aboutsummaryrefslogtreecommitdiff
path: root/src/lj_strfmt.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/lj_strfmt.c')
-rw-r--r--src/lj_strfmt.c549
1 files changed, 549 insertions, 0 deletions
diff --git a/src/lj_strfmt.c b/src/lj_strfmt.c
new file mode 100644
index 00000000..0003de70
--- /dev/null
+++ b/src/lj_strfmt.c
@@ -0,0 +1,549 @@
1/*
2** String formatting.
3** Copyright (C) 2005-2013 Mike Pall. See Copyright Notice in luajit.h
4*/
5
6#include <stdio.h>
7
8#define lj_strfmt_c
9#define LUA_CORE
10
11#include "lj_obj.h"
12#include "lj_buf.h"
13#include "lj_str.h"
14#include "lj_state.h"
15#include "lj_char.h"
16#include "lj_strfmt.h"
17
18/* -- Format parser ------------------------------------------------------- */
19
20static const uint8_t strfmt_map[('x'-'A')+1] = {
21 STRFMT_A,0,0,0,STRFMT_E,0,STRFMT_G,0,0,0,0,0,0,
22 0,0,0,0,0,0,0,0,0,0,STRFMT_X,0,0,
23 0,0,0,0,0,0,
24 STRFMT_A,0,STRFMT_C,STRFMT_D,STRFMT_E,STRFMT_F,STRFMT_G,0,STRFMT_I,0,0,0,0,
25 0,STRFMT_O,STRFMT_P,STRFMT_Q,0,STRFMT_S,0,STRFMT_U,0,0,STRFMT_X
26};
27
28SFormat LJ_FASTCALL lj_strfmt_parse(FormatState *fs)
29{
30 const uint8_t *p = fs->p, *e = fs->e;
31 fs->str = (const char *)p;
32 for (; p < e; p++) {
33 if (*p == '%') { /* Escape char? */
34 if (p[1] == '%') { /* '%%'? */
35 fs->p = ++p+1;
36 goto retlit;
37 } else {
38 SFormat sf = 0;
39 uint32_t c;
40 if (p != (const uint8_t *)fs->str)
41 break;
42 for (p++; (uint32_t)*p - ' ' <= (uint32_t)('0' - ' '); p++) {
43 /* Parse flags. */
44 if (*p == '-') sf |= STRFMT_F_LEFT;
45 else if (*p == '+') sf |= STRFMT_F_PLUS;
46 else if (*p == '0') sf |= STRFMT_F_ZERO;
47 else if (*p == ' ') sf |= STRFMT_F_SPACE;
48 else if (*p == '#') sf |= STRFMT_F_ALT;
49 else break;
50 }
51 if ((uint32_t)*p - '0' < 10) { /* Parse width. */
52 uint32_t width = (uint32_t)*p++ - '0';
53 if ((uint32_t)*p - '0' < 10)
54 width = (uint32_t)*p++ - '0' + width*10;
55 sf |= (width << STRFMT_SH_WIDTH);
56 }
57 if (*p == '.') { /* Parse precision. */
58 uint32_t prec = 0;
59 p++;
60 if ((uint32_t)*p - '0' < 10) {
61 prec = (uint32_t)*p++ - '0';
62 if ((uint32_t)*p - '0' < 10)
63 prec = (uint32_t)*p++ - '0' + prec*10;
64 }
65 sf |= ((prec+1) << STRFMT_SH_PREC);
66 }
67 /* Parse conversion. */
68 c = (uint32_t)*p - 'A';
69 if (LJ_LIKELY(c <= (uint32_t)('x' - 'A'))) {
70 uint32_t sx = strfmt_map[c];
71 if (sx) {
72 fs->p = p+1;
73 return (sf | sx | ((c & 0x20) ? 0 : STRFMT_F_UPPER));
74 }
75 }
76 /* Return error location. */
77 if (*p >= 32) p++;
78 fs->len = (MSize)(p - (const uint8_t *)fs->str);
79 fs->p = fs->e;
80 return STRFMT_ERR;
81 }
82 }
83 }
84 fs->p = p;
85retlit:
86 fs->len = (MSize)(p - (const uint8_t *)fs->str);
87 return fs->len ? STRFMT_LIT : STRFMT_EOF;
88}
89
90/* -- Raw conversions ----------------------------------------------------- */
91
92/* Write number to bufer. */
