|[3 earlier articles]|
|Re: Java Comment-Preserving Grammar firstname.lastname@example.org (Chris Dodd) (2004-05-30)|
|Re: Java Comment-Preserving Grammar email@example.com (Dobes Vandermeer) (2004-05-30)|
|Re: Java Comment-Preserving Grammar firstname.lastname@example.org (Tim Bauer) (2004-05-30)|
|Re: Java Comment-Preserving Grammar email@example.com (2004-06-06)|
|Re: Java Comment-Preserving Grammar firstname.lastname@example.org (Clint Olsen) (2004-06-06)|
|Re: Java Comment-Preserving Grammar email@example.com (glen herrmannsfeldt) (2004-06-09)|
|Re: Java Comment-Preserving Grammar cfc@shell01.TheWorld.com (Chris F Clark) (2004-06-11)|
|Re: Java Comment-Preserving Grammar firstname.lastname@example.org (Alex Colvin) (2004-06-13)|
|Re: Java Comment-Preserving Grammar email@example.com (glen herrmannsfeldt) (2004-06-13)|
|Re: Java Comment-Preserving Grammar cfc@shell01.TheWorld.com (Chris F Clark) (2004-06-15)|
|Re: PL/I syntax, was Java Comment-Preserving Grammar firstname.lastname@example.org (Alex Colvin) (2004-06-21)|
|Re: PL/I syntax, was Java Comment-Preserving Grammar Peter_Flass@Yahoo.com (Peter Flass) (2004-06-26)|
|From:||Chris F Clark <cfc@shell01.TheWorld.com>|
|Date:||11 Jun 2004 02:56:42 -0400|
|Organization:||The World Public Access UNIX, Brookline, MA|
|References:||04-05-075 04-06-004 04-06-022|
|Posted-Date:||11 Jun 2004 02:56:42 EDT|
Jens Troeger wrote:
> As Far As I Know, A Grammar Has Nothing To Do With Comments. The
> Scanner Usually Skips Over Comments In The Source Code Already,
> Such That The Parser Doesn'T Even See Them. However, You Can Write
> Your Own Scanner And Parser, That Make Comments Explicit In The Grammer.
> But If You Want To Allow Comments _Anywhere_ In Yout Language, The
> Good Luck!! :-) That Grammar Is Going To Be A Mess....
Glen Herrmannsfeldt replied:
> Would it be that hard? Most languages that I know
> of allow comments in places where blank space is allowed.
> I suppose in most cases blanks are removed by the lexer, and so not
> included in the grammar, though I don't think it would be all that
> hard to add them in in the places they are allowed. It might not
> look so nice, though.
Not look so nice = hard.
Most grammars ignore all spacing and comment issues because they make
the grammar not look so nice (i.e. they obscure the real intent of the
grammar) and because the simplest forms of LL and LR parsing have
trouble parsing grammars that don't have all that "extraneous cruft"
In particular, each place a whitespace or comment token is allowed, it
is almost always optional. Unless your parser generator has been
extended to support regular expressions (and standard Yacc/Bison is
not, and prior to about 1985 there were no commonly available LL+RE
tools and about 1990 no LR+RE tools), you can't express such optinal
clauses "inline". And, if you can't do them inline, then you need a
"nullable" rule to express them. And, nullable rules tend to
introduce additional conflicts into the language.
Thus, prior to 1985, it was not only inconvenient to introduce
whitespace into your language, but it was likely to cause your parser
generator to fall over.
Around 1990 the situation improved dramatically. Not only were tools
introduced the integrated regular expressions into both the LL and LR
parsing methodology, which meant that it was possible to introduce the
whitspace rules into your grammar without breaking the generator that
processed it. The same tools started introducing special notations to
handle the whitespace problem explicitly. Over the years since 1990,
I have seen several advancements in that notation.
Of course, if one still uses garden variety Yacc/Bison, one is still
stuck in the pre-1990 era, but other tools like PCCTS, JavaCC, Yacc++,
and Meta-S all have some solution for the problem.
Hope this helps,
Chris Clark Internet : email@example.com
Compiler Resources, Inc. Web Site : http://world.std.com/~compres
23 Bailey Rd voice : (508) 435-5016
Berlin, MA 01503 USA fax : (978) 838-0263 (24 hours)
Return to the
Search the comp.compilers archives again.