xref: /freebsd/contrib/bc/src/parse.c (revision e9ac41698b2f322d55ccf9da50a3596edb2c1800)
1 /*
2  * *****************************************************************************
3  *
4  * SPDX-License-Identifier: BSD-2-Clause
5  *
6  * Copyright (c) 2018-2024 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  * Code common to the parsers.
33  *
34  */
35 
36 #include <assert.h>
37 #include <stddef.h>
38 #include <stdlib.h>
39 #include <string.h>
40 
41 #include <limits.h>
42 
43 #include <parse.h>
44 #include <program.h>
45 #include <vm.h>
46 
47 void
48 bc_parse_updateFunc(BcParse* p, size_t fidx)
49 {
50 	p->fidx = fidx;
51 	p->func = bc_vec_item(&p->prog->fns, fidx);
52 }
53 
54 inline void
55 bc_parse_pushName(const BcParse* p, char* name, bool var)
56 {
57 	bc_parse_pushIndex(p, bc_program_search(p->prog, name, var));
58 }
59 
60 /**
61  * Updates the function, then pushes the instruction and the index. This is a
62  * convenience function.
63  * @param p     The parser.
64  * @param inst  The instruction to push.
65  * @param idx   The index to push.
66  */
67 static inline void
68 bc_parse_pushInstIdx(BcParse* p, uchar inst, size_t idx)
69 {
70 	bc_parse_push(p, inst);
71 	bc_parse_pushIndex(p, idx);
72 }
73 
74 void
75 bc_parse_addString(BcParse* p)
76 {
77 	size_t idx;
78 
79 	idx = bc_program_addString(p->prog, p->l.str.v);
80 
81 	// Push the string info.
82 	bc_parse_pushInstIdx(p, BC_INST_STR, idx);
83 }
84 
85 static void
86 bc_parse_addNum(BcParse* p, const char* string)
87 {
88 	BcProgram* prog = p->prog;
89 	size_t idx;
90 
91 	// XXX: This function has an implicit assumption: that string is a valid C
92 	// string with a nul terminator. This is because of the unchecked array
93 	// accesses below. I can't check this with an assert() because that could
94 	// lead to out-of-bounds access.
95 	//
96 	// XXX: In fact, just for safety's sake, assume that this function needs a
97 	// non-empty string with a nul terminator, just in case bc_parse_zero or
98 	// bc_parse_one change in the future, which I doubt.
99 
100 	BC_SIG_ASSERT_LOCKED;
101 
102 	// Special case 0.
103 	if (bc_parse_zero[0] == string[0] && bc_parse_zero[1] == string[1])
104 	{
105 		bc_parse_push(p, BC_INST_ZERO);
106 		return;
107 	}
108 
109 	// Special case 1.
110 	if (bc_parse_one[0] == string[0] && bc_parse_one[1] == string[1])
111 	{
112 		bc_parse_push(p, BC_INST_ONE);
113 		return;
114 	}
115 
116 	if (bc_map_insert(&prog->const_map, string, prog->consts.len, &idx))
117 	{
118 		BcConst* c;
119 		BcId* id = bc_vec_item(&prog->const_map, idx);
120 
121 		// Get the index.
122 		idx = id->idx;
123 
124 		// Push an empty constant.
125 		c = bc_vec_pushEmpty(&prog->consts);
126 
127 		// Set the fields. We reuse the string in the ID (allocated by
128 		// bc_map_insert()), because why not?
129 		c->val = id->name;
130 		c->base = BC_NUM_BIGDIG_MAX;
131 
132 		// We need this to be able to tell that the number has not been
133 		// allocated.
134 		bc_num_clear(&c->num);
135 	}
136 	else
137 	{
138 		BcId* id = bc_vec_item(&prog->const_map, idx);
139 		idx = id->idx;
140 	}
141 
142 	bc_parse_pushInstIdx(p, BC_INST_NUM, idx);
143 }
144 
145 void
146 bc_parse_number(BcParse* p)
147 {
148 #if BC_ENABLE_EXTRA_MATH
149 	char* exp = strchr(p->l.str.v, 'e');
150 	size_t idx = SIZE_MAX;
151 
152 	// Do we have a number in scientific notation? If so, add a nul byte where
153 	// the e is.
