author | sheepluva |
Fri, 16 Sep 2016 00:22:05 +0200 | |
changeset 11853 | 6a0d6b00dac8 |
parent 10017 | de822cd3df3a |
permissions | -rw-r--r-- |
2812 | 1 |
/* |
2 |
** $Id: llex.c,v 2.20.1.1 2007/12/27 13:02:25 roberto Exp $ |
|
3 |
** Lexical Analyzer |
|
4 |
** See Copyright Notice in lua.h |
|
5 |
*/ |
|
6 |
||
7 |
||
8 |
#include <ctype.h> |
|
9 |
#include <locale.h> |
|
10 |
#include <string.h> |
|
11 |
||
12 |
#define llex_c |
|
13 |
#define LUA_CORE |
|
14 |
||
15 |
#include "lua.h" |
|
16 |
||
17 |
#include "ldo.h" |
|
18 |
#include "llex.h" |
|
19 |
#include "lobject.h" |
|
20 |
#include "lparser.h" |
|
21 |
#include "lstate.h" |
|
22 |
#include "lstring.h" |
|
23 |
#include "ltable.h" |
|
24 |
#include "lzio.h" |
|
25 |
||
26 |
||
27 |
||
28 |
#define next(ls) (ls->current = zgetc(ls->z)) |
|
29 |
||
30 |
||
31 |
||
32 |
||
10017 | 33 |
#define currIsNewline(ls) (ls->current == '\n' || ls->current == '\r') |
2812 | 34 |
|
35 |
||
36 |
/* ORDER RESERVED */ |
|
37 |
const char *const luaX_tokens [] = { |
|
38 |
"and", "break", "do", "else", "elseif", |
|
39 |
"end", "false", "for", "function", "if", |
|
40 |
"in", "local", "nil", "not", "or", "repeat", |
|
41 |
"return", "then", "true", "until", "while", |
|
42 |
"..", "...", "==", ">=", "<=", "~=", |
|
43 |
"<number>", "<name>", "<string>", "<eof>", |
|
44 |
NULL |
|
45 |
}; |
|
46 |
||
47 |
||
48 |
#define save_and_next(ls) (save(ls, ls->current), next(ls)) |
|
49 |
||
50 |
||
51 |
static void save (LexState *ls, int c) { |
|
52 |
Mbuffer *b = ls->buff; |
|
53 |
if (b->n + 1 > b->buffsize) { |
|
54 |
size_t newsize; |
|
55 |
if (b->buffsize >= MAX_SIZET/2) |
|
56 |
luaX_lexerror(ls, "lexical element too long", 0); |
|
57 |
newsize = b->buffsize * 2; |
|
58 |
luaZ_resizebuffer(ls->L, b, newsize); |
|
59 |
} |
|
60 |
b->buffer[b->n++] = cast(char, c); |
|
61 |
} |
|
62 |
||
63 |
||
64 |
void luaX_init (lua_State *L) { |
|
65 |
int i; |
|
66 |
for (i=0; i<NUM_RESERVED; i++) { |
|
67 |
TString *ts = luaS_new(L, luaX_tokens[i]); |
|
68 |
luaS_fix(ts); /* reserved words are never collected */ |
|
69 |
lua_assert(strlen(luaX_tokens[i])+1 <= TOKEN_LEN); |
|
70 |
ts->tsv.reserved = cast_byte(i+1); /* reserved word */ |
|
71 |
} |
|
72 |
} |
|
73 |
||
74 |
||
75 |
#define MAXSRC 80 |
|
76 |
||
77 |
||
78 |
const char *luaX_token2str (LexState *ls, int token) { |
|
79 |
if (token < FIRST_RESERVED) { |
|
80 |
lua_assert(token == cast(unsigned char, token)); |
|
81 |
return (iscntrl(token)) ? luaO_pushfstring(ls->L, "char(%d)", token) : |
|
82 |
luaO_pushfstring(ls->L, "%c", token); |
|
83 |
} |
|
84 |
else |
|
85 |
return luaX_tokens[token-FIRST_RESERVED]; |
|
86 |
} |
|
87 |
||
88 |
||
89 |
static const char *txtToken (LexState *ls, int token) { |
|
90 |
switch (token) { |
|
91 |
case TK_NAME: |
|
92 |
case TK_STRING: |
|
93 |
case TK_NUMBER: |
|
94 |
save(ls, '\0'); |
|
95 |
return luaZ_buffer(ls->buff); |
|
96 |
default: |
|
97 |
return luaX_token2str(ls, token); |
|
98 |
} |
|
99 |
} |
|
100 |
||
101 |
||
102 |
void luaX_lexerror (LexState *ls, const char *msg, int token) { |
