xref: /freebsd/bin/sh/mksyntax.c (revision eaf771995719ae91e7f195390a8c14ffc2a19002)
1 /*-
2  * Copyright (c) 1991, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * This code is derived from software contributed to Berkeley by
6  * Kenneth Almquist.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 4. Neither the name of the University nor the names of its contributors
17  *    may be used to endorse or promote products derived from this software
18  *    without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  */
32 
33 #if 0
34 #ifndef lint
35 static char const copyright[] =
36 "@(#) Copyright (c) 1991, 1993\n\
37 	The Regents of the University of California.  All rights reserved.\n";
38 #endif /* not lint */
39 
40 #ifndef lint
41 static char sccsid[] = "@(#)mksyntax.c	8.2 (Berkeley) 5/4/95";
42 #endif /* not lint */
43 #endif
44 #include <sys/cdefs.h>
45 __FBSDID("$FreeBSD$");
46 
47 /*
48  * This program creates syntax.h and syntax.c.
49  */
50 
51 #include <stdio.h>
52 #include <stdlib.h>
53 #include <string.h>
54 #include "parser.h"
55 
56 
57 struct synclass {
58 	const char *name;
59 	const char *comment;
60 };
61 
62 /* Syntax classes */
63 struct synclass synclass[] = {
64 	{ "CWORD",	"character is nothing special" },
65 	{ "CNL",	"newline character" },
66 	{ "CBACK",	"a backslash character" },
67 	{ "CSBACK",	"a backslash character in single quotes" },
68 	{ "CSQUOTE",	"single quote" },
69 	{ "CDQUOTE",	"double quote" },
70 	{ "CENDQUOTE",	"a terminating quote" },
71 	{ "CBQUOTE",	"backwards single quote" },
72 	{ "CVAR",	"a dollar sign" },
73 	{ "CENDVAR",	"a '}' character" },
74 	{ "CLP",	"a left paren in arithmetic" },
75 	{ "CRP",	"a right paren in arithmetic" },
76 	{ "CEOF",	"end of file" },
77 	{ "CCTL",	"like CWORD, except it must be escaped" },
78 	{ "CSPCL",	"these terminate a word" },
79 	{ "CIGN",       "character should be ignored" },
80 	{ NULL,		NULL }
81 };
82 
83 
84 /*
85  * Syntax classes for is_ functions.  Warning:  if you add new classes
86  * you may have to change the definition of the is_in_name macro.
87  */
88 struct synclass is_entry[] = {
89 	{ "ISDIGIT",	"a digit" },
90 	{ "ISUPPER",	"an upper case letter" },
91 	{ "ISLOWER",	"a lower case letter" },
92 	{ "ISUNDER",	"an underscore" },
93 	{ "ISSPECL",	"the name of a special parameter" },
94 	{ NULL, 	NULL }
95 };
96 
97 static char writer[] = "\
98 /*\n\
99  * This file was generated by the mksyntax program.\n\
100  */\n\
101 \n";
102 
103 
104 static FILE *cfile;
105 static FILE *hfile;
106 static const char *syntax[513];
107 static int base;
108 static int size;	/* number of values which a char variable can have */
109 static int nbits;	/* number of bits in a character */
110 
111 static void filltable(const char *);
112 static void init(void);
113 static void add(const char *, const char *);
114 static void print(const char *);
115 static void output_type_macros(void);
116 
117 int
118 main(int argc __unused, char **argv __unused)
119 {
120 	char c;
121 	char d;
122 	int sign;
123 	int i;
124 	char buf[80];
125 	int pos;
126 
127 	/* Create output files */
128 	if ((cfile = fopen("syntax.c", "w")) == NULL) {
129 		perror("syntax.c");
130 		exit(2);
131 	}
132 	if ((hfile = fopen("syntax.h", "w")) == NULL) {
133 		perror("syntax.h");
134 		exit(2);
135 	}
136 	fputs(writer, hfile);
137 	fputs(writer, cfile);
138 
139 	/* Determine the characteristics of chars. */
140 	c = -1;
141 	sign = (c > 0) ? 0 : 1;
