fixup commit for tag 'llgen-1-0'

UnGetChar call is wrong after skipspaces().

first/local.h.src

@@ -1,7 +0,0 @@
-/* collection of options, selected by including or excluding 'defines' */
-
-/* Version number of the EM object code */
-#	define	VERSION		3	/* 16 bits number */
-
-/* The default machine used by ack, acc, apc */
-#	define	ACKM		"pdp"

mach/m68k2/cg/mach.c

@@ -1,274 +0,0 @@
-#ifndef NORCSID
-static char rcsid[] = "$Header$";
-#endif
-
-/*
- * (c) copyright 1983 by the Vrije Universiteit, Amsterdam, The Netherlands.
- *
- *          This product is part of the Amsterdam Compiler Kit.
- *
- * Permission to use, sell, duplicate or disclose this software must be
- * obtained in writing. Requests for such permissions may be sent to
- *
- *      Dr. Andrew S. Tanenbaum
- *      Wiskundig Seminarium
- *      Vrije Universiteit
- *      Postbox 7161
- *      1007 MC Amsterdam
- *      The Netherlands
- *
- */
-
-/*
- * machine dependent back end routines for the Motorola 68000
- */
-
-con_part(sz,w) register sz; word w; {
-
-	while (part_size % sz)
-		part_size++;
-	if (part_size == TEM_WSIZE)
-		part_flush();
-	if (sz == 1) {
-		w &= 0xFF;
-		if (part_size == 0)
-			w <<= 8;
-		part_word |= w;
-	} else {
-		assert(sz == 2);
-		part_word = w;
-	}
-	part_size += sz;
-}
-
-con_mult(sz) word sz; {
-
-	if (sz != 4)
-		fatal("bad icon/ucon size");
-	fprintf(codefile,".long %s\n",str);
-}
-
-con_float() {
-
-static int been_here;
-	if (argval != 4 && argval != 8)
-		fatal("bad fcon size");
-	fprintf(codefile,".long\t");
-	if (argval == 8)
-		fprintf(codefile,"F_DUM,");
-	fprintf(codefile,"F_DUM\n");
-	if ( !been_here++)
-	{
-	fprintf(stderr,"Warning : dummy float-constant(s)\n");
-	}
-}
-
-#ifdef REGVARS
-
-
-regscore(off,size,typ,score,totyp)
-	long off;
-{
-	if (score == 0) return -1;
-	switch(typ) {
-		case reg_float:
-			return -1;
-		case reg_pointer:
-			if (size != 4 || totyp != reg_pointer) return -1;
-			score *= 2;
-			break;
-		case reg_loop:
-			score += 5;
-			/* fall through .. */
-		case reg_any:
-			if (size != 2 || totyp == reg_pointer) return -1;
-			break;
-	}
-	if (off >= 0) {
-		/* parameters must be initialised with an instruction
-		 * like "move.w 4(a6),d0", which costs 2 words.
-		 */
-		score -= 2;
-	}
-	score -= 1; /* take save/restore into account */
-	return score;
-}
-
-
-struct regsav_t {
-	char	*rs_reg;	/* e.g. "a3" or "d5" */
-	long	rs_off;		/* offset of variable */
-	int	rs_size;	/* 2 or 4 bytes */
-} regsav[9];
-
-
-int regnr;
-int nr_a_regs,nr_d_regs;
-int TEM_BSIZE;
-static long nlocals;
-
-prolog(n)
-{	nlocals = n; }
-
-
-i_regsave()
-{
-	regnr = 0;
-	nr_a_regs = 0;
-	nr_d_regs = 0;
-	TEM_BSIZE = 0;
-}
-
-#define MOVEM_LIMIT	2
-/* If #registers to be saved exceeds MOVEM_LIMIT, we
-* use the movem instruction to save registers; else
-* we simply use several move.l's.
-*/
-
-
-save()
-{
-	register struct regsav_t *p;
-	int i;
-
-	if (regnr > MOVEM_LIMIT) {
-		fprintf(codefile,"movem.l ");
-		for (p = regsav; ;) {
-			fprintf(codefile,"%s",p->rs_reg);
-			if (++p == &regsav[regnr]) break;
-			putc('/',codefile);
-		}
-		fprintf(codefile,",-(sp)\n");
-	} else {
-		/* Note that the order in which the registers are saved
-		 * is important; it is used by gto.s.
-		 */
-		for (i = 0; i < nr_a_regs; i++) {
-			fprintf(codefile,"move.l a%d,-(sp)\n",5-i);
-		}
-		for (i = 0; i < nr_d_regs; i++) {
-			fprintf(codefile,"move.l d%d,-(sp)\n",7-i);
-		}
-	}
-	/* Push a mask that indicates which registers were saved */
-	assert(nr_d_regs < 8 && nr_a_regs < 8);
-	if (nr_d_regs == 0 && nr_a_regs == 0) {
-		fprintf(codefile,"clr.w -(sp)\n");
-	} else {
-		fprintf(codefile,"move.w #%d,-(sp)\n",
-			nr_d_regs + (nr_a_regs<<3));
-	}
-
-	/* Compute AB - LB */
-	TEM_BSIZE = 4 * (nr_d_regs + nr_a_regs) + 10;
-
-	/* allocate space for local variables */
-	fprintf(codefile,"tst.b -%D(sp)\nlink\ta6,#-%D\n",nlocals+40,nlocals);
-
-	/* initialise register-parameters */
-	for (p = regsav; p < &regsav[regnr]; p++) {
-		if (p->rs_off >= 0) {
-			fprintf(codefile,"move.%c %ld(a6),%s\n",
-				(p->rs_size == 4 ? 'l' : 'w'),
-				p->rs_off + TEM_BSIZE,
-				p->rs_reg);
-		}
-	}
-}
-
-restr()
-{
-	register struct regsav_t *p;
-	int i;
-
-	fprintf(codefile,"unlk a6\n");
-	fprintf(codefile,"add.l #2,sp\n"); /* pop mask */
-	if (regnr > MOVEM_LIMIT)  {
-		fprintf(codefile,"movem.l (sp)+,");
-		for (p = regsav; ;) {
-			fprintf(codefile,"%s",p->rs_reg);
-			if (++p == &regsav[regnr]) break;
-			putc('/',codefile);
-		}
-		putc('\n',codefile);
-	} else {
-		for (i = nr_d_regs - 1; i >= 0; i--) {
-			fprintf(codefile,"move.l (sp)+,d%d\n",7-i);
-		}
-		for (i = nr_a_regs - 1; i >= 0; i--) {
-			fprintf(codefile,"move.l (sp)+,a%d\n",5-i);
-		}
-	}
-	fprintf(codefile,"rts\n");
-}
-
-
-f_regsave()
-{
-	save();
-}
-
-regsave(str,off,size)
-	char *str;
-	long off;
-{
-	assert (regnr < 9);
-	regsav[regnr].rs_reg = str;
-	if (str[0] == 'a') {
-		nr_a_regs++;
-	} else {
-		assert(str[0] == 'd');
-		nr_d_regs++;
-	}
-	regsav[regnr].rs_off = off;
-	regsav[regnr++].rs_size = size;
-	fprintf(codefile, "!Local %ld into %s\n",off,str);
-}
-
-regreturn()
-{
-	restr();
-}
-
-#endif
-
-#ifndef REGVARS
-
-prolog(nlocals) full nlocals; {
-
-	fprintf(codefile,"tst.b -%D(sp)\nlink\ta6,#-%D\n",nlocals+40,nlocals);
-}
-
-#endif
-
-
-
-mes(type) word type ; {
-	int argt ;
-
-	switch ( (int)type ) {
-	case ms_ext :
-		for (;;) {
-			switch ( argt=getarg(
-			    ptyp(sp_cend)|ptyp(sp_pnam)|sym_ptyp) ) {
-			case sp_cend :
-				return ;
-			default:
-				strarg(argt) ;
-				fprintf(codefile,".define %s\n",argstr) ;
-				break ;
-			}
-		}
-	default :
-		while ( getarg(any_ptyp) != sp_cend ) ;
-		break ;
-	}
-}
-
-
-char    *segname[] = {
-	".text",        /* SEGTXT */
-	".data",        /* SEGCON */
-	".data",        /* SEGROM */
-	".bss"          /* SEGBSS */
-};

mach/m68k4/libem/dvi.s

@@ -1,42 +0,0 @@
-.define .dvi
-
- ! signed long divide
-	.text
-.dvi:
-	movem.l	d0/d2,.savdvi
-	move.l	(sp)+,.retdvi
-	move.l	(sp)+,d0
-	move.l	(sp)+,d1
-	clr.l	d4
-	tst.l	d0	! divisor
-	bpl	1f
-	neg.l	d0
-	not	d4
-1:
-	tst.l	d1	! dividend
-	bpl	2f
-	neg.l	d1
-	not	d4
-	swap	d4
-	not	d4
-	swap	d4
-2:
-	move.l	d1,-(sp)
-	move.l	d0,-(sp)
-	jsr	.dvu
-	tst	d4
-	beq	5f
-	neg.l	d1	! quotient
-5:
-	tst.l	d4
-	bpl	6f
-	neg.l	d3	! remainder
-6:
-	movem.l .savdvi,d0/d2
-	move.l	.retdvi,-(sp)
-	rts
-.data
-.savdvi:	.space 8
-.retdvi:	.long 0
-.text
-.align 2

util/LLgen/COPYING

@@ -0,0 +1,32 @@
+Copyright (c) 1987, 1990, 1993, 2005 Vrije Universiteit, Amsterdam, The Netherlands.
+All rights reserved.
+
+Redistribution and use of the Amsterdam Compiler Kit in source and
+binary forms, with or without modification, are permitted provided
+that the following conditions are met:
+
+   * Redistributions of source code must retain the above copyright
+     notice, this list of conditions and the following disclaimer.
+
+   * Redistributions in binary form must reproduce the above
+     copyright notice, this list of conditions and the following
+     disclaimer in the documentation and/or other materials provided
+     with the distribution.
+
+   * Neither the name of Vrije Universiteit nor the names of the
+     software authors or contributors may be used to endorse or
+     promote products derived from this software without specific
+     prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS, AUTHORS, AND
+CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
+INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+IN NO EVENT SHALL VRIJE UNIVERSITEIT OR ANY AUTHORS OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

util/LLgen/READ_ME

@@ -0,0 +1,84 @@
+                                 LLGEN V1.0
+                                 ==========
+
+  Copyright © 1991-2005 by the Vrije Universiteit, Amsterdam, the Netherlands.
+                                 2006-02-03
+
+INTRODUCTION
+============
+
+LLgen is a LL(1) parser in the style of yacc. It will generate an efficient
+recursive descent parser for Extended Context-Free grammars, with an optional
+non-correcting error recovery mechanism.
+
+For more information, see the white papers in the doc directory. (You may need
+to build LLgen first to be able to read them.)
+
+INSTALLATION
+============
+
+LLgen depends on the following software:
+
+    gcc: currently, LLgen only builds on gcc, due largely to the lack of any
+        non-gcc test systems. Please contact the support mailing list (see
+        below) if this is a problem.
+        
+    groff: used to build the documentation.
+
+LLgen uses Prime Mover as its build tool. To build, do:
+
+    ./pm install
+
+This will compile LLgen and install it into /usr/local (which you must be able
+to write to). If you want it installed elsewhere, use -D to change the PREFIX
+variable:
+
+    ./pm -DPREFIX="/home/dg/" install
+    
+(If you do change the prefix, you must supply it whenever you invoke pm. And
+there must be a trailing / in the path you give it.)
+
+Invoking ./pm by itself will compile it but not install it; look in the bin
+directory for an image of 
+
+USING LLGEN
+===========
+
+Please see the white papers in PREFIX/share/doc/LLgen, or the man page (called
+LLgen).
+
+LLgen itself uses a parser compiled in LLgen, so you can study the files in
+the src directory for examples. If you wish to modify LLgen's own parser, the
+bootstrap.sh script will update LLgen's source from the parser *.g files.
+LLgen itself may then be recompiled with pm in the usual way. (Although you
+may want to keep a copy of a known good LLgen around in case you break
+something!)
+
+SUPPORT
+=======
+
+LLgen is part of the Amsterdam Compiler Kit (although standalone and
+distributed seperately). For more details, please see:
+
+    http://tack.sourceforge.net
+
+There is a mailing list available at:
+
+	http://sourceforge.net/mail/?group_id=130811
+
+LICENSE
+=======
+
+LLgen is © 1991-2005 by the Vrije Universiteit and is distributed under a
+license equivalent to the three-clause BSD license. See the COPYING file
+for details.
+
+Prime Mover is © 2006 David Given and is distributed under the MIT license.
+Do './pm --license' for details.
+
+-----------------------------------------------------------------------------
+David Given
+dg@cowlark.com
+2006-02-04
+
+/* $Id$ */

util/LLgen/bootstrap.sh

@@ -0,0 +1,16 @@
+#!/bin/sh
+# $Id$
+#
+# Script to rebuild LLgen's own parser with LLgen.
+
+set -e
+
+if ! (which LLgen > /dev/null); then
+	echo "You can only bootstrap the LLgen parser if you've already got"
+	echo "LLgen installed."
+fi
+
+(cd src && LLgen -vvv -x tokens.g LLgen.g)
+
+echo ""
+echo "You should now be able to rebuild LLgen with the new parser."

util/LLgen/c.pm

@@ -0,0 +1,144 @@
+-- This file is part of Prime Mover v0.1pre1.
+-- (C) 2006 David Given
+--
+-- pm includefile to compile *host* C programs.
+
+-- Standard Lua boilerplate.
+
+local io_open = io.open
+local string_gsub = string.gsub
+local string_gfind = string.gfind
+local table_insert = table.insert
+local table_getn = table.getn
+local filetime = pm.filetime
+
+-- Define some variables.
+
+CC = "gcc %CBUILDFLAGS% %CDYNINCLUDES% %CINCLUDES% %CEXTRAFLAGS% -c -o %out% %in%"
+CPROGRAM = "gcc %CBUILDFLAGS% %CLINKFLAGS% %CEXTRAFLAGS% -o %out% %in% %CLIBRARIES%"
+CDEPENDS = "gcc %CBUILDFLAGS% %CDYNINCLUDES% %CINCLUDES% %CEXTRAFLAGS% -MM -MG -MF %out% %in%"
+AR = "%RM% %out% && ar cr %out% %in%"
+
+CBUILDFLAGS = "-g -Os"
+CINCLUDES = {}
+CEXTRAFLAGS = ""
+CLINKFLAGS = ""
+CDYNINCLUDES = ""
+CLIBRARIES = ""
+
+--- Manage C file dependencies ----------------------------------------------
+
+local dependency_cache = {}
+local function load_dependency_file(fn)
+	local o = dependency_cache[fn]
+	if o then
+		return o
+	end
+	
+	-- Read in the dependency file.
+	
+	local f = io_open(fn)
+	if not f then
+		print("failed to open "..fn)
+		return nil
+	end
+	f = f:read("*a")
+	
+	-- Massage the dependency file into a string containing one unescaped
+	-- filename per line.
+	
+	f = string_gsub(f, "^.*[^\\]: *", "")
+	f = string_gsub(f, "\\\r?\n", "")
+	f = string_gsub(f, "([^\\]) +", "%1\n")
+	f = string_gsub(f, "\\", "")
+	
+	-- Parse the string.
+	
+	o = {}
+	for l in string_gfind(f, "[^\n]+") do
+		table_insert(o, l)
+	end
+	
+	dependency_cache[fn] = o
+	return o
+end
+
+-- This clause specialises 'simple' to add support for smart dependencies of C
+-- files.
+
+simple_with_clike_dependencies = simple {
+	class = "simple_with_clike_dependencies",
+	makedepends = {"%CDEPENDS%"},
+
+	__init = function(self, p)
+ 		simple.__init(self, p)
+  		
+		-- If we're a class, don't verify.
+		
+		if ((type(p) == "table") and p.class) then
+			return
+		end
+
+		-- If dynamicheaders is an object, turn it into a singleton list.
+		
+		if self.dynamicheaders then
+			if (type(self.dynamicheaders) ~= "table") then
+				self:__error("doesn't know what to do with dynamicheaders, which ",
+					"should be a list or an object but was a ", type(self.dynamicheaders))
+			end
+			if self.dynamicheaders.class then
+				self.dynamicheaders = {self.dynamicheaders}
+			end
+		end
+	end,
+	
+	__dependencies = function(self, inputs, outputs)
+		local obj = simple {
+			CDYNINCLUDES = self.CDYNINCLUDES,
+			command = self.makedepends,
+			outputs = {"%U%-%I%.d"},
+			unpack(inputs)
+		}
+		local o = obj:__build()
+		local depends = load_dependency_file(o[1])
+		if not depends then
+			self:__error("could not determine the dependencies for ",
+				pm.rendertable(inputs))
+		end
+		return depends
+	end,
+	
+	__buildadditionalchildren = function(self)
+		self.CDYNINCLUDES = ""
+		if self.dynamicheaders then
+			for _, i in ipairs(self.dynamicheaders) do
+				local o = i:__build()
+				if o[1] then
+					self.CDYNINCLUDES = self.CDYNINCLUDES..' "-I'..string_gsub(o[1], "/[^/]*$", "")..'"'
+				end
+			end
+		end
+	end
+}
+
+-- These are the publically useful clauses.
+
+cfile = simple_with_clike_dependencies {
+	class = "cfile",
+	command = {"%CC%"},
+	outputs = {"%U%-%I%.o"},
+}
+
+cprogram = simple {
+	class = "cprogram",
+	command = {"%CPROGRAM%"},
+	outputs = {"%U%-%I%"},
+}
+
+clibrary = simple {
+	class = "clibrary",
+	command = {
+		"%AR%"
+	},
+	outputs = {"%U%-%I%.a"},
+}

util/LLgen/doc/LLgen.1

@@ -0,0 +1,139 @@
+.\" $Id$
+.TH LLGEN 1 "$Revision$"
+.ad
+.SH NAME
+LLgen, an extended LL(1) parser generator
+.SH SYNOPSIS
+LLgen [ \-vxwans ] [ \-j[\fInum\fP] ] [ \-l\fInum\fP ] [ \-h\fInum\fP ] file ...
+.SH DESCRIPTION
+\fILLgen\fP
+converts a context-free grammar into a set of
+functions which form a recursive descent parser with no backtrack.
+The grammar may be ambiguous;
+ambiguities can be broken by user specifications.
+.PP
+\fILLgen\fP
+reads each
+\fIfile\fP
+in sequence.
+Together, these files must constitute a context-free grammar.
+For each file,
+\fILLgen\fP
+generates an output file, which must be compiled by the
+C-compiler.
+In addition, it generates the files
+\fILpars.c\fP
+and
+\fILpars.h.\fP
+\fILpars.h\fP
+contains the
+\fIdefine\fP
+statements that associate the
+\fILLgen\fP-assigned `token-codes' with user declared `token-names'.
+This allows other source files, for instance the source file
+containing the lexical analyzer,
+to access the token-codes by
+using the token-names.
+\fILpars.c\fP
+contains the error recovery routines and tables. It must also
+be compiled by the C-compiler. When the generated parser uses non-correcting
+error recovery (
+\fB\-n\fP 
+option) 
+\fILLgen\fP
+also generates a file 
+\fILncor.c\fP
+that contains the non-correcting recovery mechanism.
+.PP
+\fILLgen\fP
+will only update those output files that differ from their previous
+version.
+This allows
+\fILLgen\fP
+to be used with
+\fImake\fP
+(1) convieniently.
+.PP
+To obtain a working program, the user must also supply a
+lexical analyzer, as well as
+\fImain\fP
+and
+\fILLmessage\fP,
+an error reporting routine;
+\fILex\fP
+(1) is a useful program for creating lexical analysers usable
+by
+\fILLgen\fP.
+.PP
+\fILLgen\fP accepts the following flags:
+.IP \fB\-v\fP
+create a file called
+\fILL.output\fP,
+which contains a description of the conflicts that
+were not resolved.
+If the flag is given more than once,
+\fILLgen\fP
+will be more "verbose".
+If it is given three times, a complete description of the
+grammar will be supplied.
+.IP \fB\-x\fP
+the sets that are computed are extended with the nonterminal
+symbols and these extended sets are also included in the
+\fILL.output\fP
+file.
+.IP \fB\-w\fP
+no warnings are given.
+.IP \fB\-a\fP
+Produce ANSI C function headers and prototypes.
+.IP \fB\-n\fP
+Produce a parser with non-correcting error recovery.
+.IP \fB\-s\fP
+Simulate the calling of all defined subparsers in all semantic actions. When 
+using non-correcting error recovery, subparsers that are called in semantic
+actions may cause problems; this flag provides a `brute-force' solution. 
+.IP \fB\-j\fP[\fInum\fP]
+when this flag is given, \fILLgen\fP will generate dense switches,
+so that the compiler can generate a jump table for it. This will only be
+done for switches that have density between
+\fIlow_percentage\fP and \fIhigh_percentage\fP, as explained below.
+Usually, compilers generate a jumptable when the density of the switch
+is above a certain threshold. When jump tables are to be used more often,
+\fIhigh_percentage\fP must be set to this threshold, and \fIlow_percentage\fP
+must be set to a minimum threshold. There is a time-space trade-off here.
+.I num
+is the minimum number of cases in a switch for the \fB\-j\fP option to be
+effective. The default value (if
+.I num
+is not given) is 8.
+.IP \fB\-l\fP\fInum\fP
+The \fIlow_percentage\fP, as described above. Default value is 10.
+.IP \fB\-h\fP\fInum\fP
+The \fIhigh_percentage\fP, as described above. Default value is 30.
+.SH FILES
+LL.output                 verbose output file
+.br
+Lpars.c                   the error recovery routines
+.br
+Lncor.c                   non-correcting error recovery mechanism
+.br
+Lpars.h                   defines for token names
+.SH "SEE ALSO"
+\fIlex\fP(1)
+.br
+\fImake\fP(1)
+.br
+\fILLgen, an Extended LL(1) Parser Generator\fP
+by C.J.H. Jacobs.
+.br
+\fITop-down Non-Correcting Error Recovery in LLgen\fP
+by A.W van Deudekom and P.J. Kooiman
+.SH DIAGNOSTICS
+Are intended to be self-explanatory. They are reported
+on standard error. A more detailed report is found in the
+\fILL.output\fP
+file.
+.SH AUTHOR
+Ceriel J. H. Jacobs
+.br
+The non-correcting error recovery mechanism is written by
+A.W van Deudekom and P.J. Kooiman.

