benson@HARVARD.HARVARD.EDU (10/22/88)
Sun documentation claims that K&R requires that all float calculations be performed in double precision. Sun's compiler provides -fsingle to use single precision. gcc appears to use single precision for floats. This is fine by me, but is it what you really want? another question: I could leave out -traditional except for comments in preprocessor macros. I don't suppose that you would consider distinguishing more finely. --benson