142 	for (nbits = 1 ; ; nbits++) {
143 		d = (1 << nbits) - 1;
144 		if (d == c)
145 			break;
146 	}
147 #if 0
148 	printf("%s %d bit chars\n", sign? "signed" : "unsigned", nbits);
149 #endif
150 	if (nbits > 9) {
151 		fputs("Characters can't have more than 9 bits\n", stderr);
152 		exit(2);
153 	}
154 	size = (1 << nbits) + 1;
155 	base = 1;
156 	if (sign)
157 		base += 1 << (nbits - 1);
158 
159 	fputs("#include <sys/cdefs.h>\n", hfile);
160 
161 	/* Generate the #define statements in the header file */
162 	fputs("/* Syntax classes */\n", hfile);
163 	for (i = 0 ; synclass[i].name ; i++) {
164 		sprintf(buf, "#define %s %d", synclass[i].name, i);
165 		fputs(buf, hfile);
166 		for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07)
167 			putc('\t', hfile);
168 		fprintf(hfile, "/* %s */\n", synclass[i].comment);
169 	}
170 	putc('\n', hfile);
171 	fputs("/* Syntax classes for is_ functions */\n", hfile);
172 	for (i = 0 ; is_entry[i].name ; i++) {
173 		sprintf(buf, "#define %s %#o", is_entry[i].name, 1 << i);
174 		fputs(buf, hfile);
175 		for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07)
176 			putc('\t', hfile);
177 		fprintf(hfile, "/* %s */\n", is_entry[i].comment);
178 	}
179 	putc('\n', hfile);
180 	fprintf(hfile, "#define SYNBASE %d\n", base);
181 	fprintf(hfile, "#define PEOF %d\n\n", -base);
182 	putc('\n', hfile);
183 	fputs("#define BASESYNTAX (basesyntax + SYNBASE)\n", hfile);
184 	fputs("#define DQSYNTAX (dqsyntax + SYNBASE)\n", hfile);
185 	fputs("#define SQSYNTAX (sqsyntax + SYNBASE)\n", hfile);
186 	fputs("#define ARISYNTAX (arisyntax + SYNBASE)\n", hfile);
187 	putc('\n', hfile);
188 	output_type_macros();		/* is_digit, etc. */
189 	putc('\n', hfile);
190 
191 	/* Generate the syntax tables. */
192 	fputs("#include \"shell.h\"\n", cfile);
193 	fputs("#include \"syntax.h\"\n\n", cfile);
194 	init();
195 	fputs("/* syntax table used when not in quotes */\n", cfile);
196 	add("\n", "CNL");
197 	add("\\", "CBACK");
198 	add("'", "CSQUOTE");
199 	add("\"", "CDQUOTE");
200 	add("`", "CBQUOTE");
201 	add("$", "CVAR");
202 	add("}", "CENDVAR");
203 	add("<>();&| \t", "CSPCL");
204 	print("basesyntax");
205 	init();
206 	fputs("\n/* syntax table used when in double quotes */\n", cfile);
207 	add("\n", "CNL");
208 	add("\\", "CBACK");
209 	add("\"", "CENDQUOTE");
210 	add("`", "CBQUOTE");
211 	add("$", "CVAR");
212 	add("}", "CENDVAR");
213 	/* ':/' for tilde expansion, '-^]' for [a\-x] pattern ranges */
214 	add("!*?[]=~:/-^", "CCTL");
215 	print("dqsyntax");
216 	init();
217 	fputs("\n/* syntax table used when in single quotes */\n", cfile);
218 	add("\n", "CNL");
219 	add("\\", "CSBACK");
220 	add("'", "CENDQUOTE");
221 	/* ':/' for tilde expansion, '-^]' for [a\-x] pattern ranges */
222 	add("!*?[]=~:/-^", "CCTL");
223 	print("sqsyntax");
224 	init();
225 	fputs("\n/* syntax table used when in arithmetic */\n", cfile);
226 	add("\n", "CNL");
227 	add("\\", "CBACK");
228 	add("`", "CBQUOTE");
229 	add("\"", "CIGN");
230 	add("$", "CVAR");
231 	add("}", "CENDVAR");
232 	add("(", "CLP");
233 	add(")", "CRP");
234 	print("arisyntax");
235 	filltable("0");
236 	fputs("\n/* character classification table */\n", cfile);
237 	add("0123456789", "ISDIGIT");
238 	add("abcdefghijklmnopqrstuvwxyz", "ISLOWER");
239 	add("ABCDEFGHIJKLMNOPQRSTUVWXYZ", "ISUPPER");
240 	add("_", "ISUNDER");
241 	add("#?$!-*@", "ISSPECL");
242 	print("is_type");
243 	exit(0);
244 }
245 
246 
247 
248 /*
249  * Clear the syntax table.