util/LLgen/doc/LLgen.refs

@@ -0,0 +1,54 @@
+%T An ALL(1) Compiler Generator
+%A D. R. Milton
+%A L. W. Kirchhoff
+%A B. R. Rowland
+%B Proc. of the SIGPLAN '79 Symposium on Compiler Construction
+%D August 1979 
+%J SIGPLAN Notices
+%N 8
+%P 152-157
+%V 14
+
+%T Lex - A Lexical Analyser Generator
+%A M. E. Lesk
+%I Bell Laboratories
+%D October 1975
+%C Murray Hill, New Jersey
+%R Comp. Sci. Tech. Rep. No. 39
+
+%T Yacc: Yet Another Compiler Compiler
+%A S. C. Johnson
+%I Bell Laboratories
+%D 1975
+%C Murray Hill, New Jersey
+%R Comp. Sci. Tech. Rep. No. 32
+
+%T The C Programming Language
+%A B. W. Kernighan
+%A D. M. Ritchie
+%I Prentice-Hall, Inc.
+%C Englewood Cliffs, New Jersey
+%D 1978
+
+%A M. Griffiths
+%T LL(1) Grammars and Analysers
+%E F. L. Bauer and J. Eickel
+%B Compiler Construction, An Advanced Course
+%I Springer-Verlag
+%C New York, N.Y.
+%D 1974
+
+%T Make - A Program for Maintaining Computer Programs
+%A S. I. Feldman
+%J Software - Practice and Experience
+%V 10
+%N 8
+%P 255-265
+%D August 1979
+
+%T Methods for the Automatic Construction of Error Correcting Parsers
+%A J. R\*:ohrich
+%J Acta Informatica
+%V 13
+%P 115-139
+%D 1980

util/LLgen/lib/incl

@@ -0,0 +1,68 @@
+/* $Id$ */
+#ifdef LL_DEBUG
+#include <assert.h>
+#include <stdio.h>
+#define LL_assert(x)	assert(x)
+#else
+#define LL_assert(x)	/* nothing */
+#endif
+
+extern int LLsymb;
+
+#define LL_SAFE(x)	/* Nothing */
+#define LL_SSCANDONE(x)	if (LLsymb != x) LLsafeerror(x)
+#define LL_SCANDONE(x)	if (LLsymb != x) LLerror(x)
+#define LL_NOSCANDONE(x) LLscan(x)
+#ifdef LL_FASTER
+#define LLscan(x)	if ((LLsymb = LL_LEXI()) != x) LLerror(x)
+#endif
+
+extern unsigned int LLscnt[];
+extern unsigned int LLtcnt[];
+extern int LLcsymb;
+
+#if LL_NON_CORR
+extern int LLstartsymb;
+#endif
+
+#define LLsdecr(d)	{LL_assert(LLscnt[d] > 0); LLscnt[d]--;}
+#define LLtdecr(d)	{LL_assert(LLtcnt[d] > 0); LLtcnt[d]--;}
+#define LLsincr(d)	LLscnt[d]++
+#define LLtincr(d)	LLtcnt[d]++
+
+#if LL_ANSI_C
+extern int LL_LEXI(void);
+extern void LLread(void);
+extern int LLskip(void);
+extern int LLnext(int);
+extern void LLerror(int);
+extern void LLsafeerror(int);
+extern void LLnewlevel(unsigned int *);
+extern void LLoldlevel(unsigned int *);
+#ifndef LL_FASTER
+extern void LLscan(int);
+#endif
+#ifndef LLNOFIRSTS
+extern int LLfirst(int, int);
+#endif
+#if LL_NON_CORR
+extern void LLnc_recover(void);
+#endif
+#else /* not LL_ANSI_C */
+extern LLread();
+extern int LLskip();
+extern int LLnext();
+extern LLerror();
+extern LLsafeerror();
+extern LLnewlevel();
+extern LLoldlevel();
+#ifndef LL_FASTER
+extern LLscan();
+#endif
+#ifndef LLNOFIRSTS
+extern int LLfirst();
+#endif
+#if LL_NON_CORR
+extern LLnc_recover();
+#endif
+#endif /* not LL_ANSI_C */

util/LLgen/lib/nc_incl

@@ -0,0 +1,70 @@
+#define LLALT 9999			/* Alternative is following */
+
+#define LLTERMINAL 1			/* Symbol is a terminal */
+#define LLNONTERMINAL 2			/* Symbol is a nonterminal */
+#define LLEORULE 0			/* No more alternatives */
+
+
+struct lhs {				/* LHS of a rule */
+	int nr;				/* Nr of the nonterminal */
+	struct symbol *rhs;		/* Pointer to RHS */
+	char first[LLSETSIZE];		/* First set */
+	char follow[LLSETSIZE];		/* Follow set */
+	char empty;			/* Set if nonterminal produces empty */
+};
+
+struct symbol {				/* Symbol in the RHS of a rule */
+	int x;				/* LLTERMINAL or LLNONTERMINAL */
+	int nr;				/* Nr of the symbol */
+	struct symbol *link;		/* Ptr to next rule with this symbol */
+	struct symbol *next;		/* Ptr to next symbol in this rule */
+	struct lhs *lhs;		/* Ptr to LHS */
+};
+
+struct terminal {			/* Array with links to terminals in a */
+	struct symbol *link;		/* rule */	
+};
+
+struct nonterminal {			/* Array with links to nt's in a rule */
+	struct symbol *link;		/* and pointer to LHS's */
+	struct lhs *rule;
+};
+
+struct stack_elt {			/* Stack element */
+	int flags;			/* Some flags */
+	int nr;				/* Nr of symbol */
+	int ref_count;			/* Nr of predecessors */
+	int hyp_ref_count;		/* Temporary nr of predecessors */
+	int matched;			/* Nr of LHS trying to match */
+	int nr_nexts;			/* Nr of successors */	
+	struct edge *edges;		/* Array of edges to other stack elt's*/
+};
+
+/* Possible flags in a stack element */
+#define LLHEAD		1		/* Stack element is a head */
+#define LLDUMMY		2		/* Stack element is substituted */
+#define LLGEN_SEARCH	8		/* Set by 'generate_heads()' */
+
+ 
+struct edge {				/* Edges of a stack element */
+	char flags;			/* Some flags */
+	struct stack_elt *ptr;		/* Array with pointers to stack elt's */
+};
+
+/* Possible flags in an edge */
+#define LLLOOP		 1		/* Belongs to a loop */
+#define LLLOOP_SEARCH	 2		/* Used by 'loop()' */
+#define LLHYP_SEARCH 	 4		/* Used by 'hyp_run()' */
+#define PRINT_SEARCH	 8		/* DEBUG */
+#define LLMARK_SEARCH	16		/* Used by 'mark_loop()' */
+#define LLYES		32	
+#define LLNO		64
+
+#define LLEOSTACK -1			/* Indicates last element of a stack */
+#define LLHEADS_BUF_INCR 10		/* Nr of elements the buffer will be */
+#define LLCLEANUP_BUF_INCR 25 		/* increased by */
+#define LL_VIS_INCR 200
+
+/* Macro's to manipulate bit sets */
+#define LLIN(a, i)	((a)[(i)/8] & (1 << ((i) % 8)))
+#define LLPUTIN(a, i)	((a)[(i)/8] |= (1 << ((i) % 8)))

util/LLgen/lib/rec

@@ -0,0 +1,442 @@
+/* 
+ * Some grammar independent code.
+ * This file is copied into Lpars.c.
+ */
+
+#ifndef lint
+static char *rcsid = "$Id$";
+#endif
+
+unsigned int	LLtcnt[LL_NTERMINALS];
+unsigned int	LLscnt[LL_NSETS];
+int		LLcsymb, LLsymb;
+static int	LLlevel;
+
+#if LL_NON_CORR
+int		LLstartsymb;
+static int	fake_eof = 0;
+#endif
+
+#if LL_ANSI_C
+#define LL_VOIDCST (void)
+void	LLmessage(int);
+#else
+#define LL_VOIDCST
+#endif
+#ifdef LL_USERHOOK
+#if LL_ANSI_C
+static int	LLdoskip(int);
+static int	LLuserhook(int, int*);
+#else
+static int	LLdoskip();
+static int	LLuserhook();
+#endif
+#endif
+
+#ifndef LL_FASTER
+#if LL_ANSI_C
+void LLscan(int t)
+#else
+LLscan(t)
+	int	t;
+#endif
+{
+	/*
+	 * Check if the next symbol is equal to the parameter
+	 */
+
+#if LL_NON_CORR
+	/* See if the error recovery has eaten an eof */
+	if (fake_eof) {
+		LLsymb = EOFILE;
+		fake_eof = 0;
+	}
+	else {
+		LLsymb = LL_LEXI();
+	}
+
+	if (LLsymb == t) {
+#else
+	if ((LLsymb = LL_LEXI()) == t) {
+#endif
+
+#if LL_NON_CORR
+		/* Check if a previous parser has 'crashed', in that
+		 * case continue with non-correcting parser 
+		 */
+		if (err_seen && !nc_done) {
+			LLnc_recover();
+			nc_done = 1;
+			/* Remember that the error recovery has eaten an eof */
+			fake_eof = 1;
+			if (t != LLsymb) {
+				LLerror(t);
+			}
+			else
+				return;
+		}
+#endif
+		return;
+	}
+	/*
+	 * If we come here, an error has been detected
+	 */
+	LLerror(t);
+}
+#endif
+
+#if LL_ANSI_C
+void LLread(void) {
+#else
+LLread() {
+#endif
+
+#if LL_NON_CORR
+	/* Again, check if another parser has crashed,
+	 * in that case intercept and go to the
+	 * non-correcting parser
+	 */
+
+	if (err_seen && !nc_done) {
+		LLnc_recover();
+		nc_done = 1;
+		/* Pretend we read end of file */
+		LLsymb = EOFILE;
+		LLcsymb = LLindex[EOFILE];
+		fake_eof = 0;
+		return;
+	}
+
+	if (fake_eof) {
+		LLsymb = EOFILE;
+		LLcsymb = LLindex[EOFILE];
+		fake_eof = 0;
+		return;
+	}
+#endif
+
+	for (;;) {
+		if ((LLcsymb = LLindex[(LLsymb = LL_LEXI())]) >= 0) return;
+		LLmessage(0);
+	}
+	/* NOTREACHED */
+}
+
+#if LL_ANSI_C
+void LLerror(int t)
+#else
+LLerror(t)
+	int	t;
+#endif
+{
+	register int i;
+
+	if (t == EOFILE && LLsymb <= 0) return;
+#ifdef LL_NEWMESS
+	if (t == EOFILE) {
+#ifdef LL_USERHOOK
+		static int lst[] = { EOFILE, 0 };
+		LL_VOIDCST LLuserhook(EOFILE, lst);
+#endif /* LL_USERHOOK */
+		if (LLsymb != EOFILE && LLsymb > 0) {
+			LLmessage(-1);
+			while ((LLsymb = LL_LEXI()) > 0 && LLsymb != EOFILE) 
+				/* nothing */ ;
+		}
+		return;
+	}
+#endif
+
+#if LL_NON_CORR
+	if ((!nc_done) && (LLsymb > 0) && (LLsymb != EOFILE)) {
+		LLmessage(0);
+		LLnc_recover();
+		nc_done = 1;
+		LLsymb = EOFILE;
+	}
+#endif
+	
+	if ((LLcsymb = LLindex[LLsymb]) < 0) {
+		LLmessage(0);
+		LLread();
+	}
+	i = LLindex[t];
+	LLtcnt[i]++;
+#ifdef LL_USERHOOK
+	LL_VOIDCST LLdoskip(t);
+#else
+	LL_VOIDCST LLskip();
+#endif
+	LLtcnt[i]--;
+	if (LLsymb != t) {
+#if LL_NON_CORR
+		/* A little kludge here; when using non-correcting recovery
+		 * it can happen that a program is correct but incomplete. 
+		 * Here, we test this, and make sure the appropriate 
+		 * message is generated 
+	 	 */
+		if (! nc_done) {
+			int oldLLsymb;
+			oldLLsymb = LLsymb;
+			LLsymb = EOFILE;
+			LLmessage(0);
+			nc_done = 1;	
+		/* Not really, but to prevent more than 1 error message */
+			LLsymb = oldLLsymb;
+		}
+#endif
+		LLmessage(t);
+	}
+}
+
+#if LL_ANSI_C
+void LLsafeerror(int t)
+#else
+LLsafeerror(t)
+	int	t;
+#endif
+{
+	if (t == EOFILE && LLsymb <= 0) return;
+#ifdef LL_NEWMESS
+	if (t == EOFILE) {
+#ifdef LL_USERHOOK
+		static int lst[] = { EOFILE, 0 };
+		LL_VOIDCST LLuserhook(EOFILE, lst);
+#endif /* LL_USERHOOK */
+		if (LLsymb != EOFILE && LLsymb > 0) {
+			LLmessage(-1);
+			while ((LLsymb = LL_LEXI()) > 0 && LLsymb != EOFILE) 
+				/* nothing */ ;
+		}
+		return;
+	}
+#endif
+#if LL_NON_CORR
+	if ((!nc_done) && (LLsymb > 0) && (LLsymb != EOFILE)) {
+		LLmessage(0);
+		LLnc_recover();
+		nc_done = 1;
+		LLsymb = EOFILE;
+	}
+	/* A little kludge here; when using non-correcting recovery
+         * it can happen that a program is correct but incomplete.
+         * Here, we test this, and make sure the appropriate
+         * message is generated
+         */
+        if (! nc_done) {
+                int oldLLsymb;
+                oldLLsymb = LLsymb;
+                LLsymb = EOFILE;
+                LLmessage(0);
+                nc_done = 1;
+        /* Not really, but to prevent more than 1 error message */
+                LLsymb = oldLLsymb;
+        }
+#endif
+	LLmessage(t);
+}
+
+#ifndef LLNOFIRSTS
+#if LL_ANSI_C
+int LLfirst(int x, int d) {
+#else
+int LLfirst(x, d) {
+#endif
+	register int i;
+
+	return (i = LLindex[x]) >= 0 &&
+		(LLsets[d + (i >> 3)] & (1 << (i & 07)));
+}
+#endif
+
+#if LL_ANSI_C
+int LLnext(int n)
+#else
+int LLnext(n)
+	int	n;
+#endif
+{
+	/*	returns: 0 if the current symbol is'nt skipped, and it
+			 is'nt a member of "n",
+			 1 if we have a new symbol, but it is'nt a member,
+			 2 if the current symbol is a member,
+			 and 3 if we have a new symbol and it is a member.
+		So, the low order bit indicates wether we have a new symbol,
+		and the next bit indicates wether it is a member of "n".
+	*/
+
+	int retval = 0;
+
+	if (LLskip()) retval = 1;
+	if (n <= 0 && LLsets[(LLcsymb >> 3) - n] & (1 << (LLcsymb & 07))) {
+		retval |= 2;
+	}
+	else if (n > 0 && LLcsymb == LLindex[n]) retval |= 2;
+	return retval;
+}
+
+#if LL_ANSI_C
+int LLskip(void) {
+#else
+int LLskip() {
+#endif
+	/*	returns 0 if the current symbol is'nt skipped, and
+		1 if it is, t.i., we have a new symbol
+	*/
+#ifdef LL_USERHOOK
+	return LLdoskip(0);
+}
+
+#if LL_ANSI_C
+extern void LL_USERHOOK(int, int *);
+static int LLuserhook(int e, int *list)
+#else
+static int LLuserhook(e, list)
+	int e;
+	int *list;
+#endif
+{
+	int old = LLsymb;
+	LL_USERHOOK(e, list);
+	LLread();
+	return LLsymb != old;
+}
+
+#if LL_ANSI_C
+static void LLmklist(register int *list)
+#else
+static LLmklist(list)
+	register int *list;
+#endif
+{
+	char Xset[LL_SSIZE];
+	register char *p;
+	register int i;
+
+	for (p = &Xset[0]; p < &Xset[LL_SSIZE]; ) *p++ = 0;
+	for (i = 0; i < LL_NTERMINALS; i++) {
+		if (LLtcnt[i] != 0) Xset[i >> 3] |= (1 << (i & 07));
+	}
+	for (i = LL_NSETS - 1; i >= 0; i--) if (LLscnt[i] != 0) {
+		register char *q = &LLsets[LL_SSIZE * i];
+
+		p = &Xset[0];
+		while (p < &Xset[LL_SSIZE]) *p++ |= *q++;
+	}
+	for (i = 0; i < LL_NTERMINALS; i++) {
+		if (Xset[i >> 3] & (1 << (i & 07))) {
+			*list++ = LLtok[i];
+		}
+	}
+	*list = 0;
+}
+
+#if LL_ANSI_C
+static int LLdoskip(int e)
+#else
+static int LLdoskip(e)
+	int	e;
+#endif
+{
+	int LLx;
+	int list[LL_NTERMINALS+1];
+#endif /* LL_USERHOOK */
+	register int i;
+	int retval;
+	int LLi, LLb;
+
+	retval = 0;
+#ifdef LL_USERHOOK
+	LLmklist(list);
+	LLx = LLuserhook(e, list);
+	if (LLx) retval = 1;
+#endif /* LL_USERHOOK */
+	for (;;) {
+		if (LLtcnt[LLcsymb] != 0) {
+#ifdef LL_USERHOOK
+			if (!e || !LLx || LLcsymb == LLindex[e])
+#endif
+			return retval;
+		}
+		LLi = LLcsymb >> 3;
+		LLb = 1 << (LLcsymb & 07);
+		for (i = LL_NSETS - 1; i >= 0; i--) {
+			if (LLscnt[i] != 0) {
+				if (LLsets[LL_SSIZE*i+LLi] & LLb) {
+#ifdef LL_USERHOOK
+					if (!e || !LLx || LLcsymb == LLindex[e])
+#endif
+					return retval;
+				}
+			}
+		}
+#ifdef LL_USERHOOK
+		if (LLx) {
+			LLx = LLuserhook(e, list);
+			continue;
+		}
+#endif /* LL_USERHOOK */
+#if LL_NON_CORR
+		if ((!nc_done) && (LLsymb > 0)) {
+			LLmessage(0);
+			LLnc_recover();
+			nc_done = 1;
+			fake_eof = 1;
+		}
+		else {
+			LLmessage(0);
+		}
+#else
+		LLmessage(0);
+#endif
+		retval = 1;
+		LLread();
+	}
+	/* NOTREACHED */
+}
+
+#if LL_ANSI_C
+void LLnewlevel(unsigned int *LLsinfo) {
+#else
+LLnewlevel(LLsinfo) unsigned int *LLsinfo; {
+#endif
+	register int i;
+
+	if (LLlevel++) {
+		LLsinfo[LL_NSETS+LL_NTERMINALS] = (unsigned) LLsymb;
+		LLsinfo[LL_NSETS+LL_NTERMINALS+1] = (unsigned) LLcsymb;
+		for (i = LL_NTERMINALS - 1; i >= 0; i--) {
+			LLsinfo[i] = LLtcnt[i];
+			LLtcnt[i] = 0;
+		}
+		for (i = LL_NSETS - 1; i >= 0; i--) {
+			LLsinfo[LL_NTERMINALS+i] = LLscnt[i];
+			LLscnt[i] = 0;
+		}
+	}
+	LLtincr(0);
+}
+
+#if LL_ANSI_C
+void LLoldlevel(unsigned int *LLsinfo) {
+#else
+LLoldlevel(LLsinfo) unsigned int *LLsinfo; {
+#endif
+	register int i;
+
+	LLtdecr(0);
+#ifdef LL_DEBUG
+	for (i = 0; i < LL_NTERMINALS; i++) LL_assert(LLtcnt[i] == 0);
+	for (i = 0; i < LL_NSETS; i++) LL_assert(LLscnt[i] == 0);
+#endif
+	if (--LLlevel) {
+		for (i = LL_NSETS - 1; i >= 0; i--) {
+			LLscnt[i] = LLsinfo[LL_NTERMINALS+i];
+		}
+		for (i = LL_NTERMINALS - 1; i >= 0; i--) {
+			LLtcnt[i] = LLsinfo[i];
+		}
+		LLsymb = (int) LLsinfo[LL_NSETS+LL_NTERMINALS];
+		LLcsymb = (int) LLsinfo[LL_NSETS+LL_NTERMINALS+1];
+	}
+}