154 	if (exp != NULL)
155 	{
156 		idx = ((size_t) (exp - p->l.str.v));
157 		*exp = 0;
158 	}
159 #endif // BC_ENABLE_EXTRA_MATH
160 
161 	bc_parse_addNum(p, p->l.str.v);
162 
163 #if BC_ENABLE_EXTRA_MATH
164 	// If we have a number in scientific notation...
165 	if (exp != NULL)
166 	{
167 		bool neg;
168 
169 		// Figure out if the exponent is negative.
170 		neg = (*((char*) bc_vec_item(&p->l.str, idx + 1)) == BC_LEX_NEG_CHAR);
171 
172 		// Add the number and instruction.
173 		bc_parse_addNum(p, bc_vec_item(&p->l.str, idx + 1 + neg));
174 		bc_parse_push(p, BC_INST_LSHIFT + neg);
175 	}
176 #endif // BC_ENABLE_EXTRA_MATH
177 }
178 
179 void
180 bc_parse_text(BcParse* p, const char* text, BcMode mode)
181 {
182 	BC_SIG_LOCK;
183 
184 	// Make sure the pointer isn't invalidated.
185 	p->func = bc_vec_item(&p->prog->fns, p->fidx);
186 	bc_lex_text(&p->l, text, mode);
187 
188 	BC_SIG_UNLOCK;
189 }
190 
191 void
192 bc_parse_reset(BcParse* p)
193 {
194 	BC_SIG_ASSERT_LOCKED;
195 
196 	// Reset the function if it isn't main and switch to main.
197 	if (p->fidx != BC_PROG_MAIN)
198 	{
199 		bc_func_reset(p->func);
200 		bc_parse_updateFunc(p, BC_PROG_MAIN);
201 	}
202 
203 	// Reset the lexer.
204 	p->l.i = p->l.len;
205 	p->l.t = BC_LEX_EOF;
206 
207 #if BC_ENABLED
208 	if (BC_IS_BC)
209 	{
210 		// Get rid of the bc parser state.
211 		p->auto_part = false;
212 		bc_vec_npop(&p->flags, p->flags.len - 1);
213 		bc_vec_popAll(&p->exits);
214 		bc_vec_popAll(&p->conds);
215 		bc_vec_popAll(&p->ops);
216 	}
217 #endif // BC_ENABLED
218 
219 	// Reset the program. This might clear the error.
220 	bc_program_reset(p->prog);
221 
222 	// Jump if there is an error.
223 	if (BC_ERR(vm->status)) BC_JMP;
224 }
225 
226 #if BC_DEBUG
227 void
228 bc_parse_free(BcParse* p)
229 {
230 	BC_SIG_ASSERT_LOCKED;
231 
232 	assert(p != NULL);
233 
234 #if BC_ENABLED
235 	if (BC_IS_BC)
236 	{
237 		bc_vec_free(&p->flags);
238 		bc_vec_free(&p->exits);
239 		bc_vec_free(&p->conds);
240 		bc_vec_free(&p->ops);
241 		bc_vec_free(&p->buf);
242 	}
243 #endif // BC_ENABLED
244 
245 	bc_lex_free(&p->l);
246 }
247 #endif // BC_DEBUG
248 
249 void
250 bc_parse_init(BcParse* p, BcProgram* prog, size_t func)
251 {
252 #if BC_ENABLED
253 	uint16_t flag = 0;
254 #endif // BC_ENABLED
255 
256 	BC_SIG_ASSERT_LOCKED;
257 
258 	assert(p != NULL && prog != NULL);
259 
260 #if BC_ENABLED
261 	if (BC_IS_BC)
262 	{
263 		// We always want at least one flag set on the flags stack.
264 		bc_vec_init(&p->flags, sizeof(uint16_t), BC_DTOR_NONE);
265 		bc_vec_push(&p->flags, &flag);
266 
267 		bc_vec_init(&p->exits, sizeof(BcInstPtr), BC_DTOR_NONE);
268 		bc_vec_init(&p->conds, sizeof(size_t), BC_DTOR_NONE);
269 		bc_vec_init(&p->ops, sizeof(BcLexType), BC_DTOR_NONE);
270 		bc_vec_init(&p->buf, sizeof(char), BC_DTOR_NONE);
271 
272 		p->auto_part = false;
273 	}
274 #endif // BC_ENABLED
275 
276 	bc_lex_init(&p->l);
277 
278 	// Set up the function.
279 	p->prog = prog;
280 	bc_parse_updateFunc(p, func);
281 }
282