xref: /freebsd/contrib/bc/src/lex.c (revision 9768746ba83efa02837c5b9c66348db6e900208f)
1 /*
2  * *****************************************************************************
3  *
4  * SPDX-License-Identifier: BSD-2-Clause
5  *
6  * Copyright (c) 2018-2023 Gavin D. Howard and contributors.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions are met:
10  *
11  * * Redistributions of source code must retain the above copyright notice, this
12  *   list of conditions and the following disclaimer.
13  *
14  * * Redistributions in binary form must reproduce the above copyright notice,
15  *   this list of conditions and the following disclaimer in the documentation
16  *   and/or other materials provided with the distribution.
17  *
18  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
19  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
22  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28  * POSSIBILITY OF SUCH DAMAGE.
29  *
30  * *****************************************************************************
31  *
32  * Common code for the lexers.
33  *
34  */
35 
36 #include <assert.h>
37 #include <ctype.h>
38 #include <stdbool.h>
39 #include <string.h>
40 
41 #include <lex.h>
42 #include <vm.h>
43 #include <bc.h>
44 
45 void
46 bc_lex_invalidChar(BcLex* l, char c)
47 {
48 	l->t = BC_LEX_INVALID;
49 	bc_lex_verr(l, BC_ERR_PARSE_CHAR, c);
50 }
51 
52 void
53 bc_lex_lineComment(BcLex* l)
54 {
55 	l->t = BC_LEX_WHITESPACE;
56 	while (l->i < l->len && l->buf[l->i] != '\n')
57 	{
58 		l->i += 1;
59 	}
60 }
61 
62 void
63 bc_lex_comment(BcLex* l)
64 {
65 	size_t i, nlines = 0;
66 	const char* buf;
67 	bool end = false, got_more;
68 	char c;
69 
70 	l->i += 1;
71 	l->t = BC_LEX_WHITESPACE;
72 
73 	// This loop is complex because it might need to request more data from
74 	// stdin if the comment is not ended. This loop is taken until the comment
75 	// is finished or we have EOF.
76 	do
77 	{
78 		buf = l->buf;
79 		got_more = false;
80 
81 		// If we are in stdin mode, the buffer must be the one used for stdin.
82 		assert(vm->mode != BC_MODE_STDIN || buf == vm->buffer.v);
83 
84 		// Find the end of the comment.
85 		for (i = l->i; !end; i += !end)
86 		{
87 			// While we don't have an asterisk, eat, but increment nlines.
88 			for (; (c = buf[i]) && c != '*'; ++i)
89 			{
90 				nlines += (c == '\n');
91 			}
92 
93 			// If this is true, we need to request more data.
94 			if (BC_ERR(!c || buf[i + 1] == '\0'))
95 			{
96 				// Read more, if possible.
97 				if (!vm->eof && l->mode != BC_MODE_FILE)
98 				{
99 					got_more = bc_lex_readLine(l);
100 				}
101 
102 				break;
103 			}
104 
105 			// If this turns true, we found the end. Yay!
106 			end = (buf[i + 1] == '/');
107 		}
108 	}
109 	while (got_more && !end);
110 
111 	// If we didn't find the end, barf.
112 	if (!end)
113 	{
114 		l->i = i;
115 		bc_lex_err(l, BC_ERR_PARSE_COMMENT);
116 	}
117 
118 	l->i = i + 2;
119 	l->line += nlines;
120 }
121 
122 void
123 bc_lex_whitespace(BcLex* l)
124 {
125 	char c;
126 
127 	l->t = BC_LEX_WHITESPACE;
128 
129 	// Eat. We don't eat newlines because they can be special.
130 	for (c = l->buf[l->i]; c != '\n' && isspace(c); c = l->buf[++l->i])
131 	{
132 		continue;
133 	}
134 }
135 
136 void
137 bc_lex_commonTokens(BcLex* l, char c)
138 {
139 	if (!c) l->t = BC_LEX_EOF;
140 	else if (c == '\n') l->t = BC_LEX_NLINE;
141 	else bc_lex_whitespace(l);
142 }
143 
144 /**
145  * Parses a number.