util/LLgen/mkdistr.sh

@@ -0,0 +1,64 @@
+#!/bin/sh
+# $Id$
+#
+# Script to build a distribution package.
+
+set -e
+	
+if [ "$1" = "" ]; then
+	echo "Please specify a version number!"
+	exit 1
+fi
+
+if [ -d ../LLgen-$1 ]; then
+	echo "I think you've already generated version $1."
+	exit 1
+fi
+
+mkdir ../LLgen-$1
+cp -a --parents \
+	pm \
+	c.pm \
+	pmfile \
+	doc/LLgen.refs \
+	doc/LLgen_NCER.n \
+	doc/LLgen.1 \
+	doc/LLgen.n \
+	lib/rec \
+	lib/incl \
+	lib/nc_incl \
+	lib/nc_rec \
+	src/io.h \
+	src/extern.h \
+	src/reach.c \
+	src/LLgen.c \
+	src/LLgen.g \
+	src/compute.c \
+	src/savegram.c \
+	src/global.c \
+	src/gencode.c \
+	src/main.c \
+	src/name.c \
+	src/types.h \
+	src/sets.c \
+	src/sets.h \
+	src/tokens.c \
+	src/tokens.g \
+	src/Lpars.c \
+	src/Lpars.h \
+	src/cclass.c \
+	src/cclass.h \
+	src/alloc.c \
+	src/check.c \
+	src/machdep.c \
+	mkdistr.sh \
+	bootstrap.sh \
+	COPYING \
+	READ_ME \
+\
+	../LLgen-$1
+
+(cd .. && tar cvf LLgen-$1.tar.bz2 --bzip2 LLgen-$1)
+
+echo ""
+echo "Done --- but did you remember to update the version number in the README?"

util/LLgen/pmfile

@@ -0,0 +1,103 @@
+-- $Id$
+--
+-- This is the build file used to compile LLgen. It should be run through
+-- Prime Mover (copy supplied). See the READ_ME file for more information.
+
+include "c.pm"
+
+-- Where is LLgen going to be installed eventually? (Needs trailing slash.)
+
+PREFIX = PREFIX or "/usr/local/"
+
+-- Where's LLgen's staging area? (Don't change. Needs trailing slash.)
+
+INSTALLPATH = "bin/"
+
+LLgen = cprogram {
+	CEXTRAFLAGS = '-DLIBDIR=\\"'..PREFIX..'share/LLgen\\" -DNON_CORRECTING',
+
+	cfile "src/main.c",
+	cfile "src/gencode.c",
+	cfile "src/compute.c",
+	cfile "src/check.c",
+	cfile "src/reach.c",
+	cfile "src/global.c",
+	cfile "src/name.c",
+	cfile "src/sets.c",
+	cfile "src/alloc.c",
+	cfile "src/machdep.c",
+	cfile "src/cclass.c",
+	cfile "src/savegram.c",
+
+	-- These use pre-LLgen'd version of the files. If LLgen.g gets updated,
+	-- they need rebuilding. Use the bootstrap script to do this.
+	
+	cfile "src/LLgen.c",
+	cfile "src/Lpars.c",
+	cfile "src/tokens.c",
+
+	outputs = {"%U%/LLgen"},
+	install = pm.install(              INSTALLPATH.."bin/LLgen")
+}
+
+library = group {
+	install = {
+		pm.install("lib/rec",          INSTALLPATH.."share/LLgen/rec"),
+		pm.install("lib/incl",         INSTALLPATH.."share/LLgen/incl"),
+		pm.install("lib/nc_incl",      INSTALLPATH.."share/LLgen/nc_incl"),
+		pm.install("lib/nc_rec",       INSTALLPATH.."share/LLgen/nc_rec"),
+	}
+}
+
+manpage = group {
+	install = {
+		pm.install("doc/LLgen.1", INSTALLPATH.."man/man1/LLgen.1"),
+	}
+}
+
+documentation = group {
+	simple {
+		outputs = {"%U%-%I%.ps.gz"},
+		command = "refer -sA+T -p %in[1]% %in[2]% | groff -Tps -e -t -ms "..
+			"| gzip -c9 > %out[1]%",
+		
+		file "doc/LLgen.refs",
+		file "doc/LLgen.n",
+		
+		install = {
+			pm.install(INSTALLPATH.."share/doc/LLgen/LLgen.ps.gz")
+		}
+	},
+	
+	simple {
+		outputs = {"%U%-%I%.ps.gz"},
+		command = "groff -Tps -e -t -p -ms %in% | gzip -c9 > %out[1]%",
+		
+		file "doc/LLgen_NCER.n",
+		
+		install = {
+			pm.install(INSTALLPATH.."share/doc/LLgen/NCER.ps.gz")
+		}
+	},
+}
+
+-- Default rule: builds everything into the staging area, but does nothing
+-- else.
+
+default = group {
+	LLgen,         -- build LLgen itself
+	library,       -- copy over the library
+	manpage,       -- copy over the man page
+	documentation, -- build the two white papers
+}
+
+-- This rule will build everything, and then install it to its final location.
+
+install = group {
+	default,
+	
+	install = {
+		"mkdir -p %PREFIX%",
+		"(cd bin && tar cf - .) | (cd %PREFIX% && tar xvf -)"
+	}
+}

util/LLgen/src/LLgen.g

@@ -0,0 +1,665 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * LLgen.g
+ * Defines the grammar of LLgen.
+ * Some routines that build the internal structure are also included
+ */
+
+{
+# include <stdlib.h>
+# include <string.h>
+# include "types.h"
+# include "io.h"
+# include "extern.h"
+# include "assert.h"
+# include "cclass.h"
+
+# ifndef NORCSID
+static string	rcsid = "$Id$";
+# endif
+p_mem		alloc(), ralloc();
+string		store();
+p_gram		search();
+long		ftell();
+
+static int	acount;			/* count #of global actions */
+static p_term t_list;
+static int t_cnt;
+static p_gram	alt_table;
+static int	n_alts;
+static int	max_alts;
+#define ALTINCR 32
+
+static p_gram	rule_table;
+static int	n_rules;
+static int	max_rules;
+#define RULEINCR	32
+
+/* Here are defined : */
+STATIC		newnorder();
+STATIC		newtorder();
+STATIC		mkalt();
+STATIC		mkterm();
+STATIC p_gram	copyrule();
+/* and of course LLparse() */
+
+STATIC
+newnorder(index) {
+	static int porder;
+
+	if (norder != -1) {
+		nonterms[porder].n_next = index;
+	}
+	else	norder = index;
+	porder = index;
+	nonterms[porder].n_next = -1;
+}
+
+STATIC
+newtorder(index) {
+	static int porder;
+
+	if (torder != -1) {
+		tokens[porder].t_next = index;
+	}
+	else	torder = index;
+	porder = index;
+	tokens[porder].t_next = -1;
+}
+
+p_init()
+{
+	alt_table = (p_gram )alloc(ALTINCR*sizeof(t_gram));
+	n_alts = 0;
+	max_alts = ALTINCR;
+	rule_table = (p_gram )alloc(RULEINCR*sizeof(t_gram));
+	n_rules = 0;
+	max_rules = RULEINCR;
+}
+
+}
+
+%start	LLparse, spec;
+
+spec	:		{	acount = 0; p_init(); }
+	  [ %persistent def ]*
+			{	/*
+				 * Put an endmarker in temporary file
+				 */
+				putc('\0', fact);
+				putc('\0', fact);
+				free((p_mem) rule_table);
+				free((p_mem) alt_table);
+			}
+	;
+
+def			{	register string p; }
+	: rule
+	  /*
+	   * A grammar rule
+	   */
+	| C_TOKEN listel [ ',' listel ]* ';'
+	  /*
+	   * A token declaration
+	   */
+	| C_START C_IDENT
+			{	p = store(lextoken.t_string); }
+	  ',' C_IDENT
+	  /*
+	   * A start symbol declaration
+	   */
+			{	/*
+				 * Put the declaration in the list
+				 * of start symbols
+				 */
+				register p_gram temp;
+				register p_start ff;
+
+				temp = search(NONTERM,lextoken.t_string,BOTH);
+				ff = (p_start) alloc(sizeof(t_start));
+				ff->ff_nont = g_getcont(temp);
+				ff->ff_name = p;
+				ff->ff_next = start;
+				start = ff;
+				while (ff = ff->ff_next) {
+					if (! strcmp(p, ff->ff_name)) {
+						error(linecount, "\"%s\" already used in a %%start", p);
+						break;
+					}
+				}
+			}
+	  ';'
+	| C_LEXICAL C_IDENT
+	  /*
+	   * Declaration of a name for the lexical analyser.
+	   * May appear only once
+	   */
+			{	if (!lexical) {
+					lexical = store(lextoken.t_string);
+				}
+				else	error(linecount,"Duplicate %%lexical");
+			}
+	  ';'
+	| C_PREFIX C_IDENT
+	  /*
+	   * Prefix of external names (default: LL)
+	   */
+			{	if (!prefix) {
+					prefix = store(lextoken.t_string);
+					if (strlen(prefix) > 6) {
+						error(linecount,
+							"%%prefix too long");
+						prefix[6] = 0;
+					}
+				}
+				else	error(linecount,"Duplicate %%prefix");
+			}
+	  ';'
+	| C_ONERROR C_IDENT
+			{
+#ifdef NON_CORRECTING
+				if (non_corr) {
+					warning(linecount, "%%onerror conflicts with -n option");
+				}
+				else
+#endif
+				  if (! onerror) {
+					onerror = store(lextoken.t_string);
+				}
+				else	error(linecount,"Duplicate %%onerror");
+			}
+	  ';'
+	| C_ACTION	{	acount++; }
+	  /*
+	   * A global C-declaration
+	   */
+	| firsts
+	  /*
+	   * declarations for macros
+	   */
+	;
+
+listel	: C_IDENT	{	p_gram temp = search(TERMINAL,lextoken.t_string,ENTERING);
+				newtorder(g_getcont(temp));
+				tokens[g_getcont(temp)].t_lineno = linecount;
+			}
+	;
+
+rule			{	register p_nont p;
+				p_gram rr;
+				register p_gram temp;
+			}
+	: /*
+	   * grammar for a production rule
+	   */
+	  C_IDENT	{	temp = search(NONTERM,lextoken.t_string,BOTH);
+				p = &nonterms[g_getcont(temp)];
+				if (p->n_rule) {
+					error(linecount,
+"Nonterminal %s already defined", lextoken.t_string);
+				}
+				/*
+				 * Remember the order in which the nonterminals
+				 * were defined. Code must be generated in that
+				 * order to keep track with the actions on the
+				 * temporary file
+				 */
+				newnorder(p - nonterms);
+				p->n_count = acount;
+				acount = 0;
+				p->n_lineno = linecount;
+				p->n_off = ftell(fact);
+			}
+	  [ C_PARAMS	{	if (lextoken.t_num > 0) {
+					p->n_flags |= PARAMS;
+					if (lextoken.t_num > 15) {
+						error(linecount,"Too many parameters");
+					}
+					else	setntparams(p,lextoken.t_num);
+				}
+			}
+	  ]?
+	  [ C_ACTION	{	p->n_flags |= LOCALS; }
+	  ]?
+	  ':' 		{ 	in_production = 1; }
+	  productions(&rr) ';'
+			{	in_production = 0; }
+			/*
+			 * Do not use p->n_rule now! The nonterms array
+			 * might have been re-allocated.
+			 */
+			{	nonterms[g_getcont(temp)].n_rule = rr;}
+	;
+
+productions(p_gram *p;)
+	/*
+	 * One or more alternatives
+	 */
+	{	p_gram		prod;
+		int		conflres = 0;
+		int		t = 0;
+		int		haddefault = 0;
+		int		altcnt = 0;
+		int		o_lc, n_lc;
+	} :
+			{	o_lc = linecount; }
+	  simpleproduction(p,&conflres)
+			{	if (conflres & DEF) haddefault = 1; }
+	  [
+	    [ '|'	{	n_lc = linecount; }
+	      simpleproduction(&prod,&t)
+			{	if (n_alts >= max_alts-2) {
+					alt_table = (p_gram ) ralloc(
+						(p_mem) alt_table,
+						(unsigned)(max_alts+=ALTINCR)*sizeof(t_gram));
+				}
+				if (t & DEF) {
+					if (haddefault) {
+						error(n_lc,
+		"More than one %%default in alternation");
+					}
+					haddefault = 1;
+				}
+				mkalt(*p,conflres,o_lc,&alt_table[n_alts++]);
+				altcnt++;
+				o_lc = n_lc;
+				conflres = t;
+				t = 0;
+				*p = prod;
+			}
+	    ]+		{	if (conflres & (COND|PREFERING|AVOIDING)) {
+					error(n_lc,
+		"Resolver on last alternative not allowed");
+				}
+				mkalt(*p,conflres,n_lc,&alt_table[n_alts++]);
+				altcnt++;
+				g_settype((&alt_table[n_alts]),EORULE);
+				*p = copyrule(&alt_table[n_alts-altcnt],altcnt+1);
+			}
+	  |
+			{	if (conflres & (COND|PREFERING|AVOIDING)) {
+					error(o_lc,
+		"No alternation conflict resolver allowed here");
+				}
+				/*
+				if (conflres & DEF) {
+					error(o_lc,
+		"No %%default allowed here");
+				}
+				*/
+			}
+	  ]
+			{	n_alts -= altcnt; }
+	;
+{
+
+STATIC
+mkalt(prod,condition,lc,res) p_gram prod; register p_gram res; {
+	/*
+	 * Create an alternation and initialise it.
+	 */
+	register p_link		l;
+	static p_link list;
+	static int cnt;
+
+	if (! cnt) {
+		cnt = 50;
+		list = (p_link) alloc(50 * sizeof(t_link));
+	}
+	cnt--;
+	l = list++;
+	l->l_rule = prod;
+	l->l_flag = condition;
+	g_setlink(res,l);
+	g_settype(res,ALTERNATION);
+	res->g_lineno = lc;
+	nalts++;
+}
+}
+
+simpleproduction(p_gram *p; register int *conflres;)
+	{	t_gram		elem;
+		int		elmcnt = 0;
+		int		cnt, kind;
+		int		termdeleted = 0;
+	} :
+	  [ C_DEFAULT	{	*conflres |= DEF; }
+	  ]?
+	  [
+	    /*
+	     * Optional conflict reslover
+	     */
+	      C_IF C_EXPR {	*conflres |= COND; }
+	    | C_PREFER	{	*conflres |= PREFERING; }
+	    | C_AVOID	{	*conflres |= AVOIDING; }
+	  ]?
+	  [ C_ILLEGAL	{
+#ifdef NON_CORRECTING
+				if (n_rules >= max_rules-2) {
+					rule_table = (p_gram) ralloc(
+						  (p_mem) rule_table,
+						  (unsigned)(max_rules+=RULEINCR)*sizeof(t_gram));
+				}
+				elmcnt++;
+				rule_table[n_rules++] =
+				    *search(TERMINAL, "LLILLEGAL", BOTH);
+				if (*conflres & DEF) {
+					error(linecount, "%%illegal not allowed in %%default rule");
+				}
+#endif
+			}
+	  ]?
+	  [ %persistent elem(&elem)
+			{	if (n_rules >= max_rules-2) {
+					rule_table = (p_gram) ralloc(
+						  (p_mem) rule_table,
+						  (unsigned)(max_rules+=RULEINCR)*sizeof(t_gram));
+				}
+				kind = FIXED;
+				cnt = 0;
+			}
+	    [ repeats(&kind, &cnt)
+			{	if (g_gettype(&elem) != TERM) {
+					rule_table[n_rules] = elem;
+					g_settype((&rule_table[n_rules+1]),EORULE);
+					mkterm(copyrule(&rule_table[n_rules],2),
+					       0,
+					       elem.g_lineno,
+					       &elem);
+				}
+			}
+	    |
+			{ if (g_gettype(&elem) == TERM) {
+				register p_term q = g_getterm(&elem);
+
+				if (! (q->t_flags & RESOLVER) &&
+				    g_gettype(q->t_rule) != ALTERNATION &&
+				    g_gettype(q->t_rule) != EORULE) {
+				    while (g_gettype(q->t_rule) != EORULE) {
+					rule_table[n_rules++] = *q->t_rule++;
+					elmcnt++;
+					if (n_rules >= max_rules-2) {
+					    rule_table = (p_gram) ralloc(
+						  (p_mem) rule_table,
+						  (unsigned)(max_rules+=RULEINCR)*sizeof(t_gram));
+					}
+				    }
+				    elem = *--(q->t_rule);
+				    n_rules--;
+				    elmcnt--;
+				    if (q == t_list - 1) {
+					t_list--;
+					nterms--;
+					t_cnt++;
+				    }
+				    termdeleted = 1;
+				}
+			  }
+			}
+	    ]		{	if (!termdeleted && g_gettype(&elem) == TERM) {
+					register p_term q;
+
+					q = g_getterm(&elem);
+					r_setkind(q,kind);
+					r_setnum(q,cnt);
+					if ((q->t_flags & RESOLVER) &&
+					    (kind == PLUS || kind == FIXED)) {
+						error(linecount,
+		"%%while not allowed in this term");
+					}
+					/*
+					 * A persistent fixed term is the same
+					 * as a non-persistent fixed term.
+					 * Should we complain?
+					if ((q->t_flags & PERSISTENT) &&
+					    kind == FIXED) {
+						error(linecount,
+							"Illegal %%persistent");
+					}
+					*/
+				}
+				termdeleted = 0;
+				elmcnt++;
+				rule_table[n_rules++] = elem;
+			}
+	  ]*		{	register p_term q;
+
+				g_settype((&rule_table[n_rules]),EORULE);
+				*p = 0;
+				n_rules -= elmcnt;
+				if (g_gettype(&rule_table[n_rules]) == TERM &&
+				    elmcnt == 1) {
+					q = g_getterm(&rule_table[n_rules]);
+					if (r_getkind(q) == FIXED &&
+					    r_getnum(q) == 0) {
+						*p = q->t_rule;
+					}
+				}
+				if (!*p) *p = copyrule(&rule_table[n_rules],
+						elmcnt+1);
+			}
+	;
+{
+
+STATIC
+mkterm(prod,flags,lc,result) p_gram prod; register p_gram result; {
+	/*
+	 * Create a term, initialise it and return
+	 * a grammar element containing it
+	 */
+	register p_term		q;
+
+	if (! t_cnt) {
+		t_cnt = 50;
+		t_list = (p_term) alloc(50 * sizeof(t_term));
+	}
+	t_cnt--;
+	q = t_list++;
+	q->t_rule = prod;
+	q->t_contains = 0;
+	q->t_flags = flags;
+	g_settype(result,TERM);
+	g_setterm(result,q);
+	result->g_lineno = lc;
+	nterms++;
+}
+}
+
+elem (register p_gram pres;)
+	{	register int	t = 0;
+		p_gram		p1;
+		int		ln;
+		p_gram		pe;
+#ifdef NON_CORRECTING
+		int		erroneous = 0;
+#endif
+	} :
+	  '['		{	ln = linecount; }
+	  [ C_WHILE C_EXPR	{	t |= RESOLVER; }
+	  ]?
+	  [ C_PERSISTENT	{	t |= PERSISTENT; }
+	  ]?
+	  productions(&p1)
+	  ']'		{
+				mkterm(p1,t,ln,pres);
+			}
+	|
+	  [ C_ERRONEOUS		{
+#ifdef NON_CORRECTING
+					erroneous = 1;
+#endif
+				}
+	  ]?
+
+	  [
+	  C_IDENT	{	pe = search(UNKNOWN,lextoken.t_string,BOTH);
+				*pres = *pe;
+#ifdef NON_CORRECTING
+				if (erroneous) {
+					if (g_gettype(pres) != TERMINAL){
+						warning(linecount,
+							"Erroneous only allowed on terminal");
+						erroneous = 0;
+					}
+					else
+						pres->g_erroneous = 1;
+				}
+#endif
+
+			}
+	  [ C_PARAMS	{	if (lextoken.t_num > 15) {
+					error(linecount,"Too many parameters");
+				} else	g_setnpar(pres,lextoken.t_num);
+				if (g_gettype(pres) == TERMINAL) {
+					error(linecount,
+						"Terminal with parameters");
+				}
+			}
+	  ]?
+	| C_LITERAL	{	pe = search(LITERAL,lextoken.t_string,BOTH);
+				*pres = *pe;
+#ifdef NON_CORRECTING
+				if (erroneous)
+					pres->g_erroneous = 1;
+#endif
+			}
+	  ]
+	|		{	g_settype(pres,ACTION);
+				pres->g_lineno = linecount;
+#ifdef NON_CORRECTING
+				g_setsubparse(pres, (p_start) 0);
+#endif
+			}
+
+	  [ C_SUBSTART
+
+			{
+#ifdef NON_CORRECTING
+				nsubstarts++;
+#endif
+			}
+
+	    C_IDENT
+			{
+#ifdef NON_CORRECTING
+				register p_gram temp;
+				register p_start subp;
+
+				temp = search(NONTERM,lextoken.t_string,BOTH);
+				subp = (p_start) alloc (sizeof(t_start));
+
+				subp->ff_nont = g_getcont(temp);
+				subp->ff_name = (string) 0;
+				subp->ff_next = (p_start) 0;
+				g_setsubparse(pres, subp);
+#endif
+			}
+
+		[ ',' C_IDENT
+			{
+#ifdef NON_CORRECTING
+				register p_gram temp;
+				register p_start ff;
+
+				temp = search(NONTERM,lextoken.t_string,BOTH);
+
+				ff = g_getsubparse(pres);
+				while (ff) {
+					if (ff->ff_nont == g_getcont(temp)) {
+						warning(linecount, "\"%s\" used twice in %%substart", lextoken.t_string);
+						break;
+					}
+					ff = ff->ff_next;
+
+				}
+
+				ff = (p_start) alloc(sizeof(t_start));
+				ff->ff_nont = g_getcont(temp);
+				ff->ff_name = (string) 0;
+				ff->ff_next = g_getsubparse(pres);
+				g_setsubparse(pres, ff);
+#endif
+			}
+
+	       ]* ';'
+	   ]?
+
+	  C_ACTION
+	;
+
+repeats(int *kind; int *cnt;)	{	int t1 = 0; } :
+	  [
+	    '?'		{	*kind = OPT; }
+	  | [ '*'	{	*kind = STAR; }
+	    | '+'	{	*kind = PLUS; }
+	    ]
+	    number(&t1)?
+			{	if (t1 == 1) {
+					t1 = 0;
+					if (*kind == STAR) *kind = OPT;
+					if (*kind == PLUS) *kind = FIXED;
+				}
+			}
+	  | number(&t1)
+	  ]		{	*cnt = t1; }
+	;
+
+number(int *t;)
+	: C_NUMBER
+			{	*t = lextoken.t_num;
+				if (*t <= 0 || *t >= 8192) {
+					error(linecount,"Illegal number");
+				}
+			}
+	;
+
+firsts	{	register string p; }
+	: C_FIRST C_IDENT
+			{	p = store(lextoken.t_string); }
+	  ',' C_IDENT ';'
+			{	/*
+				 * Store this %first in the list belonging
+				 * to this input file
+				 */
+				p_gram temp;
+				register p_first ff;
+
+				temp = search(NONTERM,lextoken.t_string,BOTH);
+				ff = (p_first) alloc(sizeof(t_first));
+				ff->ff_nont = g_getcont(temp);
+				ff->ff_name = p;
+				ff->ff_next = pfile->f_firsts;
+				pfile->f_firsts = ff;
+			}
+	;
+{
+
+STATIC p_gram
+copyrule(p,length) register p_gram p; {
+	/*
+	 * Returns a pointer to a grammar rule that was created in
+	 * p. The space pointed to by p can now be reused
+	 */
+	register p_gram t;
+	p_gram rule;
+
+	t = (p_gram) alloc((unsigned) length * sizeof(t_gram));
+	rule = t;
+	while (length--) {
+		*t++ = *p++;
+	}
+	return rule;
+}
+}

