Debugging API for krb5

[Sam Hartman]

>>>>> "Jeffrey" == Jeffrey Altman <jaltman at columbia.edu> writes:

    Jeffrey> Sam Hartman wrote:
    >>>>>>> "Jeffrey" == Jeffrey Altman <jaltman at columbia.edu> writes:
    >>>>>>> 
    >>
    Jeffrey> Actually, I did not leave out the mask in the
    Jeffrey> krb5int_debug_message() function.  My expectation is that
    Jeffrey> the function will only be called if the appropriate mask
    Jeffrey> bits are set.  Probably we need a macro
    >>  This seems like bad design because it semes like you are
    >> asking the caller to repeat work that is needed for every use
    >> of the function.
    >> 

    Jeffrey> I really don't want to incur the function overhead if
    Jeffrey> debugging is not being used.  Perhaps then

If we must use an inline function, we can do that.  Look at
k5-platform.h I think for examples of how we're doing that.
Personally I wouldn't expect you to need to have debugging in
sufficiently critical code paths that the function overhead is an
issue.