Pull out a little more char* from our core infrastructure. Add shims to turn down case_ignore_strmp and case_ignore_strncmp().
Tweak reference files to deal with more correct(?) ordering in edge cases.
Start sketching a shim layer for shortname, description, and text, but increasingly I'm of the opinion that our only options are either do it all in one submit (eeek!) or to add additional members of struct waypt that are allocated/copied at waypt_add time so we can then at least separate readers and writers. Instinct tells me the latter will result in harsh punishment as we'll have members that are out of sync and it will be difficult to flush them out. So substantial parts of this CL may get tossed.