util/LLgen/src/alloc.c

@@ -0,0 +1,90 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * alloc.c
+ * Interface to malloc() and realloc()
+ */
+
+#include <stdlib.h>
+# include "types.h"
+# include "extern.h"
+
+# ifndef NORCSID
+static string rcsid = "$Id$";
+# endif
+
+static string e_nomem = "Out of memory";
+
+p_mem
+alloc(size) unsigned size; {
+	/*
+	   Allocate "size" bytes. Panic if it fails
+	 */
+	p_mem	p;
+
+	if ((p = malloc(size)) == 0) fatal(linecount,e_nomem);
+	return p;
+}
+
+p_mem
+ralloc(p,size) p_mem p; unsigned size; {
+	/*
+	   Re-allocate the chunk of memory indicated by "p", to
+	   occupy "size" bytes
+	 */
+	if ((p = realloc(p,size)) == 0) fatal(linecount,e_nomem);
+	return p;
+}
+
+p_mem
+new_mem(p) register p_info p; {
+	/*
+	   This routine implements arrays that can grow.
+	   It must be called every time a new element is added to it.
+	   Also, the array has associated with it a "info_alloc" structure,
+	   which contains info on the element size, total allocated size,
+	   a pointer to the array, a pointer to the first free element,
+	   and a pointer to the top.
+	   If the base of the array is remembered elsewhere, it should
+	   be updated each time this routine is called
+	 */
+	p_mem	rp;
+	unsigned sz;
+
+	if (p->i_max >= p->i_top) {	/* No more free elements */
+		sz = p->i_size;
+		if (sizeof(char *) > 2) {
+			/*
+			   Do not worry about size. Just double it.
+			 */
+			p->i_size += p->i_size;
+			if (! p->i_size)
+				p->i_size += p->i_incr * p->i_esize;
+			}
+		else {
+			/*
+			   Worry about size, so only increment in chunks of i_incr.
+			 */
+			p->i_size += p->i_incr * p->i_esize;
+		}
+		p->i_ptr = !p->i_ptr ?
+			alloc(p->i_size) :
+			ralloc(p->i_ptr, p->i_size);
+		p->i_max = p->i_ptr + sz;
+		p->i_top = p->i_ptr + p->i_size;
+	}
+	rp = p->i_max;
+	p->i_max += p->i_esize;
+	return rp;
+}

util/LLgen/src/cclass.c

@@ -0,0 +1,138 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/* $Id$ */
+
+#include "cclass.h"
+char c_class[] = {
+	0,	/* 00 */
+	0,	/* 01 */
+	0,	/* 02 */
+	0,	/* 03 */
+	0,	/* 04 */
+	0,	/* 05 */
+	0,	/* 06 */
+	0,	/* 07 */
+	0,	/* 010 */
+	ISSPA,	/* 011 */
+	ISSPA,	/* 012 */
+	0,	/* 013 */
+	ISSPA,	/* 014 */
+	ISSPA,	/* 015 */
+	0,	/* 016 */
+	0,	/* 017 */
+	0,	/* 020 */
+	0,	/* 021 */
+	0,	/* 022 */
+	0,	/* 023 */
+	0,	/* 024 */
+	0,	/* 025 */
+	0,	/* 026 */
+	0,	/* 027 */
+	0,	/* 030 */
+	0,	/* 031 */
+	0,	/* 032 */
+	0,	/* 033 */
+	0,	/* 034 */
+	0,	/* 035 */
+	0,	/* 036 */
+	0,	/* 037 */
+	ISSPA,	/* ' ' */
+	0,	/* '!' */
+	0,	/* '"' */
+	0,	/* '#' */
+	0,	/* '$' */
+	ISKEY,	/* '%' */
+	0,	/* '&' */
+	ISLIT,	/* ''' */
+	ISACT,	/* '(' */
+	0,	/* ')' */
+	ISTOK,	/* '*' */
+	ISTOK,	/* '+' */
+	ISTOK,	/* ',' */
+	0,	/* '-' */
+	0,	/* '.' */
+	ISCOM,	/* '/' */
+	ISDIG,	/* '0' */
+	ISDIG,	/* '1' */
+	ISDIG,	/* '2' */
+	ISDIG,	/* '3' */
+	ISDIG,	/* '4' */
+	ISDIG,	/* '5' */
+	ISDIG,	/* '6' */
+	ISDIG,	/* '7' */
+	ISDIG,	/* '8' */
+	ISDIG,	/* '9' */
+	ISTOK,	/* ':' */
+	ISTOK,	/* ';' */
+	0,	/* '<' */
+	0,	/* '=' */
+	0,	/* '>' */
+	ISTOK,	/* '?' */
+	0,	/* '@' */
+	ISLET,	/* 'A' */
+	ISLET,	/* 'B' */
+	ISLET,	/* 'C' */
+	ISLET,	/* 'D' */
+	ISLET,	/* 'E' */
+	ISLET,	/* 'F' */
+	ISLET,	/* 'G' */
+	ISLET,	/* 'H' */
+	ISLET,	/* 'I' */
+	ISLET,	/* 'J' */
+	ISLET,	/* 'K' */
+	ISLET,	/* 'L' */
+	ISLET,	/* 'M' */
+	ISLET,	/* 'N' */
+	ISLET,	/* 'O' */
+	ISLET,	/* 'P' */
+	ISLET,	/* 'Q' */
+	ISLET,	/* 'R' */
+	ISLET,	/* 'S' */
+	ISLET,	/* 'T' */
+	ISLET,	/* 'U' */
+	ISLET,	/* 'V' */
+	ISLET,	/* 'W' */
+	ISLET,	/* 'X' */
+	ISLET,	/* 'Y' */
+	ISLET,	/* 'Z' */
+	ISTOK,	/* '[' */
+	0,	/* '\' */
+	ISTOK,	/* ']' */
+	0,	/* '^' */
+	ISLET,	/* '_' */
+	0,	/* '`' */
+	ISLET,	/* 'a' */
+	ISLET,	/* 'b' */
+	ISLET,	/* 'c' */
+	ISLET,	/* 'd' */
+	ISLET,	/* 'e' */
+	ISLET,	/* 'f' */
+	ISLET,	/* 'g' */
+	ISLET,	/* 'h' */
+	ISLET,	/* 'i' */
+	ISLET,	/* 'j' */
+	ISLET,	/* 'k' */
+	ISLET,	/* 'l' */
+	ISLET,	/* 'm' */
+	ISLET,	/* 'n' */
+	ISLET,	/* 'o' */
+	ISLET,	/* 'p' */
+	ISLET,	/* 'q' */
+	ISLET,	/* 'r' */
+	ISLET,	/* 's' */
+	ISLET,	/* 't' */
+	ISLET,	/* 'u' */
+	ISLET,	/* 'v' */
+	ISLET,	/* 'w' */
+	ISLET,	/* 'x' */
+	ISLET,	/* 'y' */
+	ISLET,	/* 'z' */
+	ISACT,	/* '{' */
+	ISTOK,	/* '|' */
+	0,	/* '}' */
+	0,	/* '~' */
+	0	/* 0177 */
+};

util/LLgen/src/cclass.h

@@ -0,0 +1,17 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/* $Id$ */
+
+extern char c_class[];
+
+#define ISLET 1
+#define ISDIG 2
+#define ISSPA 3
+#define ISKEY 4
+#define ISTOK 5
+#define ISCOM 6
+#define ISLIT 7
+#define ISACT 8

util/LLgen/src/check.c

@@ -0,0 +1,474 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * check.c
+ * Several routines to perform checks and printouts
+ */
+
+#include <stdlib.h>
+#include <string.h>
+# include "types.h"
+# include "extern.h"
+# include "io.h"
+# include "sets.h"
+# include "assert.h"
+
+# ifndef NORCSID
+static string rcsid1 = "$Id$";
+# endif
+
+static string	c_first = "> firstset   ";
+static string	c_contains = "> containset ";
+static string	c_follow = "> followset  ";
+p_set		setalloc();
+static int	level;
+
+/* In this file are defined : */
+extern conflchecks();
+STATIC prline();
+STATIC printset();
+STATIC int check();
+STATIC moreverbose();
+STATIC prrule();
+STATIC cfcheck();
+STATIC resolve();
+STATIC propagate();
+STATIC spaces();
+
+conflchecks() {
+	/*
+	 * Check for conflicts, that is,
+	 * in a repeating term, the FIRST and FOLLOW must be disjunct,
+	 * unless there is a disambiguating condition.
+	 * in an alternation, the sets that determine the direction to take,
+	 * must be disjunct.
+	 */
+	register p_nont	p;
+	register int s;
+	p_file		x = files;
+
+	f_input = x->f_name;
+	if (verbose >= 3) {
+		for (p = nonterms; p < maxnt; p++) p->n_flags |= VERBOSE;
+	}
+	if (verbose) {
+		if ((fout = fopen(f_out,"w")) == NULL) fatal(1,e_noopen,f_out);
+	}
+	/*
+	 * Check the rules in the order in which they are declared,
+	 * and input file by input file, to give proper error messages
+	 */
+	for (; x < maxfiles; x++) {
+	    f_input = x->f_name;
+	    for (s = x->f_nonterminals; s != -1; s = p->n_next) {
+		p = &nonterms[s];
+		if (check(p->n_rule)) p->n_flags |= VERBOSE;
+	    }
+	}
+	for (x = files; x < maxfiles; x++) {
+	    f_input = x->f_name;
+	    for (s = x->f_nonterminals; s != -1; s = p->n_next) {
+		p = &nonterms[s];
+		if (p->n_flags & RECURSIVE) {
+			error(p->n_lineno,
+				"Recursion in default for nonterminal %s",
+				p->n_name);
+		}
+		/*
+		 * If a printout is needed for this rule in
+		 * LL.output, just do it
+		 */
+		if (verbose && (p->n_flags & VERBOSE)) {
+			fprintf(fout,"\n%s :\n",p->n_name);
+			printset(p->n_first,c_first);
+			printset(p->n_contains,c_contains);
+			printset(p->n_follow,c_follow);
+			fprintf(fout,"> rule%s\n\t",
+				p->n_flags&EMPTY ? "\t(EMPTY producing)" : "");
+			level = 8;
+			prrule(p->n_rule);
+			level = 0;
+			prline("\n");
+		}
+		/*
+		 * Now, the conflicts may be resolved
+		 */
+		resolve(p->n_rule);
+	    }
+	}
+	if (verbose) fclose(fout);
+}
+
+STATIC
+prline(s) char *s; {
+	fputs(s, fout);
+	spaces();
+}
+
+STATIC
+printset(p,s) register p_set p; string s; {
+	/*
+	 * Print the elements of a set
+	 */
+	register int	i;
+	register int	j;
+	register p_token pt;
+	string		name;
+	int		k;
+	int		hulp;
+
+	k = strlen(s) + 2 + level;
+	/*
+	 * k contains relative level of indentation
+	 */
+	fprintf(fout,"%s{ ",s);
+	j = k;
+	/*
+	 * j will gather the total length of the line
+	 */
+	for (i = 0, pt = tokens; i < ntokens; i++,pt++) {
+		if (IN(p,i)) {
+			hulp = strlen(pt->t_string)+1;
+			if (pt->t_tokno < 0400) hulp += 2;
+			if ((j += hulp) >= 78) {
+				/*
+				 * Line becoming too long
+				 */
+				j = k+hulp;
+				prline("\n");
+				fprintf(fout,">%*c",k - level - 1,' ');
+			}
+			fprintf(fout, pt->t_tokno<0400 ? "'%s' " : "%s ",pt->t_string);
+		}
+	}
+	if (ntprint) for (i = 0; i < nnonterms; i++) {
+		/*
+		 * Nonterminals in the set must also be printed
+		 */
+		if (NTIN(p,i)) {
+			name = nonterms[i].n_name;
+			hulp = strlen(name) + 3;
+			if ((j += hulp) >= 78) {
+				j = k + hulp;
+				prline("\n");
+				fprintf(fout,">%*c",k - level - 1,' ');
+			}
+			fprintf(fout,"<%s> ",name);
+		}
+	}
+	prline("}\n");
+}
+
+STATIC int
+check(p) register p_gram p; {
+	/*
+	 * Search for conflicts in a grammar rule.
+	 */
+	register p_set	temp;
+	register int retval;
+
+	retval = 0;
+	for (;;) {
+		switch (g_gettype(p)) {
+		  case EORULE :
+			return retval;
+		  case NONTERM : {
+			register p_nont n;
+
+			n = &nonterms[g_getcont(p)];
+			if (g_getnpar(p) != getntparams(n)) {
+			    error(p->g_lineno,
+				"Call of %s: parameter count mismatch",
+				n->n_name);
+			}
+			break; }
+		  case TERM : {
+			register p_term q;
+
+			q = g_getterm(p);
+			retval |= check(q->t_rule);
+			if (r_getkind(q) == FIXED) break;
+			if (setempty(q->t_first)) {
+				q->t_flags |= EMPTYFIRST;
+				retval = 1;
+				error(p->g_lineno, "No symbols in term");
+			}
+			if (empty(q->t_rule)) {
+				q->t_flags |= EMPTYTERM;
+				retval = 1;
+				error(p->g_lineno, "Term with variable repetition count produces empty");
+			}
+			temp = setalloc();
+			setunion(temp,q->t_first);
+			if (!setintersect(temp,q->t_follow)) {
+				/*
+				 * q->t_first * q->t_follow != EMPTY
+				 */
+				if (!(q->t_flags & RESOLVER)) {
+					/*
+					 * No conflict resolver
+					 */
+					error(p->g_lineno,
+						"Repetition conflict");
+					retval = 1;
+					moreverbose(temp);
+				}
+			}
+			else {
+				if (q->t_flags & RESOLVER) {
+					q->t_flags |= NOCONF;
+					warning(p->g_lineno,
+						"%%while without conflict");
+				}
+			}
+			free((p_mem) temp);
+			break; }
+		  case ALTERNATION : {
+			register p_link l;
+
+			l = g_getlink(p);
+			temp = setalloc();
+			setunion(temp,l->l_symbs);
+			if(!setintersect(temp,l->l_others)) {
+				/*
+				 * temp now contains the conflicting
+				 * symbols
+				 */
+				if (!(l->l_flag & (COND|PREFERING|AVOIDING))) {
+					error(p->g_lineno,
+"Alternation conflict");
+					retval = 1;
+					moreverbose(temp);
+				}
+			} else {
+				if (l->l_flag & (COND|PREFERING|AVOIDING)) {
+					l->l_flag |= NOCONF;
+					warning(p->g_lineno,
+"Conflict resolver without conflict");
+				}
+			}
+			free( (p_mem) temp);
+			if (l->l_flag & PREFERING) propagate(l->l_symbs,p+1);
+			retval |= check(l->l_rule);
+			break; }
+		}
+		p++;
+	}
+}
+
+STATIC
+moreverbose(t) register p_set t; {
+	/*
+	 * t points to a set containing conflicting symbols and pssibly
+	 * also containing nonterminals.
+	 * Take care that a printout will be prepared for these nonterminals
+	 */
+	register int i;
+	register p_nont p;
+
+	if (verbose == 2) for (i = 0, p = nonterms; i < nnonterms; i++, p++) {
+		if (NTIN(t,i)) p->n_flags |= VERBOSE;
+	}
+}
+
+STATIC
+prrule(p) register p_gram p; {
+	/*
+	 * Create a verbose printout of grammar rule p
+	 */
+	register FILE	*f;
+	int		present = 0;
+	int		firstalt = 1;
+
+	f = fout;
+	for (;;) {
+		switch (g_gettype(p)) {
+		  case EORULE :
+			fputs("\n",f);
+			return;
+		  case TERM : {
+			register p_term	q;
+			register int	c;
+
+			q = g_getterm(p);
+			if (present) prline("\n");
+			fputs("[   ",f);
+			level += 4;
+			if (q->t_flags & RESOLVER) {
+				prline("%while (..)\n");
+			}
+			if (q->t_flags & PERSISTENT) {
+				prline("%persistent\n");
+			}
+			if (r_getkind(q) != FIXED) {
+				if (!(q->t_flags & PERSISTENT)) {
+				    prline("> continue repetition on the\n");
+				}
+				printset(q->t_first, c_first);
+				if (q->t_flags & PERSISTENT) {
+				    prline("> continue repetition on the\n");
+				}
+				printset(q->t_contains, c_contains);
+				prline("> terminate repetition on the\n");
+				printset(q->t_follow,c_follow);
+				if (q->t_flags & EMPTYFIRST) {
+				    prline(">>> empty first\n");
+				}
+				if (q->t_flags & EMPTYTERM) {
+				    prline(">>> term produces empty\n");
+				}
+				cfcheck(q->t_first,q->t_follow,
+					q->t_flags & RESOLVER);
+			}
+			prrule(q->t_rule);
+			level -= 4;
+			spaces();
+			c = r_getkind(q);
+			fputs(c == STAR ? "]*" : c == PLUS ? "]+" :
+			      c == OPT ? "]?" : "]", f);
+			if (c = r_getnum(q)) {
+				fprintf(f,"%d",c);
+			}
+			prline("\n");
+			break; }
+		  case ACTION :
+			fputs("{..} ",f);
+			break;
+		  case ALTERNATION : {
+			register p_link l;
+
+			l = g_getlink(p);
+			if (firstalt) {
+				firstalt = 0;
+			}
+			else	prline("|\n");
+			printset(l->l_symbs,"> alternative on ");
+			cfcheck(l->l_symbs,
+				l->l_others,
+				(int)(l->l_flag&(COND|PREFERING|AVOIDING)));
+			fputs("    ",f);
+			level += 4;
+			if (l->l_flag & DEF) {
+				prline("%default\n");
+			}
+			if (l->l_flag & AVOIDING) {
+				prline("%avoid\n");
+			}
+			if (l->l_flag & PREFERING) {
+				prline("%prefer\n");
+			}
+			if (l->l_flag & COND) {
+				prline("%if ( ... )\n");
+			}
+			prrule(l->l_rule);
+			level -= 4;
+			if (g_gettype(p+1) == EORULE) {
+				return;
+			}
+			spaces();
+			p++; continue; }
+		  case LITERAL :
+		  case TERMINAL : {
+			register p_token pt = &tokens[g_getcont(p)];
+
+			fprintf(f,pt->t_tokno<0400 ?
+				  "'%s' " : "%s ", pt->t_string);
+			break; }
+		  case NONTERM :
+			fprintf(f,"%s ",nonterms[g_getcont(p)].n_name);
+			break;
+		}
+		p++;
+		present = 1;
+	}
+}
+
+STATIC
+cfcheck(s1,s2,flag) p_set s1,s2; {
+	/*
+	 * Check if s1 and s2 have elements in common.
+	 * If so, flag must be non-zero, indicating that there is a
+	 * conflict resolver, otherwise, flag must be zero, indicating
+	 * that there is not.
+	 */
+	register p_set temp;
+
+	temp = setalloc();
+	setunion(temp,s1);
+	if (!setintersect(temp,s2)) {
+		if (! flag) {
+			printset(temp,">>> conflict on ");
+			prline("\n");
+		}
+	} else {
+		if (flag) {
+			prline(">>> %if/%while, no conflict\n");
+		}
+	}
+	free((p_mem) temp);
+}
+
+STATIC
+resolve(p) register p_gram p; {
+	/*
+	 * resolve conflicts, as specified by the user
+	 */
+	for (;;) {
+		switch (g_gettype(p)) {
+		  case EORULE :
+			return;
+		  case TERM :
+			resolve(g_getterm(p)->t_rule);
+			break;
+		  case ALTERNATION : {
+			register p_link	l;
+
+			l = g_getlink(p);
+			if (l->l_flag & AVOIDING) {
+				/*
+				 * On conflicting symbols, this rule
+				 * is never chosen
+				 */
+				setminus(l->l_symbs,l->l_others);
+			}
+			if (setempty(l->l_symbs)) {
+				/*
+				 * This may be caused by the statement above
+				 */
+				error(p->g_lineno,"Alternative never chosen");
+			}
+			resolve(l->l_rule);
+			break; }
+		}
+		p++;
+	}
+}
+
+STATIC
+propagate(set,p) p_set set; register p_gram p; {
+	/*
+	 * Propagate the fact that on the elements of set the grammar rule
+	 * p will not be chosen.
+	 */
+	while (g_gettype(p) != EORULE) {
+		setminus(g_getlink(p)->l_symbs,set);
+		p++;
+	}
+}
+
+STATIC
+spaces() {
+
+	if (level > 0) fprintf(fout,"%*c",level,' ');
+}