|
103 |
char buff[MAXSRC]; |
|
104 |
luaO_chunkid(buff, getstr(ls->source), MAXSRC); |
|
105 |
msg = luaO_pushfstring(ls->L, "%s:%d: %s", buff, ls->linenumber, msg); |
|
106 |
if (token) |
|
107 |
luaO_pushfstring(ls->L, "%s near " LUA_QS, msg, txtToken(ls, token)); |
|
108 |
luaD_throw(ls->L, LUA_ERRSYNTAX); |
|
109 |
} |
|
110 |
||
111 |
||
112 |
void luaX_syntaxerror (LexState *ls, const char *msg) { |
|
113 |
luaX_lexerror(ls, msg, ls->t.token); |
|
114 |
} |
|
115 |
||
116 |
||
117 |
TString *luaX_newstring (LexState *ls, const char *str, size_t l) { |
|
118 |
lua_State *L = ls->L; |
|
119 |
TString *ts = luaS_newlstr(L, str, l); |
|
120 |
TValue *o = luaH_setstr(L, ls->fs->h, ts); /* entry for `str' */ |
|
121 |
if (ttisnil(o)) |
|
122 |
setbvalue(o, 1); /* make sure `str' will not be collected */ |
|
123 |
return ts; |
|
124 |
} |
|
125 |
||
126 |
||
127 |
static void inclinenumber (LexState *ls) { |
|
128 |
int old = ls->current; |
|
129 |
lua_assert(currIsNewline(ls)); |
|
130 |
next(ls); /* skip `\n' or `\r' */ |
|
131 |
if (currIsNewline(ls) && ls->current != old) |
|
132 |
next(ls); /* skip `\n\r' or `\r\n' */ |
|
133 |
if (++ls->linenumber >= MAX_INT) |
|
134 |
luaX_syntaxerror(ls, "chunk has too many lines"); |
|
135 |
} |
|
136 |
||
137 |
||
138 |
void luaX_setinput (lua_State *L, LexState *ls, ZIO *z, TString *source) { |
|
139 |
ls->decpoint = '.'; |
|
140 |
ls->L = L; |
|
141 |
ls->lookahead.token = TK_EOS; /* no look-ahead token */ |
|
142 |
ls->z = z; |
|
143 |
ls->fs = NULL; |
|
144 |
ls->linenumber = 1; |
|
145 |
ls->lastline = 1; |
|
146 |
ls->source = source; |
|
147 |
luaZ_resizebuffer(ls->L, ls->buff, LUA_MINBUFFER); /* initialize buffer */ |
|
148 |
next(ls); /* read first char */ |
|
149 |
} |
|
150 |
||
151 |
||
152 |
||
153 |
/* |
|
154 |
** ======================================================= |
|
155 |
** LEXICAL ANALYZER |
|
156 |
** ======================================================= |
|
157 |
*/ |
|
158 |
||
159 |
||
160 |
||
161 |
static int check_next (LexState *ls, const char *set) { |
|
162 |
if (!strchr(set, ls->current)) |
|
163 |
return 0; |
|
164 |
save_and_next(ls); |
|
165 |
return 1; |
|
166 |
} |
|
167 |
||
168 |
||
169 |
static void buffreplace (LexState *ls, char from, char to) { |
|
170 |
size_t n = luaZ_bufflen(ls->buff); |
|
171 |
char *p = luaZ_buffer(ls->buff); |
|
172 |
while (n--) |
|
173 |
if (p[n] == from) p[n] = to; |
|
174 |
} |
|
175 |
||
176 |
||
177 |
static void trydecpoint (LexState *ls, SemInfo *seminfo) { |
|
178 |
/* format error: try to update decimal point separator */ |
|
6039
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
179 |
#ifndef ANDROID |
2812 | 180 |
struct lconv *cv = localeconv(); |
6039
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
181 |
#endif |
2812 | 182 |
char old = ls->decpoint; |
6039
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
183 |
#ifdef ANDROID |
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
184 |
ls->decpoint = '.'; |
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
185 |
#else |
2812 | 186 |
