🚀 go-pugleaf

In article <1992Sep25.141556.7799@news2.cis.umn.edu> fin@unet.umn.edu (Craig A. Finseth) writes:
>
>Personal opinion follows:
>
>	Considering that return from main() is handled specially by
>	the compiler, I consider _any_ recursive invocation of main()
>	to be suspect and hence not reasonable.
>
>	Why? Think for a moment: this is one of those cases that is
>	just begging to be badly handled by a compiler writer.  Who
>	needs this sort of grief?
>
>End personal opinion.  Return to usual sweeping generalizations (:-).

You are already making sweeping generalizations.  Although there are a
few compilers that have special case handling for main, there is no
requirment that this be so.  The traditional C compiliers handled
main() just like any other function.  The only thing that distinguished
main from any other function was that was the name that library setup
called after it had completed everything else.  The compiler itself did
not do anything special for the name "main".  For a lot of hardware,
the calling sequence does not really care if the actual number of
parameters passed and declared match exactly.  Prior to ANSI it was
almost amnidatory that stack/call sequences were designed to handle
such mismatches.  Certainly I have only ever used one compiler that
actually generated special code for the name "main".