[OpenAFS-devel] New Windows AKLOG...

[Rodney M Dyer]

Hi,

I'm working on radically cleaning up the AKLOG source that I have and 
making a clean-room Windows 2k/XP only implementation (probably an UNIX 
sin, but I'm doing it anyway).

Positives of the new version are...

	*  Written for Windows 2k/XP only.
	*  The original code has been cleaned and optimized extensively.
	*  Uses Microsoft's standard C libs.
	*  Maintained for the latest versions of OpenAFS 1.2.9a and MIT Kerberos 5 
1.3.
	*  Written as a pure Win32 console app.
	*  Retains all functionality in the original version.
	*  Added option flag -stlgtk to obtain a user logon token.
	*  Better logging.


The code would compile clean (zero errors) if not for a few inconsitent and 
conflicting header file definitions between OpenAFS and MIT Krb5...

http://www.coe.uncc.edu/~rmdyer/good_aklog_compile.txt


However, there's a single function call for automatically obtaining the 
Kerberos realm name from the KDC that's failing to compile...

	retval = krb5_get_host_realm(context, cellconfig->hostName[0], &hrealms);

This function call needs the "k5-int.h" header.  When I include that single 
header the compiler blows chunks...

http://www.coe.uncc.edu/~rmdyer/bad_aklog_compile.txt

I "can" just ignore that function + header and remark them out, but I would 
prefer to keep that functionality if at all possible.  What has happened to 
the Kerberos 5 build headers?  They seem to be very inconsistent.

Anybody know what the solution here might be?

Help is appreciated.

Thanks,

Rodney

Rodney M. Dyer
Windows Systems Programmer
Mosaic Computing Group
William States Lee College of Engineering
University of North Carolina at Charlotte
Email: rmdyer@uncc.edu
Web: http://www.coe.uncc.edu/~rmdyer
Phone (704)687-3518
Help Desk Line (704)687-3150
FAX (704)687-2352
Office  267 Smith Building