{~ pcc & Bell ~} - (nf)

log11 at ucbcad.UUCP log11 at ucbcad.UUCP
Sat Jun 11 21:12:02 AEST 1983

ucbcad!log11    Jun 11 04:05:00 1983

	{~	/* (flame on) */

	I spoke recently to the author of the Berkeley Pascal compiler
(not pi) about a document of his which gives details on the output of
pass 1 (i.e., from c1) of pcc.  He said that Bell lawyers ruled that
distribution of this document to non-source-licensed sites would violate
source-license agreements.  The reason being that the numbers put out by
c1 are #define's in the source of pcc.  While (to my knowledge) the
document does not quote the source directly, this does give all the
goodies away, by Bell's lights (which I think are rather dim, personally,
but let's not get into that.)

	Now: what if we were to assemble a suite of C source codes with
documentation to aid users in inferring the meanings of the manifest
constants produced from c1?  Ideally, this scheme would produce reliable
inferences regardless of the values of the constants.  Would this document
also be in violation of the rules?

	What's wrong with thinking of c2, et. al., as compiler-writing
tools, anyway?  With m4/cpp, lex, yacc, AND c2, the amount of drudgery
involved in bringing up compilers on UNIX* could be substantially
reduced.  This could only be to Ma B's ultimate benefit.

	~}	/* (flame off) */

	    Michael Turner

P.S. Personal replies should be sent to the above account, not the
     pseudonymous login "log11", under which this submission appears.
* UNIX is somebody's trademark, last I heard.

sendflame: missing closing "~}" (line 33)

More information about the Comp.lang.c mailing list