250  */
251 
252 static void
253 filltable(const char *dftval)
254 {
255 	int i;
256 
257 	for (i = 0 ; i < size ; i++)
258 		syntax[i] = dftval;
259 }
260 
261 
262 /*
263  * Initialize the syntax table with default values.
264  */
265 
266 static void
267 init(void)
268 {
269 	filltable("CWORD");
270 	syntax[0] = "CEOF";
271 	syntax[base + CTLESC] = "CCTL";
272 	syntax[base + CTLVAR] = "CCTL";
273 	syntax[base + CTLENDVAR] = "CCTL";
274 	syntax[base + CTLBACKQ] = "CCTL";
275 	syntax[base + CTLBACKQ + CTLQUOTE] = "CCTL";
276 	syntax[base + CTLARI] = "CCTL";
277 	syntax[base + CTLENDARI] = "CCTL";
278 	syntax[base + CTLQUOTEMARK] = "CCTL";
279 	syntax[base + CTLQUOTEEND] = "CCTL";
280 }
281 
282 
283 /*
284  * Add entries to the syntax table.
285  */
286 
287 static void
288 add(const char *p, const char *type)
289 {
290 	while (*p)
291 		syntax[*p++ + base] = type;
292 }
293 
294 
295 
296 /*
297  * Output the syntax table.
298  */
299 
300 static void
301 print(const char *name)
302 {
303 	int i;
304 	int col;
305 
306 	fprintf(hfile, "extern const char %s[];\n", name);
307 	fprintf(cfile, "const char %s[%d] = {\n", name, size);
308 	col = 0;
309 	for (i = 0 ; i < size ; i++) {
310 		if (i == 0) {
311 			fputs("      ", cfile);
312 		} else if ((i & 03) == 0) {
313 			fputs(",\n      ", cfile);
314 			col = 0;
315 		} else {
316 			putc(',', cfile);
317 			while (++col < 9 * (i & 03))
318 				putc(' ', cfile);
319 		}
320 		fputs(syntax[i], cfile);
321 		col += strlen(syntax[i]);
322 	}
323 	fputs("\n};\n", cfile);
324 }
325 
326 
327 
328 /*
329  * Output character classification macros (e.g. is_digit).  If digits are
330  * contiguous, we can test for them quickly.
331  */
332 
333 static const char *macro[] = {
334 	"#define is_digit(c)\t((unsigned int)((c) - '0') <= 9)",
335 	"#define is_eof(c)\t((c) == PEOF)",
336 	"#define is_alpha(c)\t((is_type+SYNBASE)[(int)c] & (ISUPPER|ISLOWER))",
337 	"#define is_name(c)\t((is_type+SYNBASE)[(int)c] & (ISUPPER|ISLOWER|ISUNDER))",
338 	"#define is_in_name(c)\t((is_type+SYNBASE)[(int)c] & (ISUPPER|ISLOWER|ISUNDER|ISDIGIT))",
339 	"#define is_special(c)\t((is_type+SYNBASE)[(int)c] & (ISSPECL|ISDIGIT))",
340 	"#define digit_val(c)\t((c) - '0')",
341 	NULL
342 };
343 
344 static void
345 output_type_macros(void)
346 {
347 	const char **pp;
348 
349 	for (pp = macro ; *pp ; pp++)
350 		fprintf(hfile, "%s\n", *pp);
351 }
352