ls->decpoint = (cv ? cv->decimal_point[0] : '.'); |
6039
d75329716a02
Removed the need for duplicate libraries, the android build now uses lua, freetype and tremor from misc/
Xeli
parents:
2812
diff
changeset
|
187 |
#endif |
2812 | 188 |
buffreplace(ls, old, ls->decpoint); /* try updated decimal separator */ |
189 |
if (!luaO_str2d(luaZ_buffer(ls->buff), &seminfo->r)) { |
|
190 |
/* format error with correct decimal point: no more options */ |
|
191 |
buffreplace(ls, ls->decpoint, '.'); /* undo change (for error message) */ |
|
192 |
luaX_lexerror(ls, "malformed number", TK_NUMBER); |
|
193 |
} |
|
194 |
} |
|
195 |
||
196 |
||
197 |
/* LUA_NUMBER */ |
|
198 |
static void read_numeral (LexState *ls, SemInfo *seminfo) { |
|
199 |
lua_assert(isdigit(ls->current)); |
|
200 |
do { |
|
201 |
save_and_next(ls); |
|
202 |
} while (isdigit(ls->current) || ls->current == '.'); |
|
203 |
if (check_next(ls, "Ee")) /* `E'? */ |
|
204 |
check_next(ls, "+-"); /* optional exponent sign */ |
|
205 |
while (isalnum(ls->current) || ls->current == '_') |
|
206 |
save_and_next(ls); |
|
207 |
save(ls, '\0'); |
|
208 |
buffreplace(ls, '.', ls->decpoint); /* follow locale for decimal point */ |
|
209 |
if (!luaO_str2d(luaZ_buffer(ls->buff), &seminfo->r)) /* format error? */ |
|
210 |
trydecpoint(ls, seminfo); /* try to update decimal point separator */ |
|
211 |
} |
|
212 |
||
213 |
||
214 |
static int skip_sep (LexState *ls) { |
|
215 |
int count = 0; |
|
216 |
int s = ls->current; |
|
217 |
lua_assert(s == '[' || s == ']'); |
|
218 |
save_and_next(ls); |
|
219 |
while (ls->current == '=') { |
|
220 |
save_and_next(ls); |
|
221 |
count++; |
|
222 |
} |
|
223 |
return (ls->current == s) ? count : (-count) - 1; |
|
224 |
} |
|
225 |
||
226 |
||
227 |
static void read_long_string (LexState *ls, SemInfo *seminfo, int sep) { |
|
228 |
int cont = 0; |
|
229 |
(void)(cont); /* avoid warnings when `cont' is not used */ |
|
230 |
save_and_next(ls); /* skip 2nd `[' */ |
|
231 |
if (currIsNewline(ls)) /* string starts with a newline? */ |
|
232 |
inclinenumber(ls); /* skip it */ |
|
233 |
for (;;) { |
|
234 |
switch (ls->current) { |
|
235 |
case EOZ: |
|
236 |
luaX_lexerror(ls, (seminfo) ? "unfinished long string" : |
|
237 |
"unfinished long comment", TK_EOS); |
|
238 |
break; /* to avoid warnings */ |
|
239 |
#if defined(LUA_COMPAT_LSTR) |
|
240 |
case '[': { |
|
241 |
if (skip_sep(ls) == sep) { |
|
242 |
save_and_next(ls); /* skip 2nd `[' */ |
|
243 |
cont++; |
|
244 |
#if LUA_COMPAT_LSTR == 1 |
|
245 |
if (sep == 0) |
|
246 |
luaX_lexerror(ls, "nesting of [[...]] is deprecated", '['); |
|
247 |
#endif |
|
248 |
} |
|
249 |
break; |
|
250 |
} |
|
251 |
#endif |
|
252 |
case ']': { |
|
253 |
if (skip_sep(ls) == sep) { |
|
254 |
save_and_next(ls); /* skip 2nd `]' */ |
|
255 |
#if defined(LUA_COMPAT_LSTR) && LUA_COMPAT_LSTR == 2 |
|
256 |
cont--; |
|
257 |
if (sep == 0 && cont >= 0) break; |
|
258 |
#endif |
|
259 |
goto endloop; |
|
260 |
} |
|
261 |
break; |
|
262 |
} |
|
263 |
case '\n': |
|
264 |
case '\r': { |
|
265 |
save(ls, '\n'); |
|
266 |