util/LLgen/src/extern.h

@@ -0,0 +1,94 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * $Id$
+ * Miscellanious constants and
+ * some variables that are visible in more than one file
+ */
+
+# define LTEXTSZ	256	/* Size of longest token */
+
+/*
+ * options for the identifier search routine
+ */
+
+# define ENTERING	1
+# define BOTH		2
+
+/*
+ * Now for some	declarations
+ */
+
+extern char	ltext[];	/* input buffer	*/
+extern int	nnonterms;	/* number of nonterminals */
+extern int	ntokens;	/* number of terminals */
+extern int	nterms;		/* number of terms */
+extern int	nalts;		/* number of alternatives */
+extern p_start	start;		/* will	contain	startsymbols */
+#ifdef NON_CORRECTING
+extern int	nsubstarts;	/* number of subparserstarts */
+extern p_set	start_firsts;	/* Will contain the union of first sets of
+				   startsymbols when -n -s option is on */
+#endif
+extern int	linecount;	/* line number */
+extern int	assval;		/* to create difference	between	literals
+				 * and other terminals
+				 */
+extern p_nont	nonterms;	/* the nonterminal array */
+extern p_nont	maxnt;		/* is filled up until here */
+extern p_token	tokens;		/* the token array */
+extern p_token	maxt;		/* is filled up until here */
+extern int	norder, torder;
+				/* order of nonterminals in the grammar,
+				 * important because actions are copied to
+				 * a temporary file in the order in which they
+				 * were read
+				 */
+extern string	e_noopen;	/* Error message string used often */
+extern int	verbose;	/* Level of verbosity */
+extern int	wflag;		/* warnings? */
+extern string	lexical;	/* name of lexical analyser */
+extern string	prefix;		/* prefix of externals */
+extern string	onerror;	/* name of user error handler */
+extern int	ntneeded;	/* ntneeded = 1 if nonterminals are included
+				 * in the sets.
+				 */
+extern int	ntprint;	/* ntprint = 1 if they must be printed too in
+				 * the LL.output file (-x option)
+				 */
+# ifndef NDEBUG
+extern int	debug;
+# endif /* not NDEBUG */
+extern p_file	files,pfile;	/* pointers to file structure.
+				 * "files" points to the start of the
+				 * list */
+extern p_file	maxfiles;
+extern string	LLgenid;	/* LLgen identification string */
+extern t_token	lextoken;	/* the current token */
+extern int	nerrors;
+extern string	rec_file, incl_file;
+#ifdef NON_CORRECTING
+extern string	nc_rec_file, nc_incl_file;
+#endif
+extern int	low_percentage, high_percentage;
+extern int	min_cases_for_jmptable;
+extern int	jmptable_option;
+extern int	ansi_c;
+#ifdef NON_CORRECTING
+extern int	non_corr;
+extern int	subpars_sim;
+extern p_gram	illegal_gram;
+#endif
+extern int	strip_grammar;
+extern int	in_production;

util/LLgen/src/global.c

@@ -0,0 +1,89 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * global.c
+ * Contains declarations visible in several other source files
+ */
+
+# include "types.h"
+# include "extern.h"
+# include "io.h"
+
+# ifndef NORCSID
+static string rcsid4 = "$Id$";
+# endif
+
+char	ltext[LTEXTSZ];
+p_nont	nonterms;
+p_nont	maxnt;
+int	nnonterms;
+p_token	tokens;
+p_token	maxt;
+int	ntokens;
+int	nterms, nalts;
+int	norder, torder;
+#ifdef NON_CORRECTING
+int	nsubstarts;
+p_set	start_firsts;
+#endif
+p_start	start;
+int	linecount;
+int	assval;
+FILE	*fout;
+FILE	*fpars;
+FILE	*finput;
+FILE	*fact;
+char	f_pars[] = PARSERFILE;
+char	f_temp[] = ACTFILE;
+#ifdef NON_CORRECTING
+char	f_nc[20];
+#endif
+char	f_out[20];
+string	f_input;
+char	f_include[20];
+char	f_rec[20];
+string	e_noopen = "Cannot open %s";
+int	verbose;
+int	wflag;
+string	lexical;
+string	prefix;
+string	onerror;
+int	ntneeded;
+int	ntprint;
+# ifndef NDEBUG
+int	debug;
+#endif /* not NDEBUG */
+p_file	files;
+p_file	maxfiles;
+p_file	pfile;
+string	LLgenid = "/* LLgen generated code from source %s */\n";
+t_token lextoken;
+int	nerrors;
+string	rec_file, incl_file;
+#ifdef NON_CORRECTING
+string	nc_rec_file, nc_incl_file;
+#endif
+int	low_percentage = 10, high_percentage = 30;
+int	min_cases_for_jmptable = 8;
+int	jmptable_option;
+int	ansi_c = 0;
+#ifdef NON_CORRECTING
+int	non_corr = 0;
+int	subpars_sim = 0;
+p_gram	illegal_gram;
+#endif
+int	strip_grammar = 0;
+int	in_production;		/* set when the parser is reading a production
+				   rule.
+				*/

util/LLgen/src/io.h

@@ -0,0 +1,43 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * $Id$
+ * Some important file names and variables
+ */
+
+# include <stdio.h>
+
+/* FILES */
+
+# define OUTFILE	"%s.output"	/* -v option */
+# define PARSERFILE	"xxxXXXXXX"	/* This is what we want */
+# define ACTFILE	"tempXXXXXX"	/* temporary file to save actions */
+# define HFILE		"%spars.h"	/* file for "#define's " */
+# define RFILE		"%spars.c"	/* Error recovery */
+#ifdef NON_CORRECTING
+# define NCFILE		"%sncor.c"	/* Non-corrcting error recovery */
+#endif
+extern FILE *finput;
+extern FILE *fpars;
+extern FILE *fact;
+extern FILE *fout;
+extern char	f_pars[];
+extern char	f_temp[];
+extern char	f_out[];
+extern string	f_input;
+extern char	f_include[];
+extern char	f_rec[];
+#ifdef NON_CORRECTING
+extern char	f_nc[];
+#endif

util/LLgen/src/machdep.c

@@ -0,0 +1,72 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * machdep.c
+ * Machine dependant things
+ */
+
+#include <stdlib.h>
+#include <string.h>
+# include "types.h"
+
+# ifndef NORCSID
+static string rcsid5 = "$Id$";
+# endif
+
+/* In this file the following routines are defined: */
+extern	UNLINK();
+extern	RENAME();
+extern string	libpath();
+
+UNLINK(x) string x; {
+	/* Must remove the file "x" */
+
+#ifdef USE_SYS
+	sys_remove(x);	/* systemcall to remove file */
+#else
+	unlink(x);
+#endif
+}
+
+RENAME(x,y) string x,y; {
+	/* Must move the file "x" to the file "y" */
+
+#ifdef USE_SYS
+	if(! sys_rename(x,y)) fatal(1,"Cannot rename to %s",y);
+#else
+	unlink(y);
+	if (link(x,y) != 0) fatal(1,"Cannot rename to %s",y);
+	unlink(x);
+#endif
+}
+
+string
+libpath(s) string s; {
+	/* Must deliver a full pathname to the library file "s" */
+
+	register string p;
+	register length;
+	p_mem alloc();
+	string strcpy(), strcat();
+
+	char* libdir = getenv("LLGEN_LIB_DIR");
+	if (!libdir)
+		libdir = LIBDIR;
+	length = strlen(libdir) + strlen(s) + 2;
+	p = (string) alloc((unsigned) length);
+	strcpy(p,libdir);
+	strcat(p,"/");
+	strcat(p,s);
+	return p;
+}

util/LLgen/src/main.c

@@ -0,0 +1,378 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * main.c
+ * Contains main program, and some error message routines
+ */
+
+#include <stdlib.h>
+#include <string.h>
+# include "types.h"
+# include "io.h"
+# include "extern.h"
+# include "sets.h"
+# include "assert.h"
+
+# ifndef NORCSID
+static string rcsid6 = "$Id$";
+# endif
+
+/* In this file the following routines are defined: */
+extern int	main();
+STATIC		readgrammar();
+STATIC		doparse();
+extern		error();
+extern		fatal();
+extern		comfatal();
+extern		copyfile();
+extern		install();
+extern char	*mktemp();
+extern char	*sbrk();
+
+main(argc,argv) register string	argv[]; {
+	register string arg;
+	string libpath();
+	char	*beg_sbrk = 0;
+
+	/* Initialize */
+
+	assval = 0400;
+	/* read options */
+
+	while (argc >= 2 && (arg = argv[1], *arg == '-')) {
+		while (*++arg) {
+			switch(*arg) {
+			  case 'j':
+			  case 'J':
+				jmptable_option = 1;
+				if (*++arg)
+					min_cases_for_jmptable = atoi(arg);
+				break;
+			  case 'w':
+			  case 'W':
+				wflag = 1;
+				continue;
+			  case 'v':
+			  case 'V':
+				verbose++;
+				continue;
+			  case 'l':
+			  case 'L':
+				low_percentage = atoi(++arg);
+				break;
+			  case 'h':
+			  case 'H':
+				high_percentage = atoi(++arg);
+				break;
+# ifndef NDEBUG
+			  case 'd':
+			  case 'D':
+				debug++;
+				continue;
+			  case 'r':
+			  case 'R':
+				if (rec_file) {
+					fprintf(stderr,"duplicate -r flag\n");
+					exit(1);
+				}
+				rec_file = ++arg;
+				break;
+			  case 'i':
+			  case 'I':
+				if (incl_file) {
+					fprintf(stderr,"duplicate -i flag\n");
+					exit(1);
+				}
+				incl_file = ++arg;
+				break;
+#endif /* not NDEBUG */
+			  case 'x':
+			  case 'X':
+				ntneeded = 1;
+				ntprint = 1;
+				continue;
+			  case 'a':
+			  case 'A':
+				ansi_c = 1;
+				continue;
+#ifdef NON_CORRECTING
+			  case 'n':
+			  case 'N':
+				non_corr = 1;
+				continue;
+			  case 's':
+			  case 'S':
+				subpars_sim = 1;
+				continue;
+#endif
+			  case 'g':
+			  case 'G':
+				strip_grammar = 1;
+				continue;
+			  default:
+				fprintf(stderr,"illegal option : %c\n",*arg);
+				exit(1);
+			}
+			break;
+		}
+		argv++;
+		argc--;
+	}
+
+	if (verbose) beg_sbrk = sbrk(0);
+
+#ifdef NON_CORRECTING
+	if ((subpars_sim) && (!non_corr)) {
+	    fprintf(stderr,"option -s illegal without -n, turned off\n");
+	    subpars_sim = 0;
+	}
+#endif
+
+	/*
+	 * Now check wether the sets should include nonterminals
+	 */
+	if (verbose == 2) ntneeded = 1;
+	/*
+	 * Initialise
+	 */
+# ifndef NDEBUG
+	if (!rec_file) {
+# endif
+		rec_file = libpath("rec");
+# ifndef NDEBUG
+	}
+	if (!incl_file) {
+# endif
+		incl_file = libpath("incl");
+# ifndef NDEBUG
+	}
+# endif
+#ifdef NON_CORRECTING
+	if (non_corr) {
+	    nc_incl_file = libpath("nc_incl");
+	    nc_rec_file = libpath ("nc_rec");
+	}
+#endif
+	mktemp(f_temp);
+	mktemp(f_pars);
+	if ((fact = fopen(f_temp,"w")) == NULL) {
+		fputs("Cannot create temporary\n",stderr);
+		exit(1);
+	}
+	name_init();
+	readgrammar(argc,argv);
+	sprintf(f_out, OUTFILE, prefix ? prefix : "LL");
+
+	/* for the following two filenames only one L is used; historical
+	   reasons ...
+	*/
+	sprintf(f_include, HFILE, prefix ? prefix : "L");
+	sprintf(f_rec, RFILE, prefix ? prefix : "L");
+#ifdef NON_CORRECTING
+	if (non_corr)
+	    sprintf(f_nc, NCFILE, prefix ? prefix : "L");
+#endif
+	setinit(ntneeded);
+	maxnt = &nonterms[nnonterms];
+	maxt = &tokens[ntokens];
+	/*
+	 * Now, the grammar is read. Do some computations
+	 */
+	co_reach();		/* Check for undefined and unreachable */
+	if (nerrors) comfatal();
+	do_compute();
+	conflchecks();
+	if (nerrors) comfatal();
+	fclose(fact);
+	if (argc-- == 1) {
+		fputs("No code generation for input from standard input\n",
+		      stderr);
+	}
+	else	gencode(argc);
+	UNLINK(f_temp);
+	UNLINK(f_pars);
+	if (verbose) {
+		fprintf(stderr, "number of nonterminals: %d\n", nnonterms);
+		fprintf(stderr, "number of tokens: %d\n", ntokens);
+		fprintf(stderr, "number of term structures: %d\n", nterms);
+		fprintf(stderr, "number of alternation structures: %d\n", nalts);
+		fprintf(stderr, "total memory used: %ld\n", (long)(sbrk(0) - beg_sbrk));
+	}
+	exit(0);
+}
+
+STATIC
+readgrammar(argc,argv) char *argv[]; {
+	/*
+	 * Do just what the name suggests : read the grammar
+	 */
+	register p_file	p;
+	p_mem		alloc();
+
+	linecount = 0;
+	f_input = "no filename";
+	/*
+	 * Build the file structure
+	 */
+	files = p = (p_file) alloc((unsigned) (argc+1) * sizeof(t_file));
+	if (argc-- == 1) {
+		finput = stdin;
+		f_input = "standard input";
+		doparse(p++);
+	} else {
+		while (argc--) {
+			if ((finput = fopen(f_input=argv[1],"r")) == NULL) {
+				fatal(0,e_noopen,f_input);
+			}
+			doparse(p++);
+			argv++;
+			fclose(finput);
+		}
+	}
+	maxfiles = p;
+	if (! lexical) lexical = "yylex";
+	/*
+	 * There must be a start symbol!
+	 */
+	if (! nerrors && start == 0) {
+		fatal(linecount,"Missing %%start");
+	}
+	if (nerrors) comfatal();
+}
+
+STATIC
+doparse(p) register p_file p; {
+
+	linecount = 0;
+	p->f_name = f_input;
+	p->f_firsts = 0;
+	pfile = p;
+	torder = -1;
+	norder = -1;
+	LLparse();
+	p->f_nonterminals = norder;
+	p->f_terminals = torder;
+}
+
+/* VARARGS1 */
+error(lineno,s,t,u) string	s,t,u; {
+	/*
+	 * Just an error message
+	 */
+
+	++nerrors;
+	if (!lineno) lineno = 1;
+	fprintf(stderr,"\"%s\", line %d: ",f_input, lineno);
+	fprintf(stderr,s,t,u);
+	fputs("\n",stderr);
+}
+
+/* VARARGS1 */
+warning(lineno,s,t,u) string	s,t,u; {
+	/*
+	 * Just a warning
+	 */
+
+	if (wflag) return;
+	if (!lineno) lineno = 1;
+	fprintf(stderr,"\"%s\", line %d: (Warning) ",f_input, lineno);
+	fprintf(stderr,s,t,u);
+	fputs("\n",stderr);
+}
+
+/* VARARGS1 */
+fatal(lineno,s,t,u) string	s,t,u; {
+	/*
+	 * Fatal error
+	 */
+	error(lineno,s,t,u);
+	comfatal();
+}
+
+comfatal() {
+	/*
+	 * Some common code for exit on errors
+	 */
+	if (fact != NULL) {
+		fclose(fact);
+		UNLINK(f_temp);
+	}
+	if (fpars != NULL) fclose(fpars);
+	UNLINK(f_pars);
+	exit(1);
+}
+
+copyfile(file) string file; {
+	/*
+	 * Copies a file indicated by the parameter to filedescriptor fpars.
+	 */
+	register int	c;
+	register FILE	*f;
+
+	if ((f = fopen(file,"r")) == NULL) {
+		fatal(0,"Cannot open library file %s, call an expert",file);
+	}
+	while ((c = getc(f)) != EOF) putc(c,fpars);
+	fclose(f);
+}
+
+install(target, source) string target, source; {
+	/*
+	 * Copy the temporary file generated from source to target
+	 * if allowed (which means that the target must be generated
+	 * by LLgen from the source, or that the target is not present
+	 */
+	register int	c1, c2;
+	register FILE	*f1, *f2;
+	int		cnt;
+
+	/*
+	 * First open temporary, generated for source
+	 */
+	if ((f1 = fopen(f_pars,"r")) == NULL) {
+		fatal(0,e_noopen,f_pars);
+	}
+	/*
+	 * Now open target for reading
+	 */
+	if ((f2 = fopen(target,"r")) == NULL) {
+		fclose(f1);
+		RENAME(f_pars, target);
+		return;
+	}
+	/*
+	 * Compute length of LLgen identification string. The target must
+	 * start with that!
+	 */
+	cnt = strlen(LLgenid) + strlen(source) - 2;
+	/*
+	 * Now compare the target with the temporary
+	 */
+	do {
+		c1 = getc(f1);
+		c2 = getc(f2);
+		if (cnt >= 0) cnt--;
+	} while (c1 == c2 && c1 != EOF);
+	fclose(f1);
+	fclose(f2);
+	/*
+	 * Here, if c1 != c2 the target must be recreated
+	 */
+	if (c1 != c2) {
+		if (cnt >= 0) {
+			fatal(0,"%s : not a file generated by LLgen",target);
+		}
+		RENAME(f_pars,target);
+	}
+}