146  * @param l         The lexer.
147  * @param start     The start character.
148  * @param int_only  Whether this function should only look for an integer. This
149  *                  is used to implement the exponent of scientific notation.
150  */
151 static size_t
152 bc_lex_num(BcLex* l, char start, bool int_only)
153 {
154 	const char* buf = l->buf + l->i;
155 	size_t i;
156 	char c;
157 	bool last_pt, pt = (start == '.');
158 
159 	// This loop looks complex. It is not. It is asking if the character is not
160 	// a nul byte and it if it a valid num character based on what we have found
161 	// thus far, or whether it is a backslash followed by a newline. I can do
162 	// i+1 on the buffer because the buffer must have a nul byte.
163 	for (i = 0; (c = buf[i]) && (BC_LEX_NUM_CHAR(c, pt, int_only) ||
164 	                             (c == '\\' && buf[i + 1] == '\n'));
165 	     ++i)
166 	{
167 		// I don't need to test that the next character is a newline because
168 		// the loop condition above ensures that.
169 		if (c == '\\')
170 		{
171 			i += 2;
172 
173 			// Make sure to eat whitespace at the beginning of the line.
174 			while (isspace(buf[i]) && buf[i] != '\n')
175 			{
176 				i += 1;
177 			}
178 
179 			c = buf[i];
180 
181 			// If the next character is not a number character, bail.
182 			if (!BC_LEX_NUM_CHAR(c, pt, int_only)) break;
183 		}
184 
185 		// Did we find the radix point?
186 		last_pt = (c == '.');
187 
188 		// If we did, and we already have one, then break because it's not part
189 		// of this number.
190 		if (pt && last_pt) break;
191 
192 		// Set whether we have found a radix point.
193 		pt = pt || last_pt;
194 
195 		bc_vec_push(&l->str, &c);
196 	}
197 
198 	return i;
199 }
200 
201 void
202 bc_lex_number(BcLex* l, char start)
203 {
204 	l->t = BC_LEX_NUMBER;
205 
206 	// Make sure the string is clear.
207 	bc_vec_popAll(&l->str);
208 	bc_vec_push(&l->str, &start);
209 
210 	// Parse the number.
211 	l->i += bc_lex_num(l, start, false);
212 
213 #if BC_ENABLE_EXTRA_MATH
214 	{
215 		char c = l->buf[l->i];
216 
217 		// Do we have a number in scientific notation?
218 		if (c == 'e')
219 		{
220 #if BC_ENABLED
221 			// Barf for POSIX.
222 			if (BC_IS_POSIX) bc_lex_err(l, BC_ERR_POSIX_EXP_NUM);
223 #endif // BC_ENABLED
224 
225 			// Push the e.
226 			bc_vec_push(&l->str, &c);
227 			l->i += 1;
228 			c = l->buf[l->i];
229 
230 			// Check for negative specifically because bc_lex_num() does not.
231 			if (c == BC_LEX_NEG_CHAR)
232 			{
233 				bc_vec_push(&l->str, &c);
234 				l->i += 1;
235 				c = l->buf[l->i];
236 			}
237 
238 			// We must have a number character, so barf if not.
239 			if (BC_ERR(!BC_LEX_NUM_CHAR(c, false, true)))
240 			{
241 				bc_lex_verr(l, BC_ERR_PARSE_CHAR, c);
242 			}
243 
244 			// Parse the exponent.
245 			l->i += bc_lex_num(l, 0, true);
246 		}
247 	}
248 #endif // BC_ENABLE_EXTRA_MATH
249 
250 	bc_vec_pushByte(&l->str, '\0');
251 }
252 
253 void
254 bc_lex_name(BcLex* l)
255 {
256 	size_t i = 0;
257 	const char* buf = l->buf + l->i - 1;
258 	char c = buf[i];
259 
260 	l->t = BC_LEX_NAME;
261 
262 	// Should be obvious. It's looking for valid characters.
