerably from 05->06,
and apparently the gssapi.h sample header file in Appendix A got corrupted.
Scanning my archives, the C-Bindings author even indicated that he
had changed his I-D tooling.  Unfortunately http://tools.ietf.org/rfcdiff
didn't exist back then (and previous versions of drafts were also no
longer accessible in the download directories (so unless you happened
to have a previous copy...).

:  From:     John Wray
:  Subject:  New GSSAPI C-bindings draft
:  Date:     Mon, 13 Jul 1998 10:41:05 -0400
:  To:       [e-mail]
:
:  I've just submitted an updated C-bindings document, so expect
:  draft-ietf-cat-gssv2-cbind-06.txt to show up on the repositories in the
:  next couple of days.
:
:  Changes from the previous version:
:
:     Reformatting.  Changing jobs meant I no longer have all
:        the nroff macros I used to use, so I took the opportunity
:        to migrate to a slightly more modern formatting program.
:        One unfortunate result of this is that diff-ing this version
:        against the previous one won't give much useful
:        information.
:
:     Incorporation of the 2078bis alignments posted to the
:        list by Martin.
:
:     A number of typos and clarification requests received
:        by e-mail (also from Martin).
:
: John


Btw. there is another problem with the gssapi.h sample header file
in rfc2744:  for 64-bit, there exists a binary-incompatibility
between the X/Open defined GSS-API C-bindings and the IETF-defined
rfc2744 C-Bindings for "count" member of the gss_OID_set type:

rfc2744:  http://tools.ietf.org/html/rfc2744#page-84

     typedef struct gss_OID_set_desc_struct  {
         size_t     count;
         gss_OID    elements;
     } gss_OID_set_desc, *gss_OID_set;
     

X/Open: http://archive.opengroup.org/publications/archive/CDROM/c441.pdf
Page 45:

     typedef struct gss_OID_set_desc_struct{
         int count;
         gss_OID elements;
     } gss_OID_set_desc, *gss_OID_set;


-Martin