util/LLgen/src/name.c

@@ -0,0 +1,251 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * name.c
+ * Defines the symboltable search routine, a name store routine and an
+ * initialising routine.
+ */
+
+# include "types.h"
+# include "extern.h"
+# include "assert.h"
+# include "io.h"
+
+# ifndef NORCSID
+static string rcsid7 = "$Id$";
+# endif
+
+# define HASHSIZE 128
+# define NMSIZ 2048	/* Room for names allocated NMSIZ bytes at a time */
+
+static char	*name, *maxname;
+static p_entry	h_root[HASHSIZE];		/* hash table */
+static string	e_literal = "Illegal literal";
+static p_entry	entries, maxentries;
+static t_info	token_info, nont_info;
+
+/* Defined in this file are: */
+extern string	store();
+extern		name_init();
+STATIC int	hash();
+STATIC p_entry	newentry();
+extern p_gram	search();
+
+p_mem alloc();
+p_mem new_mem();
+
+name_init() {
+	token_info.i_esize = sizeof (t_token);
+	token_info.i_incr = 50;
+	nont_info.i_esize = sizeof (t_nont);
+	nont_info.i_incr = 50;
+	search(TERMINAL,"EOFILE",ENTERING);
+#ifdef NON_CORRECTING
+	illegal_gram = search(TERMINAL,"LLILLEGAL",ENTERING);
+#endif
+}
+
+STATIC p_entry
+newentry(str, next) string str; p_entry next; {
+	register p_entry p;
+
+	if ((p = entries) == maxentries) {
+		p = (p_entry) alloc(50 * sizeof(t_entry));
+		maxentries = p + 50;
+	}
+	entries = p + 1;
+	p->h_name = str;
+	p->h_next = next;
+	p->h_type.g_lineno = linecount;
+#ifdef NON_CORRECTING
+	p->h_type.g_erroneous = 0;
+#endif
+	return p;
+}
+
+string
+store(s) string s; {
+	/*
+	 * Store a string s in the name table
+	 */
+	register string	s1, t ,u;
+
+	u = name;
+	t = s;
+	s1 = u;
+	do {
+		if (u >= maxname) {
+			u = alloc(NMSIZ);
+			maxname = u + NMSIZ;
+			t = s;
+			s1 = u;
+		}
+		*u++ = *t;
+	} while (*t++);
+	name = u;
+	return s1;
+}
+
+STATIC int
+hash(str) string str; {
+	/*
+	 * Compute the hash for string str
+	 */
+	register int	i;
+	register string l;
+
+	l = str;
+	i = 0;
+	while (*l != '\0') i += *l++ & 0377;
+	i += l - str;
+	return i % HASHSIZE;
+}
+
+p_gram
+search(type,str,option) register string str; {
+	/*
+	 * Search for object str.
+	 * It has type UNKNOWN, LITERAL, TERMINAL or NONTERM.
+	 * option can be ENTERING or BOTH (also looking).
+	 */
+	register int		val = 0;
+	register p_entry	p;
+	register int		i;
+	int			type1;
+
+	i = hash(str);
+	/*
+	 * Walk hash chain
+	 */
+	for (p = h_root[i]; p != (p_entry) 0; p = p->h_next) {
+		if(!strcmp(p->h_name,str)) {
+			type1 = g_gettype(&(p->h_type));
+			if (type1 != type) {
+				if (type1 == LITERAL || type == LITERAL) {
+					continue;
+				}
+				if (type == TERMINAL) {
+					error(linecount,
+						"%s: is already a nonterminal",
+						str);
+					continue;
+				}
+				else if (type == NONTERM) {
+					error(linecount,
+						"%s : is already a token",
+						str);
+					continue;
+				}
+			}
+			if (option==ENTERING)  {
+				error(linecount,
+					"%s : is already defined",str);
+			}
+			p->h_type.g_lineno = linecount;
+			return &(p->h_type);
+		}
+	}
+	p = newentry(store(str), h_root[i]);
+	h_root[i] = p;
+	if (type == TERMINAL || type == LITERAL) {
+		register p_token pt;
+
+		pt = (p_token) new_mem(&token_info);
+		tokens = (p_token) token_info.i_ptr;
+		pt->t_string = p->h_name;
+		if (type == LITERAL) {
+			if (str[0] == '\\') {
+				/*
+				 * Handle escapes in literals
+				 */
+				if (str[2] == '\0') {
+					switch(str[1]) {
+					  case 'n' :
+						val = '\n';
+						break;
+					  case 'r' :
+						val = '\r';
+						break;
+					  case 'b' :
+						val = '\b';
+						break;
+					  case 'f' :
+						val = '\f';
+						break;
+					  case 't' :
+						val = '\t';
+						break;
+					  case '\'':
+						val = '\'';
+						break;
+					  case '\\':
+						val = '\\';
+						break;
+					  default  :
+						error(linecount,e_literal);
+					}
+				} else {
+					/*
+					 * Here, str[2] != '\0'
+					 */
+					if (str[1] > '3' || str[1] < '0' ||
+					    str[2] > '7' || str[2] < '0' ||
+					    str[3] > '7' || str[3] < '0' ||
+					    str[4] != '\0') error(linecount,e_literal);
+					val = 64*str[1] - 73*'0' +
+					      8*str[2] + str[3];
+				}
+			} else {
+				/*
+				 * No escape in literal
+				 */
+				if (str[1] == '\0') val = str[0];
+				else error(linecount,e_literal);
+			}
+			pt->t_tokno = val;
+			g_settype(&(p->h_type), LITERAL);
+		} else {
+			/*
+			 * Here, type = TERMINAL
+			 */
+			pt->t_tokno = assval++;
+			g_settype(&(p->h_type), TERMINAL);
+		}
+		g_setcont(&(p->h_type), ntokens);
+		ntokens++;
+		return &(p->h_type);
+	}
+	/*
+	 * type == NONTERM || type == UNKNOWN
+	 * UNKNOWN and not yet declared means : NONTERM
+	 */
+	{
+		register p_nont q;
+
+		q = (p_nont) new_mem(&nont_info);
+		nonterms = (p_nont) nont_info.i_ptr;
+		q->n_name = p->h_name;
+		q->n_rule = 0;
+		q->n_lineno = linecount;
+		q->n_string = f_input;
+		q->n_follow = 0;
+		q->n_flags = 0;
+		q->n_contains = 0;
+		g_settype(&(p->h_type), NONTERM);
+		g_setcont(&(p->h_type), nnonterms);
+		g_setnpar(&(p->h_type), 0);
+		nnonterms++;
+		return &(p->h_type);
+	}
+}

util/LLgen/src/reach.c

@@ -0,0 +1,135 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * reach.c
+ * Determine which nonterminals are reachable, and also check that they
+ * are all defined.
+ */
+
+# include "types.h"
+# include "extern.h"
+# include "io.h"
+# include "assert.h"
+
+# ifndef NORCSID
+static string rcsid8 = "$Id$";
+# endif
+
+/* In this file the following routines are defined: */
+extern co_reach();
+STATIC reachable();
+STATIC reachwalk();
+
+co_reach() {
+	/*
+	 * Check for undefined or unreachable nonterminals.
+	 */
+	register p_nont		p;
+	register p_token	t;
+	register p_start	st;
+	register p_file		x;
+	register int		s;
+
+	/* Check for undefined nonterminals */
+	for (p = nonterms; p < maxnt; p++) {
+		if (! p->n_rule) {	/* undefined */
+			f_input = p->n_string;
+			error(p->n_lineno,"Nonterminal %s not defined",
+				p->n_name);
+		}
+	}
+
+	/*
+	 * Walk the grammar rules, starting with the startsymbols
+	 * Mark the nonterminals that are encountered with the flag
+	 * REACHABLE, and walk their rules, if not done before
+	 */
+	for (st = start; st; st = st->ff_next) {
+		reachable(&nonterms[st->ff_nont]);
+	}
+	/*
+	 * Now check for unreachable nonterminals
+	 */
+	for (x = files; x < maxfiles; x++) {
+	    f_input = x->f_name;
+	    for (s = x->f_nonterminals; s != -1; s = p->n_next) {
+		p = &nonterms[s];
+		if (! (p->n_flags & REACHABLE)) {
+			warning(p->n_lineno,"nonterminal %s unreachable",
+				p->n_name);
+		}
+	    }
+	    for (s = x->f_terminals; s != -1; s = t->t_next) {
+		t = &tokens[s];
+		if (! (t->t_flags & REACHABLE)) {
+			warning(t->t_lineno,"terminal %s not used",
+				t->t_string);
+		}
+	    }
+	}
+}
+
+STATIC
+reachable(p) register p_nont p; {
+	/*
+	 * Enter the fact that p is reachable, and look for implications
+	 */
+	if (! (p->n_flags & REACHABLE)) {
+		p->n_flags |= REACHABLE;
+		/*
+		 * Now walk its grammar rule
+		 */
+		if (p->n_rule) reachwalk(p->n_rule);
+	}
+}
+
+STATIC
+reachwalk(p) register p_gram p; {
+	/*
+	 * Walk through rule p, looking for nonterminals.
+	 * The nonterminals found are entered as reachable
+	 */
+
+	for (;;) {
+		switch(g_gettype(p)) {
+		  case ALTERNATION :
+			reachwalk(g_getlink(p)->l_rule);
+			break;
+		  case TERM :
+			reachwalk(g_getterm(p)->t_rule);
+			break;
+		  case NONTERM : {
+			register p_nont n = &nonterms[g_getcont(p)];
+
+			reachable(n);
+			if (n->n_rule && g_gettype(n->n_rule) == EORULE &&
+			    ! g_getnpar(p) && (getntparams(n) == 0)) {
+				register p_gram np = p;
+				do {
+					*np = *(np + 1);
+					np++;
+				} while (g_gettype(np) != EORULE);
+				continue;
+			}
+			break;
+		  }
+		  case TERMINAL:
+			tokens[g_getcont(p)].t_flags |= REACHABLE;
+			break;
+		  case EORULE :
+			return;
+		}
+		p++;
+	}
+}

util/LLgen/src/savegram.c

@@ -0,0 +1,385 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * All rights reserved.
+ */
+
+#ifdef NON_CORRECTING
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * savegram.c
+ * Save the input grammar for non-correcting error recovery
+ *
+ * Grammar rules are `flattened' by introducing anonymous nonterminals.
+ * [B]? becomes X; X: B | {empty}
+ * [B]+ becomes X: B Y; Y: X | {empty}
+ * [B]* becomes X; X: B X | {empty}
+ * [B | C] becomes X; X: B | C
+ * [B | C]* becomes X; X: B X | C X | {empty}  etc.
+ */
+
+
+# include "types.h"
+# include "extern.h"
+# include "io.h"
+# include "assert.h"
+# include "sets.h"
+
+#define LLALT 9999
+
+static int nt_highest;
+extern int nbytes;
+extern p_mem alloc();
+extern p_set start_firsts;
+extern p_set setalloc();
+extern p_gram search();
+
+STATIC save_rule();
+STATIC save_set();
+
+/* t_list will contain terms to be `flattened' */
+static struct t_list {
+	p_term term;
+	int t_nt_num;
+} *t_list;
+
+/* Subparse list will contain symbols in %substart */
+static struct subparse_list {
+	p_gram sub_action;
+	int sub_nt_num;
+} *sub_list;
+
+/* Index in t_list */
+static int t_list_index;
+
+/* Index in subparse_list */
+static int sub_list_index;
+
+/* File to save grammar to */
+static FILE *fgram;
+
+/* Nonterminal number to simulate parsers that get called in actions
+   used when LLgen called with -n -s options */
+int act_nt;
+
+save_grammar(f) FILE *f; {
+	/*
+	 * Save the grammar
+	 */
+	register p_nont		p;
+	register p_start	st;
+	register int		nt_nr;
+
+	fgram = f;
+
+	/* Compute highest nonterminal nr. */
+	nt_highest = nnonterms + assval - 1;
+
+
+	/* Generate some constants in the grammar file */
+
+	/* Allocate terms list */
+	t_list = (struct t_list *) alloc((unsigned) nterms * sizeof(struct t_list));
+	t_list_index = 0;
+
+	sub_list = (struct subparse_list *) alloc(nsubstarts * sizeof(struct subparse_list));
+
+	fputs("static ", fgram);
+	fputs((prefix ? prefix : "LL"), fgram);
+	fputs("grammar[] = {\n", fgram);
+
+	/* Check if -n -s option is on */
+	if (subpars_sim) {
+
+		/* Allocate action simulation nt */
+
+		act_nt = ++nt_highest;
+
+		/* write simualtion rule */
+		fprintf(fgram, "/* Simulation rule */\n");
+		fprintf(fgram, "%d,\n", act_nt);
+
+		/* Put a firstset and a fake followset */
+		/* Followset optimization is not implemented for
+		   -s because it would be hard, and does not
+		   bring enough improvement to jutify the effort
+		 */
+		save_set(start_firsts);
+		save_set(start_firsts);
+		/* Simulation rule procudes empty */
+		fprintf(fgram, "%d,\n", 1);
+		for (st = start; st; st = st->ff_next)
+		{
+			fprintf(fgram, "%d, %d, %d, \n", st->ff_nont + assval,
+								act_nt, LLALT);
+		}
+		fprintf(fgram, "%d, \n", 0);
+
+	}
+
+	/* Now process all rules */
+	for (p = nonterms, nt_nr = assval; p < maxnt; p++, nt_nr++) {
+		fprintf(fgram, "/* nr. %d %s */\n", nt_nr, p->n_name);
+		fprintf(fgram, "%d, ",nt_nr);
+		if (! p->n_rule) {	/* undefined */
+			f_input = p->n_string;
+			error(p->n_lineno,"Nonterminal %s not defined",
+				p->n_name);
+		}
+
+		/* Save the first_set and follow set */
+		save_set(p->n_nc_first);
+		save_set(p->n_nc_follow);
+
+		if (p->n_flags & EMPTY)
+			fprintf(fgram, "%d,\n", 1);
+		else
+			fprintf(fgram, "%d,\n", 0);
+
+		save_rule(p->n_rule, 0);
+
+		fprintf(fgram, "%d,\n", 0);
+	}
+
+	/* Resolve terms, they are on t_list */
+
+	fprintf(fgram, "/* Fresh nonterminals */\n");
+
+	{ int i;
+	for (i = 0; i < t_list_index; i++)
+	{
+
+	/* Terms of the form [] without + ? * or number produce
+	 a NIL pointer in the term-list */
+		if ((t_list + i)->term == (struct term *) 0) {
+			continue;
+		}
+
+		fprintf(fgram, "%d, ", (t_list + i)->t_nt_num);
+
+		/* Save the first and follow sets */
+
+		save_set((t_list + i)->term->t_nc_first);
+		save_set((t_list + i)->term->t_nc_follow);
+
+		/* NOTE: A VARIABLE REPETITION COUNT TERMS RULE IS NOT
+		   ALLOWED TO PRODUCE EMPTY IN LLGEN
+		 */
+
+		switch(r_getkind((t_list + i)->term)) {
+		case FIXED:
+		/* Already done by repeating new nonterminal */
+
+			/* FIXED term-rule may produce empty */
+			if (empty((t_list +i)->term->t_rule))
+				fprintf(fgram, "%d,\n", 1);
+			else
+				fprintf(fgram, "%d,\n", 0);
+
+			save_rule((t_list + i)->term->t_rule, 0);
+			fprintf(fgram, "%d,\n", 0);
+			break;
+		case STAR:
+		/* Save the rule, appending the new lhs for this rule */
+
+			/* Star rules always produce empty */
+			fprintf(fgram, "1,\n");
+
+			save_rule((t_list + i)->term->t_rule,
+						(t_list + i)->t_nt_num);
+			fprintf(fgram, "%d,\n%d,\n", LLALT, 0);
+			/* ALT EMPTY*/
+			break;
+		case PLUS:
+		/* Save the rule appending a fresh nonterminal */
+
+			fprintf(fgram, "%d,\n", 0);
+
+			save_rule((t_list + i)->term->t_rule, ++nt_highest);
+			fprintf(fgram, "%d,\n", 0); /* EOR */
+			fprintf(fgram, "%d, ", nt_highest);
+			/* First set of the extra nonterm is same as
+			   for the term */
+			/* Except that the new nonterm also produces empty ! */
+			save_set((t_list + i)->term->t_nc_first);
+			save_set((t_list + i)->term->t_nc_follow);
+			fprintf(fgram, "1,\n");
+			fprintf(fgram, "%d, ", (t_list+i)->t_nt_num);
+			fprintf(fgram, "%d,\n%d,\n", LLALT, 0); /* ALT EMPTY */
+			break;
+		case OPT:
+			fprintf(fgram, "1,\n");
+			save_rule((t_list + i)->term->t_rule, 0);
+			fprintf(fgram, "%d,\n%d,\n", LLALT, 0); /* ALT EMPTY */
+			break;
+		}
+	}
+	}
+
+	/* Resolve %substarts */
+	if (!subpars_sim) {
+		int i,s,check;
+		p_start ff, gg;
+		p_set temp_set;
+
+		for (i = 0; i < sub_list_index; i++) {
+			fprintf(fgram, "%d, ", (sub_list + i)->sub_nt_num);
+			/* Compute the first set */
+			temp_set = setalloc();
+			for (ff = g_getsubparse((sub_list + i)->sub_action);
+							ff; ff = ff->ff_next){
+			s = setunion(temp_set,
+					(&nonterms[ff->ff_nont])->n_first);
+			check = 0;
+			for (gg =start; gg; gg = gg->ff_next)
+				if (ff->ff_nont == gg->ff_nont)
+					check = 1;
+			if (check == 0)
+			warning((sub_list + i)->sub_action->g_lineno,
+					"\"%s\" is not a startsymbol",
+					(&nonterms[ff->ff_nont])->n_name);
+		}
+		save_set(temp_set);
+		save_set(temp_set);
+		free(temp_set);
+
+		/* Produces empty */
+		fprintf(fgram, "1,\n");
+
+		ff = g_getsubparse((sub_list + i)->sub_action);
+
+		for (; ff; ff = ff->ff_next)
+			fprintf(fgram, "%d, %d, %d, \n", ff->ff_nont + assval,
+						(sub_list + i)->sub_nt_num,
+									LLALT);
+		fprintf(fgram, "%d, \n", 0);
+		}
+	}
+
+	fprintf(fgram, "%d\n};\n", 0);
+	fprintf(fgram, "#define LLNNONTERMINALS %d\n", nt_highest - assval + 1);
+}
+
+STATIC
+save_rule(p, tail) register p_gram p; int tail; {
+/*
+ Walk through rule p, saving it. The non-terminal tail is
+ appended to the rule. It needs to be appended in this function
+ to process alt-rules correctly. Tail == 0 means don't append.
+ */
+
+	int in_alt;
+	int illegal_num;
+	/* Processing an alt needs some special care. When processing the
+	   first alternative, we don't want to write the alt-code;
+	   When appending something to the alt, it needs to be appended to
+	   every alternative and not at the end of the rule.
+	*/
+
+	/* Look up the ILLEGAL token number */
+	illegal_num = tokens[g_getcont(illegal_gram)].t_tokno;
+
+	in_alt = 0;
+	for (;;) {
+		switch(g_gettype(p)) {
+		case ALTERNATION :
+			if (in_alt)
+				fprintf(fgram, "%d,\n", LLALT);
+			else
+				in_alt = 1;
+			save_rule(g_getlink(p)->l_rule, tail);
+			break;
+		case TERM :
+			/* Make entry in term list */
+			(t_list + t_list_index)->term = g_getterm(p);
+			/* Test for [] without specifier */
+			if (g_getterm(p) == (struct term *) 0) {
+				t_list_index++;
+				break;
+			}
+			(t_list + t_list_index++)->t_nt_num = ++nt_highest;
+			fprintf(fgram, "%d, ", nt_highest);
+			/* Check if repetition,	 if so handle here */
+			if (r_getkind(g_getterm(p)) == FIXED)
+			{
+				int k;
+				for (k = 1; k < r_getnum(g_getterm(p)); k++)
+					fprintf(fgram, "%d, ", nt_highest);
+			}
+			break;
+		case NONTERM :
+			fprintf(fgram, "%d, ", g_getcont(p) + assval);
+			break;
+		case TERMINAL:
+			if (g_getcont(p) == g_getcont(illegal_gram)) {
+				/* %illegal. Ignore. */
+				break;
+			}
+			if (p->g_erroneous)
+				fprintf(fgram, "%d, ", illegal_num);
+			else
+				fprintf(fgram, "%d, ",
+						tokens[g_getcont(p)].t_tokno);
+			break;
+		case LITERAL:
+			if (p->g_erroneous)
+				fprintf(fgram, "%d, ", illegal_num);
+			else
+				fprintf(fgram, "%d, ",
+						tokens[g_getcont(p)].t_tokno);
+			break;
+		case ACTION:
+			if (subpars_sim) {
+				fprintf(fgram, "%d, ", act_nt);
+			}
+			else if (g_getsubparse(p)) {
+				/* Allocate nonterminal that will simulate
+				   subparser
+				 */
+				(sub_list + sub_list_index)->sub_nt_num =
+								++nt_highest;
+				(sub_list + sub_list_index++)->sub_action = p;
+
+				fprintf(fgram, "%d, ", nt_highest);
+			}
+			break;
+		case EORULE :
+			if ((! in_alt) && tail )
+			/* If this rule is not an alt, append tail now.
+			   If it is an alt, the recursive call of this function
+			   has appended tail to each alternative
+			 */
+			fprintf(fgram, "%d, ", tail);
+			return;
+		}
+		p++;
+	}
+}
+
+STATIC
+save_set(p) p_set p; {
+	register int k;
+	register unsigned i;
+	int j;
+
+	j = nbytes;
+	for (;;) {
+		i = (unsigned) *p++;
+		for (k = 0; k < sizeof(int); k++) {
+			fprintf(fgram,"0%o,",(int)(i & 0377));
+			i >>= 8;
+			if (--j == 0) {
+				fputs("\n",fgram);
+				return;
+			}
+		}
+	}
+	/* NOTREACHED */
+}
+#endif

