Related articles |
---|
Re: Compiler support for preprocessors (was: Using noweb with sml) nr@cs.purdue.edu (1996-02-23) |
Re: Compiler support for preprocessors (was: Using noweb with sml) nr@cs.purdue.edu (Norman Ramsey) (1996-03-01) |
From: | nr@cs.purdue.edu (Norman Ramsey) |
Newsgroups: | comp.compilers |
Date: | 23 Feb 1996 18:27:00 -0500 |
Organization: | Compilers Central |
References: | <4gii2t$ovf@cantaloupe.srv.cs.cmu.edu>, |
Keywords: | design, standards |
Michael Siff <siff@cs.wisc.edu> wrote:
>I am attempting to use noweb with sml. The notangle program has an
>optional argument, -L, which according to the noweb documentation, "emit
>pragmas that inform the compilers of the placement of lines in the noweb
>soruce." ...These pragmas are fine for C. But how can I get the SML compiler
>to recognize them?
Years ago, I lobbied the implementors of Standard ML of New Jersey for
a formal comment or other feature by which to achieve this goal. They
were not interested, in part I think because of a belief that ``one
should not have preprocessors generate ML code; one should use
functors or higher-order functions.'' I love functors and
higher-order functions, but I don't believe they are the best solution
to every problem.
Pester your favorite compiler writers and language designers to get
support for line-number pragmas. It is relatively easy to implement
(I added it to SRC Modula-3), and it is useful for all kinds of
preprocessors (e.g., yacc clones), not just for literate-programming
tools.
--
Norman Ramsey
http://www.cs.purdue.edu/homes/nr
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.