2 * Copyright (C) 2013-2014 Jo-Philipp Wich <jo@mein.io>
4 * Permission to use, copy, modify, and/or distribute this software for any
5 * purpose with or without fee is hereby granted, provided that the above
6 * copyright notice and this permission notice appear in all copies.
8 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
32 int (*parse)(const char *buf, struct jp_opcode *op, struct jp_state *s);
39 (((x) >= 'a') ? (10 + (x) - 'a') : \
40 (((x) >= 'A') ? (10 + (x) - 'A') : dec(x)))
43 * Stores the given codepoint as a utf8 multibyte sequence into the given
44 * output buffer and substracts the required amount of bytes from the given
47 * Returns false if the multibyte sequence would not fit into the buffer,
52 utf8enc(char **out, int *rem, int code)
54 if (code > 0 && code <= 0x7F)
59 *(*out)++ = code; (*rem)--;
62 else if (code > 0 && code <= 0x7FF)
67 *(*out)++ = ((code >> 6) & 0x1F) | 0xC0; (*rem)--;
68 *(*out)++ = ( code & 0x3F) | 0x80; (*rem)--;
71 else if (code > 0 && code <= 0xFFFF)
76 *(*out)++ = ((code >> 12) & 0x0F) | 0xE0; (*rem)--;
77 *(*out)++ = ((code >> 6) & 0x3F) | 0x80; (*rem)--;
78 *(*out)++ = ( code & 0x3F) | 0x80; (*rem)--;
81 else if (code > 0 && code <= 0x10FFFF)
86 *(*out)++ = ((code >> 18) & 0x07) | 0xF0; (*rem)--;
87 *(*out)++ = ((code >> 12) & 0x3F) | 0x80; (*rem)--;
88 *(*out)++ = ((code >> 6) & 0x3F) | 0x80; (*rem)--;
89 *(*out)++ = ( code & 0x3F) | 0x80; (*rem)--;
98 * Parses a string literal from the given buffer.
100 * Returns a negative value on error, otherwise the amount of consumed
101 * characters from the given buffer.
104 * -1 Unterminated string
105 * -2 Invalid escape sequence
106 * -3 String literal too long
110 parse_string(const char *buf, struct jp_opcode *op, struct jp_state *s)
113 char str[128] = { 0 };
115 const char *in = buf;
117 int rem = sizeof(str) - 1;
122 /* continuation of escape sequence */
128 if (isxdigit(in[1]) && isxdigit(in[2]) &&
129 isxdigit(in[3]) && isxdigit(in[4]))
131 if (!utf8enc(&out, &rem,
132 hex(in[1]) * 16 * 16 * 16 +
133 hex(in[2]) * 16 * 16 +
137 s->error_pos = s->off + (in - buf);
145 s->error_pos = s->off + (in - buf);
151 else if (in[0] == 'x')
153 if (isxdigit(in[1]) && isxdigit(in[2]))
155 if (!utf8enc(&out, &rem, hex(in[1]) * 16 + hex(in[2])))
157 s->error_pos = s->off + (in - buf);
165 s->error_pos = s->off + (in - buf);
170 /* \377, \77 or \7 */
171 else if (in[0] >= '0' && in[0] <= '7')
174 if (in[1] >= '0' && in[1] <= '7' &&
175 in[2] >= '0' && in[2] <= '7')
177 code = dec(in[0]) * 8 * 8 +
183 s->error_pos = s->off + (in - buf);
187 if (!utf8enc(&out, &rem, code))
189 s->error_pos = s->off + (in - buf);
197 else if (in[1] >= '0' && in[1] <= '7')
199 if (!utf8enc(&out, &rem, dec(in[0]) * 8 + dec(in[1])))
201 s->error_pos = s->off + (in - buf);
211 if (!utf8enc(&out, &rem, dec(in[0])))
213 s->error_pos = s->off + (in - buf);
221 /* single character escape */
226 s->error_pos = s->off + (in - buf);
232 case 'a': *out = '\a'; break;
233 case 'b': *out = '\b'; break;
234 case 'e': *out = '\e'; break;
235 case 'f': *out = '\f'; break;
236 case 'n': *out = '\n'; break;
237 case 'r': *out = '\r'; break;
238 case 't': *out = '\t'; break;
239 case 'v': *out = '\v'; break;
241 /* in regexp mode, retain backslash */
246 s->error_pos = s->off + (in - buf);
264 /* begin of escape sequence */
265 else if (*in == '\\')
271 /* terminating quote */
274 op->str = strdup(str);
275 return (in - buf) + 2;
283 s->error_pos = s->off + (in - buf);
296 * Parses a regexp literal from the given buffer.