util/LLgen/src/sets.c

@@ -0,0 +1,218 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * sets.c
+ * Set manipulation and allocation routines.
+ */
+
+# include "types.h"
+# include "extern.h"
+# include "sets.h"
+# include "assert.h"
+
+# ifndef NORCSID
+static string rcsid9 = "$Id$";
+# endif
+
+/* In this file the following routines are defined: */
+extern		setinit();
+extern p_set	setalloc();
+extern p_set	get_set();
+extern int	setunion();
+extern int	setintersect();
+extern		setminus();
+extern int	setempty();
+extern int	findindex();
+extern int	setcount();
+
+int		nbytes;
+static int	setsize;
+int		tsetsize;
+p_set		*setptr, *maxptr;
+static t_info	set_info;
+p_mem		alloc();
+
+setinit(nt_needed) {
+	/*
+	 * Initialises some variables needed for setcomputations
+	 */
+	register int	 bitset;
+
+	nbytes = NBYTES(ntokens);
+	bitset = ALIGN(nbytes);
+	tsetsize = NINTS(bitset);
+	if (nt_needed) {
+		/* nonterminals must be included in the sets */
+		bitset += NBYTES(nnonterms);
+	}
+	setsize = NINTS(bitset);
+	set_info.i_esize = sizeof(p_set);
+	set_info.i_incr = 20;
+}
+
+p_set
+get_set() {
+	/*
+	 * Allocate a set that cannot be freed
+	 */
+	register p_set p, q;
+	static p_set sets, maxsets;
+
+	if ((p = sets) >= maxsets) {
+		q = p = (p_set) alloc((unsigned) (50*setsize*sizeof(*sets)));
+		maxsets = p + 50 * setsize;
+		do {
+			*q++ = 0;
+		} while (q < maxsets);
+	}
+	sets = p + setsize;;
+	return p;
+}
+
+p_set
+setalloc() {
+	/*
+	 * Allocate a set which can later be freed.
+	 */
+	register p_set	p;
+	register int	size = setsize;
+
+	p = (p_set) alloc((unsigned) (size * sizeof(*p))) + size;
+	do {
+		*--p = 0;
+	} while (--size);
+	return p;
+}
+
+int
+setunion(a,b) register p_set a,b; {
+	/*
+	 * a = a union b.
+	 * Return 1 if the set a changed
+	 */
+	register int	i;
+	register int	j;
+	register int	nsub = 0;
+
+	i = setsize;
+	do {
+		*a = (j = *a) | *b++;
+		if (*a++ != j) {
+			nsub = 1;
+		}
+	} while (--i);
+	return nsub;
+}
+
+int
+setintersect(a,b) register p_set a,b; {
+	/*
+	 * a = a intersect b.
+	 * return 1 if the result is empty
+	 */
+	register int	i;
+	register int	nempty;
+
+	nempty = 1;
+	i =  setsize;
+	do {
+		if (*a++ &= *b++) nempty = 0;
+	} while (--i);
+	return nempty;
+}
+
+setminus(a,b) register p_set a,b; {
+	/*
+	 * a = a setminus b
+	 */
+	register int	i;
+
+	i = setsize;
+	do {
+		*a++ &= ~(*b++);
+	} while (--i);
+}
+
+int
+setempty(p) register p_set p; {
+	/*
+	 * Return 1 if the set p is empty
+	 */
+	register int	i;
+
+	i = tsetsize;
+	do {
+		if (*p++) return 0;
+	} while (--i);
+	return 1;
+}
+
+int
+findindex(set) p_set set; {
+	/*
+	 * The set "set" will serve as a recovery set.
+	 * Search for it in the table. If not present, enter it.
+	 * Here is room for improvement. At the moment, the list of
+	 * sets is examined with linear search.
+	 */
+	register p_set	*t;
+	p_mem		new_mem();
+	register p_set	a;
+	register p_set	b;
+	register int	i;
+	int		saved;
+
+	/*
+	 * First search for the set in the table
+	 */
+	for (t = setptr; t < maxptr; t++) {
+		a = *t;
+		b = set;
+		i = tsetsize;
+		do {
+			if (*a++ != *b++) break;
+		} while (--i);
+		if (i) continue;
+		/*
+		 * Here, the sets are equal.
+		 */
+		return nbytes * (t - setptr);
+	}
+	/*
+	 * Now check if the set consists of only one element.
+	 * It would be a waste to use a set for that
+	 */
+	if (setcount(set, &saved) == 1) return -(saved + 1);
+	/*
+	 * If it does, return its number as a negative number.
+	 */
+	maxptr = (p_set *) new_mem(&set_info);
+	setptr = (p_set *) set_info.i_ptr;
+	*maxptr = setalloc();
+	setunion(*maxptr, set);
+	return nbytes * (maxptr++ - setptr);
+}
+
+int
+setcount(set, saved) register p_set set; int *saved; {
+	register int i, j;
+
+	for (j = 0, i = 0; i < ntokens; i++) {
+		if (IN(set,i)) {
+			j++;
+			*saved = i;
+		}
+	}
+	return j;
+}

util/LLgen/src/sets.h

@@ -0,0 +1,33 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * $Id$
+ * Some macros that deal with bitsets and their size
+ */
+
+# define BITS		(8 * sizeof (int))
+# define IN(a,i)	((a)[(i)/BITS] & (1<<((i) % BITS)))
+# define NTIN(a,i)	((a)[(i)/BITS+tsetsize]&(1<<((i)%BITS)))
+# define PUTIN(a,i)	((a)[(i)/BITS] |=(1<<((i) % BITS)))
+# define NTPUTIN(a,i)	((a)[(i)/BITS+tsetsize]|=(1<<((i)%BITS)))
+# define NBYTES(n)	(((n) + 7) / 8)
+/*
+ * The next two macros operate on byte counts!
+ */
+# define NINTS(n)	(((n) + (int) (sizeof(int) - 1)) / (int) sizeof(int))
+# define ALIGN(n)	(NINTS(n) * (int) sizeof (int))
+
+extern int	tsetsize;
+extern p_set	*setptr, *maxptr;
+extern int	nbytes;

util/LLgen/src/tokens.g

@@ -0,0 +1,604 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * tokens.g
+ * Defines the tokens for the grammar of LLgen.
+ * The lexical analyser and LLmessage are also included here.
+ */
+
+{
+# include "types.h"
+# include "io.h"
+# include "extern.h"
+# include "assert.h"
+# include "cclass.h"
+
+# ifndef NORCSID
+static string	rcsidc = "$Id$";
+# endif
+
+/* Here are defined : */
+extern int	scanner();
+extern		LLmessage();
+extern int	input();
+extern		unput();
+extern		skipcomment();
+# ifdef LINE_DIRECTIVE
+STATIC		linedirective();
+# endif
+STATIC string	cpy();
+STATIC string	vallookup();
+STATIC		copyact();
+
+static int	nparams;
+}
+/* Classes */
+
+%token	C_IDENT ;	/* lextoken.t_string contains the identifier read */
+%token	C_NUMBER ;	/* lextoken.t_num contains the number read */
+%token	C_LITERAL ;	/* lextoken.t_string contains the literal read */
+%token	C_EXPR ;	/* A C expression (%if or %while) */
+%token	C_PARAMS ;	/* formal or actual parameters */
+%token	C_ACTION ;	/* a C action */
+
+/* Keywords */
+
+%token	C_TOKEN ;
+%token	C_START ;
+%token	C_IF ;
+%token	C_WHILE ;
+%token	C_PERSISTENT ;
+%token	C_FIRST ;
+%token	C_LEXICAL ;
+%token	C_PREFIX ;
+%token	C_ONERROR ;
+%token	C_AVOID ;
+%token	C_PREFER ;
+%token	C_DEFAULT ;
+%token	C_SUBSTART ;
+%token	C_ERRONEOUS ;
+%token	C_ILLEGAL ;
+
+%lexical scanner ;
+
+{
+
+/*
+ * Structure for a keyword
+ */
+
+typedef struct keyword {
+	string	w_word;
+	int	w_value;
+} t_keyw, *p_keyw;
+
+/*
+ * The list of keywords, the most often used keywords come first.
+ * Linear search is used, as there are not many keywords
+ */
+
+static t_keyw resword[] = {
+	{ "token",	C_TOKEN		},
+	{ "avoid",	C_AVOID		},
+	{ "prefer",	C_PREFER	},
+	{ "persistent", C_PERSISTENT	},
+	{ "default",	C_DEFAULT	},
+	{ "if",		C_IF		},
+	{ "while",	C_WHILE		},
+	{ "first",	C_FIRST		},
+	{ "start",	C_START		},
+	{ "lexical",	C_LEXICAL	},
+	{ "onerror",	C_ONERROR	},
+	{ "prefix",	C_PREFIX	},
+#ifdef NON_CORRECTING
+	{ "substart",	C_SUBSTART	},
+	{ "erroneous",	C_ERRONEOUS	},
+	{ "illegal",	C_ILLEGAL	},
+#endif
+	{ 0,		0		}
+};
+
+static t_token	savedtok;	/* to save lextoken in case of an insertion */
+# ifdef LINE_DIRECTIVE
+static	int	nostartline;	/* = 0 if at the start of a line */
+# endif
+
+STATIC
+copyact(ch1,ch2,flag,level) char ch1,ch2; {
+	/*
+	 * Copy an action to file f. Opening bracket is ch1, closing bracket
+	 * is ch2.
+	 * If flag & 1, copy opening and closing parameters too.
+	 * If flag & 2, don't allow ','.
+	 */
+	static int	text_seen = 0;
+	register	FILE *f;
+	register	ch;		/* Current char */
+	register	match;		/* used to read strings */
+	int		saved = linecount;
+					/* save linecount */
+	int		sav_strip = strip_grammar;
+
+	f = fact;
+	if (ch1 == '{' || flag != 1) strip_grammar = 0;
+	if (!level) {
+		text_seen = 0;
+		nparams = 0;			/* count comma's */
+		putc('\0',f);
+		fprintf(f,"# line %d \"%s\"\n", linecount,f_input);
+	}
+	if (level || (flag & 1)) putc(ch1,f);
+	for (;;) {
+		ch = input();
+		if (ch == ch2) {
+			if (!level) {
+				if (text_seen) nparams++;
+			}
+			if (level || (flag & 1)) putc(ch,f);
+			if (strip_grammar != sav_strip) {
+				if (ch1 == '{' || flag != 1) putchar(ch);
+			}
+			strip_grammar = sav_strip;
+			return;
+		}
+		switch(ch) {
+		  case ')':
+		  case '}':
+		  case ']':
+			error(linecount,"Parentheses mismatch");
+			break;
+		  case '(':
+			text_seen = 1;
+			copyact('(',')',flag,level+1);
+			continue;
+		  case '{':
+			text_seen = 1;
+			copyact('{','}',flag,level+1);
+			continue;
+		  case '[':
+			text_seen = 1;
+			copyact('[',']',flag,level+1);
+			continue;
+		  case '/':
+			ch = input();
+			unput(ch);
+			if (ch == '*') {
+				putc('/', f);
+				skipcomment(1);
+				continue;
+			}
+			ch = '/';
+			text_seen = 1;
+			break;
+		  case ';':
+		  case ',':
+			if (! level && text_seen) {
+				text_seen = 0;
+				nparams++;
+				if (ch == ',' && (flag & 2)) {
+					warning(linecount, "Parameters may not be separated with a ','");
+					ch = ';';
+				}
+			}
+			break;
+		  case '\'':
+		  case '"' :
+			/*
+			 * watch out for brackets in strings, they do not
+			 * count !
+			 */
+			text_seen = 1;
+			match = ch;
+			putc(ch,f);
+			while((ch = input())) {
+				if (ch == match) break;
+				if (ch == '\\') {
+					putc(ch,f);
+					ch = input();
+				}
+				if (ch == '\n') {
+					error(linecount,"Newline in string");
+					unput(match);
+				}
+				putc(ch,f);
+			}
+			if (ch == match) break;
+			/* Fall through */
+		    case EOF :
+			if (!level) error(saved,"Action does not terminate");
+			strip_grammar = sav_strip;
+			return;
+		    default:
+			if (c_class[ch] != ISSPA) text_seen = 1;
+		}
+		putc(ch,f);
+	}
+}
+
+scanner() {
+	/*
+	 * Lexical analyser, what else
+	 */
+	register int	ch;		/* Current char */
+	register char *p = ltext;
+	int		reserved = 0;	/* reserved word? */
+	char		*max = &ltext[LTEXTSZ - 1];
+	static int	nextexpr;
+	int		expect_expr = nextexpr;
+	long		off;
+
+	nextexpr = 0;
+	if (savedtok.t_tokno) {
+				/* A token has been inserted.
+				 * Now deliver the last lextoken again
+				 */
+		lextoken = savedtok;
+		savedtok.t_tokno = 0;
+		return lextoken.t_tokno;
+	}
+	for (;;) {
+		ch = input();
+		if (ch == EOF) return ch;
+# ifdef LINE_DIRECTIVE
+		if (ch == '#' && !nostartline) {
+			linedirective();
+			continue;
+		}
+# endif
+		switch(c_class[ch]) {
+		  case ISACT :
+			if (ch == '{') {
+				copyact('{', '}', in_production, 0);
+				return C_ACTION;
+			}
+			assert(ch == '(');
+			if (expect_expr) {
+				copyact('(', ')', 1, 0);
+				return C_EXPR;
+			}
+			off = ftell(fact);
+			copyact('(', ')', in_production != 0 ? 0 : 2, 0);
+			if (nparams == 0) fseek(fact, off, 0);
+			lextoken.t_num = nparams;
+			return C_PARAMS;
+		  case ISLIT :
+			for (;;) {
+				ch = input();
+				if (ch == '\n' || ch == EOF) {
+					error(linecount,"Missing '");
+					break;
+				}
+				if (ch == '\'') break;
+				if (ch == '\\') {
+					*p++ = ch;
+					ch = input();
+				}
+				*p++ = ch;
+				if (p > max) p--;
+			}
+			*p = '\0';
+			lextoken.t_string = ltext;
+			return C_LITERAL;
+		  case ISCOM :
+			skipcomment(0);
+			/* Fall through */
+		  case ISSPA :
+			continue;
+		  case ISDIG : {
+			register i = 0;
+			do {
+				i = 10 * i + (ch - '0');
+				ch= input();
+			} while (c_class[ch] == ISDIG);
+			lextoken.t_num = i;
+			unput(ch);
+			return C_NUMBER; }
+		  default:
+			return ch;
+		  case ISKEY :
+			reserved = 1;
+			ch = input();
+			/* Fall through */
+		  case ISLET :
+			do {
+				if (reserved && ch >= 'A' && ch <= 'Z') {
+					ch += 'a' - 'A';
+				}
+				*p++ = ch;
+				if (p > max) p--;
+				ch = input();
+			} while (c_class[ch] == ISDIG || c_class[ch] == ISLET);
+			unput(ch);
+			*p = '\0';
+			if (reserved) {	/*
+					 * Now search for the keyword
+					 */
+				register p_keyw w;
+
+				w = resword;
+				while (w->w_word) {
+					if (! strcmp(ltext,w->w_word)) {
+						/*
+						 * Return token number.
+						 */
+						if (w->w_value == C_IF ||
+						    w->w_value == C_WHILE) {
+							nextexpr = 1;
+						}
+						return w->w_value;
+					}
+					w++;
+				}
+				error(linecount,"Illegal reserved word");
+			}
+			lextoken.t_string = ltext;
+			return C_IDENT;
+		}
+	}
+}
+
+static int	backupc;	/* for unput() */
+static int	nonline;	/* = 1 if last char read was a newline */
+
+input() {
+	/*
+	 * Low level input routine, used by all other input routines
+	 */
+	register	c;
+
+	if (c = backupc) {
+			/* Last char was "unput()". Deliver it again
+			 */
+		backupc = 0;
+		return c;
+	}
+	if ((c = getc(finput)) == EOF) {
+		nonline = 0;
+		return c;
+	}
+# ifdef LINE_DIRECTIVE
+	nostartline = 1;
+# endif
+	if (!nonline) {
+		linecount++;
+# ifdef LINE_DIRECTIVE
+		nostartline = 0;
+# endif
+		nonline = 1;
+	}
+	if (c == '\n') nonline = 0;
+	if (strip_grammar) putchar(c);
+	return c;
+}
+
+unput(c) {
+	/*
+	 * "unread" c
+	 */
+	backupc = c;
+}
+
+skipcomment(flag) {
+	/*
+	 * Skip comment. If flag != 0, the comment is inside a fragment
+	 * of C-code, so keep it.
+	 */
+	register int	ch;
+	int		saved;	/* line count on which comment starts */
+
+	saved = linecount;
+	if (input() != '*') error(linecount,"Illegal comment");
+	if (flag) putc('*', fact);
+	do {
+		ch = input();
+		if (flag) putc(ch, fact);
+		while (ch == '*') {
+			ch = input();
+			if (flag) putc(ch, fact);
+			if (ch == '/') return;
+		}
+	} while (ch != EOF);
+	error(saved,"Comment does not terminate");
+}
+
+# ifdef LINE_DIRECTIVE
+STATIC
+linedirective() {
+	/*
+	 * Read a line directive
+	 */
+	register int	ch;
+	register int	i;
+	string		s_error = "Illegal line directive";
+	string		store();
+	register string	c;
+
+	do {	/*
+		 * Skip to next digit
+		 * Do not skip newlines
+		 */
+		ch = input();
+	} while (ch != '\n' && c_class[ch] != ISDIG);
+	if (ch == '\n') {
+		error(linecount,s_error);
+		return;
+	}
+	i = 0;
+	do  {
+		i = i*10 + (ch - '0');
+		ch = input();
+	} while (c_class[ch] == ISDIG);
+	while (ch != '\n' && ch != '"') ch = input();
+	if (ch == '"') {
+		c = ltext;
+		do {
+			*c++ = ch = input();
+		} while (ch != '"' && ch != '\n');
+		if (ch == '\n') {
+			error(linecount,s_error);
+			return;
+		}
+		*--c = '\0';
+		do {
+			ch = input();
+		} while (ch != '\n');
+		/*
+		 * Remember the file name
+		 */
+		if (strcmp(f_input,ltext)) f_input = store(ltext);
+	}
+	linecount = i;
+}
+# endif
+
+STATIC string
+vallookup(s) {
+	/*
+	 * Look up the keyword that has token number s
+	 */
+	register p_keyw p = resword;
+
+	while (p->w_value) {
+		if (p->w_value == s) return p->w_word;
+		p++;
+	}
+	return 0;
+}
+
+STATIC string
+cpy(s,p,inserted) register string p; {
+	/*
+	 * Create a piece of error message for token s and put it at p.
+	 * inserted = 0 if the token s was deleted (in which case we have
+	 * attributes), else it was inserted
+	 */
+	register string t = 0;
+
+	switch(s) {
+	  case C_IDENT :
+		if (!inserted) t = lextoken.t_string;
+		else t = "identifier";
+		break;
+	  case C_NUMBER :
+		t = "number";
+		break;
+	  case C_LITERAL :
+		if (!inserted) {
+			*p++ = '\'';
+			t = lextoken.t_string;
+			break;
+		}
+		t = "literal";
+		break;
+	  case C_ACTION:
+		t = "C action";
+		break;
+	  case C_PARAMS:
+		t = "C parameter section";
+		break;
+	  case C_EXPR:
+		t = "C expression";
+		break;
+	  case EOFILE :
+		t = "end-of-file";
+		break;
+	}
+	if (!t && (t = vallookup(s))) {
+		*p++ = '%';
+	}
+	if (t) {	/*
+			 * We have a string for the token. Copy it
+			 */
+		while (*t) *p++ = *t++;
+		if (s == C_LITERAL && !inserted) {
+			*p++ = '\'';
+		}
+		return p;
+	}
+	/*
+	 * The token is a literal
+	 */
+	*p++ = '\'';
+	if (s >= 040 && s <= 0176) *p++ = s;
+	else {
+	    *p++ = '\\';
+	    switch(s) {
+	      case '\b' : *p++ = 'b'; break;
+	      case '\f' : *p++ = 'f'; break;
+	      case '\n' : *p++ = 'n'; break;
+	      case '\r' : *p++ = 'r'; break;
+	      case '\t' : *p++ = 't'; break;
+	      default : *p++='0'+((s&0377)>>6); *p++='0'+((s>>3)&07);
+			*p++='0'+(s&07);
+	    }
+	}
+	*p++ = '\'';
+	return p;
+}
+
+string strcpy();
+
+LLmessage(d) {
+	/*
+	 * d is either 0, in which case the current token has been deleted,
+	 * or non-zero, in which case it represents a token that is inserted
+	 * before the current token
+	 */
+	register string	s,t;
+	char		buf[128];
+
+	nerrors++;
+	s = buf;
+	if (d < 0) {
+		strcpy(buf, "end-of-file expected");
+	}
+	else if (d == 0) {
+#ifdef LLNONCORR
+		t = " unexpected";
+#else
+		t = " deleted";
+#endif
+		s = cpy(LLsymb,s,0);
+		do *s++ = *t; while (*t++);
+	} else {
+		s = cpy(d,s,1);
+		t = " inserted in front of ";
+		do *s++ = *t++; while (*t);
+		s = cpy(LLsymb,s,0);
+		*s = '\0';
+	}
+	if (d > 0) {	/*
+			 * Save the current token and make up some
+			 * attributes for the inserted token
+			 */
+		savedtok = lextoken;
+		savedtok.t_tokno = LLsymb;
+		if (d == C_IDENT) lextoken.t_string = "dummy_identifier";
+		else if (d == C_LITERAL) lextoken.t_string = "dummy_literal";
+		else if (d == C_NUMBER) lextoken.t_num = 1;
+	}
+#ifdef LLNONCORR
+	else
+#endif
+	error(linecount, "%s", buf);
+			/* Don't change this line to 
+			 * error(linecount, buf).
+			 * The string in "buf" might contain '%' ...
+			 */
+#ifdef LLNONCORR
+	in_production = 1;
+			/* To prevent warnings from copyact */
+#endif
+}
+}