inclinenumber(ls); |
|
267 |
if (!seminfo) luaZ_resetbuffer(ls->buff); /* avoid wasting space */ |
|
268 |
break; |
|
269 |
} |
|
270 |
default: { |
|
271 |
if (seminfo) save_and_next(ls); |
|
272 |
else next(ls); |
|
273 |
} |
|
274 |
} |
|
275 |
} endloop: |
|
276 |
if (seminfo) |
|
277 |
seminfo->ts = luaX_newstring(ls, luaZ_buffer(ls->buff) + (2 + sep), |
|
278 |
luaZ_bufflen(ls->buff) - 2*(2 + sep)); |
|
279 |
} |
|
280 |
||
281 |
||
282 |
static void read_string (LexState *ls, int del, SemInfo *seminfo) { |
|
283 |
save_and_next(ls); |
|
284 |
while (ls->current != del) { |
|
285 |
switch (ls->current) { |
|
286 |
case EOZ: |
|
287 |
luaX_lexerror(ls, "unfinished string", TK_EOS); |
|
288 |
continue; /* to avoid warnings */ |
|
289 |
case '\n': |
|
290 |
case '\r': |
|
291 |
luaX_lexerror(ls, "unfinished string", TK_STRING); |
|
292 |
continue; /* to avoid warnings */ |
|
293 |
case '\\': { |
|
294 |
int c; |
|
295 |
next(ls); /* do not save the `\' */ |
|
296 |
switch (ls->current) { |
|
297 |
case 'a': c = '\a'; break; |
|
298 |
case 'b': c = '\b'; break; |
|
299 |
case 'f': c = '\f'; break; |
|
300 |
case 'n': c = '\n'; break; |
|
301 |
case 'r': c = '\r'; break; |
|
302 |
case 't': c = '\t'; break; |
|
303 |
case 'v': c = '\v'; break; |
|
304 |
case '\n': /* go through */ |
|
305 |
case '\r': save(ls, '\n'); inclinenumber(ls); continue; |
|
306 |
case EOZ: continue; /* will raise an error next loop */ |
|
307 |
default: { |
|
308 |
if (!isdigit(ls->current)) |
|
309 |
save_and_next(ls); /* handles \\, \", \', and \? */ |
|
310 |
else { /* \xxx */ |
|
311 |
int i = 0; |
|
312 |
c = 0; |
|
313 |
do { |
|
314 |
c = 10*c + (ls->current-'0'); |
|
315 |
next(ls); |
|
316 |
} while (++i<3 && isdigit(ls->current)); |
|
317 |
if (c > UCHAR_MAX) |
|
318 |
luaX_lexerror(ls, "escape sequence too large", TK_STRING); |
|
319 |
save(ls, c); |
|
320 |
} |
|
321 |
continue; |
|
322 |
} |
|
323 |
} |
|
324 |
save(ls, c); |
|
325 |
next(ls); |
|
326 |
continue; |
|
327 |
} |
|
328 |
default: |
|
329 |
save_and_next(ls); |
|
330 |
} |
|
331 |
} |
|
332 |
save_and_next(ls); /* skip delimiter */ |
|
333 |
seminfo->ts = luaX_newstring(ls, luaZ_buffer(ls->buff) + 1, |
|
334 |
luaZ_bufflen(ls->buff) - 2); |
|
335 |
} |
|
336 |
||
337 |
||
338 |
static int llex (LexState *ls, SemInfo *seminfo) { |
|
339 |
luaZ_resetbuffer(ls->buff); |
|
340 |
for (;;) { |
|
341 |
switch (ls->current) { |
|
342 |
case '\n': |
|
343 |
case '\r': { |
|
344 |
inclinenumber(ls); |
|
345 |
continue; |
|
346 |
} |
|
347 |
case '-': { |
|
348 |
next(ls); |
|
349 |
if (ls->current != '-') return '-'; |
|
350 |
/* else is a comment */ |
|
351 |
next(ls); |
|
352 |
if (ls->current == '[') { |
|
353 |
int sep = skip_sep(ls); |
|
354 |
luaZ_resetbuffer(ls->buff); /* `skip_sep' may dirty the buffer */ |
|
355 |
if (sep >= 0) { |
|
356 |
read_long_string(ls, NULL, sep); /* long comment */ |
|
357 |
luaZ_resetbuffer(ls->buff); |
|
358 |
continue; |
|
359 |
} |
|
360 |
} |
|
361 |
/* else short comment */ |
|
362 |
while (!currIsNewline(ls) && ls->current != EOZ) |
|
363 |
next(ls); |
