Andre Poenitz wrote: > > This unifies a few types used in LCursor and the math cursor. > > Note that this unifies on the 'wrong side': paroffset_type should be > unsigned in the long run but there seems to be a place or two where > the core code relies on signedness (a downward counting loop most > likely). > > It works as-is and could be corrected later by hunting down that > loop and setting paroffset_type to size_type in support/types.h.
Add a comment to this effect then. Presumably John would like it made visible with a "FIXME" ;-) authors: aren't the same people resposible for both the .C and the .h files? boundary_: _nice_ comment ;-) -- Angus