util/LLgen/src/types.h

@@ -0,0 +1,272 @@
+/* Copyright (c) 1991 by the Vrije Universiteit, Amsterdam, the Netherlands.
+ * For full copyright and restrictions on use see the file COPYING in the top
+ * level of the LLgen tree.
+ */
+
+/*
+ *  L L G E N
+ *
+ *  An Extended LL(1) Parser Generator
+ *
+ *  Author : Ceriel J.H. Jacobs
+ */
+
+/*
+ * $Id$
+ * Type and structure definitions
+ */
+
+typedef	int	*p_set;		/* pointer to bitset */
+typedef char	*p_mem;		/* pointer to some core */
+typedef char	*string;
+
+/*
+ * structure for a token
+ */
+typedef struct token {
+	int	t_tokno;	/* Lexical token number */
+	union {
+		string	t_s;	/* Attribute is either a string or */
+		int	t_v;	/* an integer */
+	} t_x;
+# define t_string t_x.t_s
+# define t_num t_x.t_v
+	int	t_flags;
+	int	t_next;
+	int	t_lineno;
+} t_token, *p_token;
+
+/*
+ * structure for the grammar elements
+ */
+typedef struct gram {
+	int	x;		/* for lay-out see comment below */
+	int	g_lineno;	/* element found on this line number */
+#ifdef NON_CORRECTING
+	int	g_erroneous;	/* 1 if element declared erroneous */
+#endif
+	union {
+		int	g_index;
+		struct term *	g_term;
+		struct link *	g_link;
+#ifdef NON_CORRECTING
+		/* If this is an action with a %substart g_subparse
+		   points to the list of startsymbols of the subparser */
+		struct ff_firsts *g_subparse;
+#endif
+	} g_i;
+} t_gram,*p_gram;
+
+/*
+ * Layout of the field x:
+ *
+ * 15  ....... 7 6 ........ 3 2 .... 0
+ * -----------------------------------
+ * | unused    | | nparams  | | type |
+ * -----------------------------------
+ */
+# define EORULE		00	/* End of (sub)rule */
+# define ACTION		01	/* Imbedded action */
+# define NONTERM	02	/* A nonterminal */
+# define TERMINAL	03	/* A terminal */
+# define TERM		04	/* Something between square brackets */
+# define ALTERNATION	05	/* Alternation (|) */
+# define LITERAL	06	/* Also a terminal */
+
+/*
+ * Access macros for the x-field of a grammar element
+ */
+# define g_gettype(p)	((p)->x&07)
+# define g_getcont(p)	((p)->g_i.g_index)
+# define g_getterm(p)	((p)->g_i.g_term)
+# define g_getlink(p)	((p)->g_i.g_link)
+# define g_getnpar(p)	(((p)->x>>3)&017)
+# define g_settype(p,s)	{ assert(((unsigned)(s))<=07);(p)->x=((p)->x&~07)|(s);}
+# define g_setcont(p,s) ((p)->g_i.g_index=(s))
+# define g_setterm(p,s)	((p)->g_i.g_term = (s))
+# define g_setlink(p,s)	((p)->g_i.g_link = (s))
+# define g_setnpar(p,s) { assert(((unsigned)(s))<=017);(p)->x=((p)->x&~0170)|((s)<<3);}
+#ifdef NON_CORRECTING
+# define g_getsubparse(p)	((p)->g_i.g_subparse)
+# define g_setsubparse(p,s)	((p)->g_i.g_subparse = (s))
+#endif
+/*
+ * Some constants to communicate with the symbol table search routine
+ */
+# define UNKNOWN	00	/* Not equal to NONTERM, TERMINAL or LITERAL */
+
+/*
+ * Some constants for safety
+ */
+# define SAFE		0	/* Indicates that a scan is done, and that the
+				 * token is correct
+				 */
+# define SAFESCANDONE	1	/* Indicates that a scan is done, and that the
+				 * token will not be skipped
+				 */
+# define SCANDONE	2	/* Indicates that a scan is done */
+# define NOSCANDONE	3	/* Indicates that no scan is done */
+# define NOSAFETY	4	/* Safety not yet computed */
+
+/*
+ * nonterminal structure
+ */
+typedef	struct {
+	int	n_flags;	/* low order four bits are reserved
+				 * the parameter count
+				 */
+# define getntparams(p)	((p)->n_flags&017)
+# define setntparams(p,i)	{assert(((unsigned)(i))<=017);(p)->n_flags&=~017;(p)->n_flags|=(i);}
+# define GENSTATIC	01000	/* set if routine can be made static */
+# define RECURSIVE	02000	/* Set if the default rule is recursive */
+# define PARAMS		04000	/* tells if a nonterminal has parameters */
+# define EMPTY		010000	/* tells if a nonterminal produces empty */
+# define LOCALS		020000	/* local declarations ? */
+# define REACHABLE	040000	/* can this nonterminal be reached ? */
+# define VERBOSE	0100000	/* Set if in LL.output file */
+	char	n_insafety;
+	char	n_outsafety;
+# define getntsafe(p)	((p)->n_insafety)
+# define setntsafe(p,i)	{assert(((unsigned)(i))<=NOSAFETY);(p)->n_insafety=(i);}
+# define getntout(p)	((p)->n_outsafety)
+# define setntout(p,i)	{assert(((unsigned)(i))<=NOSAFETY);(p)->n_outsafety=(i);}
+	int	n_count;	/* pieces of code before this rule */
+	int	n_lineno;	/* declared on line ... */
+	p_gram	n_rule;		/* pointer to right hand side of rule */
+	union {
+		p_set	n_f;	/* ptr to "first" set */
+		string	n_s;	/* If this nonterminal is not declared,
+				 * This field indicates the filename in
+				 * which it occurred
+				 */
+	} n_x;
+# define n_first  n_x.n_f
+# define n_string n_x.n_s
+#ifdef NON_CORRECTING
+	p_set	n_nc_first;	/* Pointer to non-corr first set */
+	p_set	n_nc_follow;	/* Pointer to non-corr follow set */
+#endif
+	p_set	n_follow;	/* pointer to the "follow" set	*/
+	p_set	n_contains;	/* pointer to symbols that can be produced */
+	string	n_name;		/* name of nonterminal */
+	int	n_next;		/* index of next nonterminal */
+	long	n_off;		/* index of parameters in action file */
+} t_nont, *p_nont;
+
+/*
+ * hash table structure
+ */
+typedef struct h_entry {
+	string		h_name;		/* pointer to name */
+	t_gram		h_type;		/* Type and index */
+	struct h_entry	*h_next;	/* next in hash chain */
+} t_entry, *p_entry;
+
+/*
+ * link	structure to link alternations
+ */
+typedef struct link {
+	unsigned int	l_flag;
+# define COND		01		/* Set if this alternative has a %if */
+# define DEF		02		/* This alternative is default */
+# define PREFERING	010		/* %prefer */
+# define AVOIDING	020		/* %avoid */
+# define NOCONF		01000		/* Set if there is a resolver without
+					 * conflict
+					 */
+	p_gram		l_rule;		/* pointer to this rule	*/
+	p_set		l_symbs;	/* set,	when to take this rule */
+#ifdef NON_CORRECTING
+	p_set		l_nc_symbs;
+#endif
+	p_set		l_others;	/* set, when to take another rule */
+} t_link, *p_link;
+
+/*
+ * Structure for a repitition specification
+ */
+typedef int t_reps,*p_reps;
+
+# define FIXED		00	/* a fixed number */
+# define STAR		01	/* 0 or more times */
+# define PLUS		02	/* 1 or more times */
+# define OPT		03	/* 0 or 1 times */
+
+/*
+ * Access macros for repitition in term
+ */
+# define r_getkind(q)	((q)->t_repeats&03)
+# define r_getnum(q)	(((q)->t_repeats>>2)&037777)
+# define r_setkind(q,s)	{ assert(((unsigned)(s))<=03);(q)->t_repeats=((q)->t_repeats&0177774)|(s);}
+# define r_setnum(q,s)	{ assert(((unsigned)(s))<=037777);(q)->t_repeats=((q)->t_repeats&03)|((s)<<2);}
+
+/*
+ * header structure for	a term
+ */
+typedef struct term {
+	t_reps	t_repeats;
+	int	t_flags;	/* Low order three bits for safety */
+# define gettout(q)	((q)->t_flags&07)
+# define settout(q,i)	{assert(((unsigned)(i))<=NOSAFETY);(q)->t_flags&=~07;(q)->t_flags|=i;}
+# define PERSISTENT	010	/* Set if this term has %persistent */
+# define RESOLVER	020	/* Set if this term has %while */
+# define EMPTYFIRST	0100	/* Error, empty first */
+# define EMPTYTERM	0200	/* Error, term can produce empty */
+/* # define NOCONF	01000	see link structure */
+
+	p_gram	t_rule;		/* pointer to this term	*/
+	p_set	t_follow;	/* set of followers */
+	p_set	t_first;	/* set of firsts */
+#ifdef NON_CORRECTING
+	p_set	t_nc_first;	/* set of non corr firsts */
+	p_set	t_nc_follow;	/* set of non corr followers */
+#endif
+	p_set	t_contains;	/* contains set */
+} t_term, *p_term;
+
+/*
+ * structure for firstmacros list
+ */
+typedef struct ff_firsts {
+	string	ff_name;	/* Name of the macro */
+	int	ff_nont;	/* for this nonterminal */
+	struct ff_firsts *ff_next;
+} t_first, *p_first;
+
+/*
+ * structure for start symbol list
+ */
+typedef t_first t_start;
+typedef p_first p_start;
+
+/*
+ * structure for file names and info
+ */
+typedef struct f_file {
+	string	f_name;		/* File name */
+	p_first f_firsts;	/* ptr to list of firstmacros that must be
+				 * generated in the target file for this
+				 * grammar file
+				 */
+	int f_nonterminals;	/* list of nonterminals in this file  */
+	int f_terminals;	/* list of terminals in this file  */
+	p_set f_used;		/* set of nonterminals used in this file */
+} t_file, *p_file;
+
+typedef struct info_alloc {
+	/*
+	 * Structure used for dynamically growing arrays
+	 */
+	unsigned i_size;	/* Size of the array */
+	unsigned i_esize;	/* Size of an element */
+	unsigned i_incr;	/* When filled, add room for i_incr elements */
+	p_mem	i_ptr;		/* ptr to base of array */
+	p_mem	i_max;		/* ptr to first free */
+	p_mem	i_top;		/* ptr to top of array */
+} t_info, *p_info;
+
+# ifdef NDEBUG
+# define STATIC static
+# else /* not NDEBUG */
+# define STATIC extern
+# endif /* not NDEBUG */

util/ass/Makefile

@@ -1,78 +0,0 @@
-d=../..
-l=$d/mach/m68k2/lib
-h=$d/h
-ASS_PATH=$l/em_ass
-CC=pmds
-
-SEP_OPT=-n
-
-CFLAGS=-O -.c
-
-all:            ass$(SEP_OPT)
-
-clean:
-		-rm -f ass-i ass-n *.o maktab *.old asstb.c
-
-install :       all
-		cp ass$(SEP_OPT) $(ASS_PATH)
-
-cmp :           all
-		cmp ass$(SEP_OPT) $(ASS_PATH)
-
-lint:           ass00.c ass30.c ass40.c ass50.c ass60.c ass70.c \
-			ass80.c assci.c assda.c assrl.c asstb.c asscm.c
-		lint -hpvbx \
-			ass00.c ass30.c ass40.c ass50.c ass60.c ass70.c \
-			ass80.c assci.c assda.c assrl.c asstb.c asscm.c
-
-
-ass-n:       ass00.o ass30.o ass40.o ass50.o ass60.o ass70.o \
-			ass80.o assci.o assda.o assrl.o asstb.o asscm.o \
-			$l/em_data.a
-		$(CC) $(CFLAGS) -s -o ass-n \
-			ass00.o ass30.o ass40.o ass50.o ass60.o ass70.o \
-			ass80.o assci.o assda.o assrl.o asstb.o asscm.o \
-			$l/em_data.a >ass.symb
-
-ass-i:       ass00.o ass30.o ass40.o ass50.o ass60.o ass70.o \
-			ass80.o assci.o assda.o assrl.o asstb.o asscm.o \
-			$l/em_data.a
-		$(CC) -i $(CFLAGS) -o ass-i \
-			ass00.o ass30.o ass40.o ass50.o ass60.o ass70.o \
-			ass80.o assci.o assda.o assrl.o asstb.o asscm.o \
-			$l/em_data.a
-
-ass00.o ass40.o ass60.o ass70.o ass80.o assrl.o: \
-		$h/local.h $h/em_spec.h $h/as_spec.h \
-		$h/em_flag.h $h/arch.h ass00.h assex.h
-
-assci.o:        $h/local.h $h/em_spec.h $h/as_spec.h \
-		$h/em_flag.h $h/em_mes.h $h/em_pseu.h \
-		$h/em_ptyp.h $h/arch.h ass00.h assex.h
-
-ass30.o ass50.o : \
-		$h/local.h $h/em_spec.h $h/as_spec.h \
-		$h/em_flag.h ip_spec.h ass00.h assex.h
-
-ass80.o:        $h/em_path.h
-
-assda.o:        $h/local.h $h/em_spec.h $h/as_spec.h \
-		$h/em_flag.h $h/arch.h ass00.h
-
-asscm.o:        ass00.h
-
-asstb.o:        asstb.c
-
-asstb.c:        maktab ip_spec.t
-		maktab ip_spec.t asstb.c
-
-maktab:         maktab.c $h/em_spec.h ip_spec.h $h/em_flag.h \
-		$l/em_data.a
-		$(CC) -O -o maktab maktab.c $l/em_data.a
-
-opr:
-		make pr ^ opr
-
-pr:
-		@(pr ass00.h assex.h ip_spec.h ass?0.c ass[rcd]?.c \
-		maktab.c ; pr -3 ip_spec.t)