263 	while ((c >= 'a' && c <= 'z') || isdigit(c) || c == '_')
264 	{
265 		c = buf[++i];
266 	}
267 
268 	// Set the string to the identifier.
269 	bc_vec_string(&l->str, i, buf);
270 
271 	// Increment the index. We minus 1 because it has already been incremented.
272 	l->i += i - 1;
273 }
274 
275 void
276 bc_lex_init(BcLex* l)
277 {
278 	BC_SIG_ASSERT_LOCKED;
279 	assert(l != NULL);
280 	bc_vec_init(&l->str, sizeof(char), BC_DTOR_NONE);
281 }
282 
283 void
284 bc_lex_free(BcLex* l)
285 {
286 	BC_SIG_ASSERT_LOCKED;
287 	assert(l != NULL);
288 	bc_vec_free(&l->str);
289 }
290 
291 void
292 bc_lex_file(BcLex* l, const char* file)
293 {
294 	assert(l != NULL && file != NULL);
295 	l->line = 1;
296 	vm->file = file;
297 }
298 
299 void
300 bc_lex_next(BcLex* l)
301 {
302 	BC_SIG_ASSERT_LOCKED;
303 
304 	assert(l != NULL);
305 
306 	l->last = l->t;
307 
308 	// If this wasn't here, the line number would be off.
309 	l->line += (l->i != 0 && l->buf[l->i - 1] == '\n');
310 
311 	// If the last token was EOF, someone called this one too many times.
312 	if (BC_ERR(l->last == BC_LEX_EOF)) bc_lex_err(l, BC_ERR_PARSE_EOF);
313 
314 	l->t = BC_LEX_EOF;
315 
316 	// We are done if this is true.
317 	if (l->i == l->len) return;
318 
319 	// Loop until failure or we don't have whitespace. This
320 	// is so the parser doesn't get inundated with whitespace.
321 	do
322 	{
323 		vm->next(l);
324 	}
325 	while (l->t == BC_LEX_WHITESPACE);
326 }
327 
328 /**
329  * Updates the buffer and len so that they are not invalidated when the stdin
330  * buffer grows.
331  * @param l     The lexer.
332  * @param text  The text.
333  * @param len   The length of the text.
334  */
335 static void
336 bc_lex_fixText(BcLex* l, const char* text, size_t len)
337 {
338 	l->buf = text;
339 	l->len = len;
340 }
341 
342 bool
343 bc_lex_readLine(BcLex* l)
344 {
345 	bool good;
346 
347 	// These are reversed because they should be already locked, but
348 	// bc_vm_readLine() needs them to be unlocked.
349 	BC_SIG_UNLOCK;
350 
351 	// Make sure we read from the appropriate place.
352 	switch (l->mode)
353 	{
354 		case BC_MODE_EXPRS:
355 		{
356 			good = bc_vm_readBuf(false);
357 			break;
358 		}
359 
360 		case BC_MODE_FILE:
361 		{
362 			good = false;
363 			break;
364 		}
365 
366 		case BC_MODE_STDIN:
367 		{
368 			good = bc_vm_readLine(false);
369 			break;
370 		}
371 
372 #ifdef __GNUC__
373 #ifndef __clang__
374 		default:
375 		{
376 			// We should never get here.
377 			abort();
378 		}
379 #endif // __clang__
380 #endif // __GNUC__
381 	}
382 
383 	BC_SIG_LOCK;
384 
385 	bc_lex_fixText(l, vm->buffer.v, vm->buffer.len - 1);
386 
387 	return good;
388 }
389 
390 void
391 bc_lex_text(BcLex* l, const char* text, BcMode mode)
392 {
393 	BC_SIG_ASSERT_LOCKED;
394 
395 	assert(l != NULL && text != NULL);
396 
397 	bc_lex_fixText(l, text, strlen(text));
398 	l->i = 0;
399 	l->t = l->last = BC_LEX_INVALID;
400 	l->mode = mode;
401 
402 	bc_lex_next(l);
403 }
404