298 * Returns a negative value on error, otherwise the amount of consumed
299 * characters from the given buffer.
302 * -1 Unterminated regexp
303 * -2 Invalid escape sequence
304 * -3 Regexp literal too long
308 parse_regexp(const char *buf, struct jp_opcode *op, struct jp_state *s)
310 int len = parse_string(buf, op, s);
315 op->num = REG_NOSUB | REG_NEWLINE;
317 for (p = buf + len; p; p++)
322 op->num |= REG_EXTENDED;
327 op->num |= REG_ICASE;
332 op->num &= ~REG_NEWLINE;
348 * Parses a label from the given buffer.
350 * Returns a negative value on error, otherwise the amount of consumed
351 * characters from the given buffer.
358 parse_label(const char *buf, struct jp_opcode *op, struct jp_state *s)
360 char str[128] = { 0 };
362 const char *in = buf;
363 int rem = sizeof(str) - 1;
365 while (*in == '_' || isalnum(*in))
369 s->error_pos = s->off + (in - buf);
376 if (!strcmp(str, "true") || !strcmp(str, "false"))
378 op->num = (str[0] == 't');
383 op->str = strdup(str);
391 * Parses a number literal from the given buffer.
393 * Returns a negative value on error, otherwise the amount of consumed
394 * characters from the given buffer.
397 * -2 Invalid number character
401 parse_number(const char *buf, struct jp_opcode *op, struct jp_state *s)
404 int n = strtol(buf, &e, 10);
408 s->error_pos = s->off;
417 static const struct token tokens[] = {
427 { T_BROPEN, "[", 1 },
428 { T_BRCLOSE, "]", 1 },
430 { T_PCLOSE, ")", 1 },
439 { T_WILDCARD, "*", 1 },
440 { T_REGEXP, "/", 1, parse_regexp },
441 { T_STRING, "'", 1, parse_string },
442 { T_STRING, "\"", 1, parse_string },
443 { T_LABEL, "_", 1, parse_label },
444 { T_LABEL, "az", 0, parse_label },
445 { T_LABEL, "AZ", 0, parse_label },
446 { T_NUMBER, "-", 1, parse_number },
447 { T_NUMBER, "09", 0, parse_number },
450 const char *tokennames[25] = {
467 [T_WILDCARD] = "'*'",
468 [T_REGEXP] = "/.../",
472 [T_NUMBER] = "Number",
473 [T_STRING] = "String",
480 match_token(const char *ptr, struct jp_opcode *op, struct jp_state *s)
483 const struct token *tok;
485 for (i = 0, tok = &tokens[0];
486 i < sizeof(tokens) / sizeof(tokens[0]);
487 i++, tok = &tokens[i])
489 if ((tok->plen > 0 && !strncmp(ptr, tok->pat, tok->plen)) ||
490 (tok->plen == 0 && *ptr >= tok->pat[0] && *ptr <= tok->pat[1]))
492 op->type = tok->type;
495 return tok->parse(ptr, op, s);
501 s->error_pos = s->off;
506 jp_get_token(struct jp_state *s, const char *input, int *mlen)
508 struct jp_opcode op = { 0 };
510 *mlen = match_token(input, &op, s);
514 s->error_code = *mlen;
517 else if (op.type == 0)
522 return jp_alloc_op(s, op.type, op.num, op.str, NULL);