|
364 |
continue; |
|
365 |
} |
|
366 |
case '[': { |
|
367 |
int sep = skip_sep(ls); |
|
368 |
if (sep >= 0) { |
|
369 |
read_long_string(ls, seminfo, sep); |
|
370 |
return TK_STRING; |
|
371 |
} |
|
372 |
else if (sep == -1) return '['; |
|
373 |
else luaX_lexerror(ls, "invalid long string delimiter", TK_STRING); |
|
374 |
} |
|
375 |
case '=': { |
|
376 |
next(ls); |
|
377 |
if (ls->current != '=') return '='; |
|
378 |
else { next(ls); return TK_EQ; } |
|
379 |
} |
|
380 |
case '<': { |
|
381 |
next(ls); |
|
382 |
if (ls->current != '=') return '<'; |
|
383 |
else { next(ls); return TK_LE; } |
|
384 |
} |
|
385 |
case '>': { |
|
386 |
next(ls); |
|
387 |
if (ls->current != '=') return '>'; |
|
388 |
else { next(ls); return TK_GE; } |
|
389 |
} |
|
390 |
case '~': { |
|
391 |
next(ls); |
|
392 |
if (ls->current != '=') return '~'; |
|
393 |
else { next(ls); return TK_NE; } |
|
394 |
} |
|
395 |
case '"': |
|
396 |
case '\'': { |
|
397 |
read_string(ls, ls->current, seminfo); |
|
398 |
return TK_STRING; |
|
399 |
} |
|
400 |
case '.': { |
|
401 |
save_and_next(ls); |
|
402 |
if (check_next(ls, ".")) { |
|
403 |
if (check_next(ls, ".")) |
|
404 |
return TK_DOTS; /* ... */ |
|
405 |
else return TK_CONCAT; /* .. */ |
|
406 |
} |
|
407 |
else if (!isdigit(ls->current)) return '.'; |
|
408 |
else { |
|
409 |
read_numeral(ls, seminfo); |
|
410 |
return TK_NUMBER; |
|
411 |
} |
|
412 |
} |
|
413 |
case EOZ: { |
|
414 |
return TK_EOS; |
|
415 |
} |
|
416 |
default: { |
|
417 |
if (isspace(ls->current)) { |
|
418 |
lua_assert(!currIsNewline(ls)); |
|
419 |
next(ls); |
|
420 |
continue; |
|
421 |
} |
|
422 |
else if (isdigit(ls->current)) { |
|
423 |
read_numeral(ls, seminfo); |
|
424 |
return TK_NUMBER; |
|
425 |
} |
|
426 |
else if (isalpha(ls->current) || ls->current == '_') { |
|
427 |
/* identifier or reserved word */ |
|
428 |
TString *ts; |
|
429 |
do { |
|
430 |
save_and_next(ls); |
|
431 |
} while (isalnum(ls->current) || ls->current == '_'); |
|
432 |
ts = luaX_newstring(ls, luaZ_buffer(ls->buff), |
|
433 |
luaZ_bufflen(ls->buff)); |
|
434 |
if (ts->tsv.reserved > 0) /* reserved word? */ |
|
435 |
return ts->tsv.reserved - 1 + FIRST_RESERVED; |
|
436 |
else { |
|
437 |
seminfo->ts = ts; |
|
438 |
return TK_NAME; |
|
439 |
} |
|
440 |
} |
|
441 |
else { |
|
442 |
int c = ls->current; |
|
443 |
next(ls); |
|
444 |
return c; /* single-char tokens (+ - / ...) */ |
|
445 |
} |
|
446 |
} |
|
447 |
} |
|
448 |
} |
|
449 |
} |
|
450 |
||
451 |
||
452 |
void luaX_next (LexState *ls) { |
|
453 |
ls->lastline = ls->linenumber; |
|
454 |
if (ls->lookahead.token != TK_EOS) { /* is there a look-ahead token? */ |
|
455 |
ls->t = ls->lookahead; /* use this one */ |
|
456 |
ls->lookahead.token = TK_EOS; /* and discharge it */ |
|
457 |
} |
|
458 |
else |
|
459 |
ls->t.token = llex(ls, &ls->t.seminfo); /* read next token */ |
|
460 |
} |
|
461 |
||
462 |
||
463 |
void luaX_lookahead (LexState *ls) { |
|
464 |
lua_assert(ls->lookahead.token == TK_EOS); |
|
465 |
ls->lookahead.token = llex(ls, &ls->lookahead.seminfo); |
|
466 |
} |
|
467 |