Re: [Perlgssapi-developer] [rt.cpan.org #32788] Enhancement to use OpenSolaris/Solaris 10 native gss library

[Achim G. <ac...@gr...>]

On Tuesday 29 January 2008, Leif Johansson wrote:

> > 'gss_nt_service_name' is never used in GSSAPI.pm,
> > I don't understand why there ist a need to define it?
>
> I don't remember exactly but I think gss_nt_service_name is an MITism

I agree.

> and I know that Solaris kerberos is essentially MIT kerberos.=20

I agree.

> I think you=20
> may need the definition to be compatible with older versions of MIT krb.

GSSAPI.pm internally always uses GSS_C_NT_HOSTBASED_SERVICE.
No direct dependency to 'gss_nt_service_name'.

If GSSAPI.pm is build against old MIT libraries ( 1.2.x )
GSS_C_NT_HOSTBASED_SERVICE is not available, but defined
by gss_nt_service_name instead.

=46rom GSSAPI.xs :

#if defined(MITKERB12)
/* symbols not defined in MIT Kerberos 1.2.x */
#define GSS_C_NT_USER_NAME gss_nt_user_name
#define GSS_C_NT_MACHINE_UID_NAME gss_nt_machine_uid_name
#define GSS_C_NT_STRING_UID_NAME gss_nt_string_uid_name
#define GSS_C_NT_HOSTBASED_SERVICE gss_nt_service_name
#define GSS_C_NT_EXPORT_NAME gss_nt_exported_name
#endif


But *if* GSS_C_NT_HOSTBASED_SERVICE is available and
gss_nt_service_name is not used...
Why the need to define it in GSSAPI.xs?
Why to define it in "the other direction"?

#define gss_nt_service_name GSS_C_NT_HOSTBASED_SERVICE

Hmm. Am I wrong?

Thank you,
Achim