[net.lang.c] Desireability of command-line options to cc

kendall@wjh12.UUCP (Kendall) (09/11/83)

There have been almost no instances of command-line options to cc
which alter the language.  The Berkeley -R does so, but is reasonable
in that (1) it subtracts rather than adds to the langauge, and (2)
violation of the -R-induced restriction will lead to an error message.
Zilog has two C compilers, cc and cc2, which implement different
pointer sizes.  One could argue that other switches alter the language
in minor ways.

This restraint on the part of cc implementors is great.  Having a
program work if compiled one way, not if compiled another, would be a
disaster.

	Sam Kendall			{allegra,ihnp4}!wjh12!kendall
	Delft Consulting Corporation	  decvax!genrad!wjh12!kendall

crl@CSvax:Pucc-H:pur-phy.UUCP (09/13/83)

A great command line option for VAXen would be the ability to do
"standard" identifier and structure-member naming.  Trying to make
something portable to any C compiler and adhering to those silly
7/8 character limit on identifiers and unique structure members is
very difficult without an ability to have your compiler act that way
too.

Charles LaBrec
UUCP:		pur-ee!Physics:crl, purdue!Physics:crl
INTERNET:	crl @ pur-phy.UUCP