93char * LJ_FASTCALL lj_strfmt_wnum(char *p, cTValue *o)
94{
95 if (LJ_LIKELY((o->u32.hi << 1) < 0xffe00000)) { /* Finite? */
96#if __BIONIC__
97 if (tvismzero(o)) { *p++ = '-'; *p++ = '0'; return p; }
98#endif
99 return p + lua_number2str(p, o->n);
100 } else if (((o->u32.hi & 0x000fffff) | o->u32.lo) != 0) {
101 *p++ = 'n'; *p++ = 'a'; *p++ = 'n';
102 } else if ((o->u32.hi & 0x80000000) == 0) {
103 *p++ = 'i'; *p++ = 'n'; *p++ = 'f';
104 } else {
105 *p++ = '-'; *p++ = 'i'; *p++ = 'n'; *p++ = 'f';
106 }
107 return p;
108}
109
110#define WINT_R(x, sh, sc) \
111 { uint32_t d = (x*(((1<<sh)+sc-1)/sc))>>sh; x -= d*sc; *p++ = (char)('0'+d); }
112
113/* Write integer to buffer. */
114char * LJ_FASTCALL lj_strfmt_wint(char *p, int32_t k)
115{
116 uint32_t u = (uint32_t)k;
117 if (k < 0) { u = (uint32_t)-k; *p++ = '-'; }
118 if (u < 10000) {
119 if (u < 10) goto dig1; if (u < 100) goto dig2; if (u < 1000) goto dig3;
120 } else {
121 uint32_t v = u / 10000; u -= v * 10000;
122 if (v < 10000) {
123 if (v < 10) goto dig5; if (v < 100) goto dig6; if (v < 1000) goto dig7;
124 } else {
125 uint32_t w = v / 10000; v -= w * 10000;
126 if (w >= 10) WINT_R(w, 10, 10)
127 *p++ = (char)('0'+w);
128 }
129 WINT_R(v, 23, 1000)
130 dig7: WINT_R(v, 12, 100)
131 dig6: WINT_R(v, 10, 10)
132 dig5: *p++ = (char)('0'+v);
133 }
134 WINT_R(u, 23, 1000)
135 dig3: WINT_R(u, 12, 100)
136 dig2: WINT_R(u, 10, 10)
137 dig1: *p++ = (char)('0'+u);
138 return p;
139}
140#undef WINT_R
141
142/* Write pointer to buffer. */
143char * LJ_FASTCALL lj_strfmt_wptr(char *p, const void *v)
144{
145 ptrdiff_t x = (ptrdiff_t)v;
146 MSize i, n = STRFMT_MAXBUF_PTR;
147 if (x == 0) {
148 *p++ = 'N'; *p++ = 'U'; *p++ = 'L'; *p++ = 'L';
149 return p;
150 }
151#if LJ_64
152 /* Shorten output for 64 bit pointers. */
153 n = 2+2*4+((x >> 32) ? 2+2*(lj_fls((uint32_t)(x >> 32))>>3) : 0);
154#endif
155 p[0] = '0';
156 p[1] = 'x';
157 for (i = n-1; i >= 2; i--, x >>= 4)
158 p[i] = "0123456789abcdef"[(x & 15)];
159 return p+n;
160}
161
162/* Write ULEB128 to buffer. */
163char * LJ_FASTCALL lj_strfmt_wuleb128(char *p, uint32_t v)
164{
165 for (; v >= 0x80; v >>= 7)
166 *p++ = (char)((v & 0x7f) | 0x80);
167 *p++ = (char)v;
168 return p;
169}
170
171/* Return string or write number to buffer and return pointer to start. */
172const char *lj_strfmt_wstrnum(char *buf, cTValue *o, MSize *lenp)
173{
174 if (tvisstr(o)) {
175 *lenp = strV(o)->len;
176 return strVdata(o);
177 } else if (tvisint(o)) {
178 *lenp = (MSize)(lj_strfmt_wint(buf, intV(o)) - buf);
179 return buf;
180 } else if (tvisnum(o)) {
181 *lenp = (MSize)(lj_strfmt_wnum(buf, o) - buf);
182 return buf;
183 } else {
184 return NULL;
185 }
186}
187
188/* -- Unformatted conversions to buffer ----------------------------------- */
189
190/* Add integer to buffer. */
191SBuf * LJ_FASTCALL lj_strfmt_putint(SBuf *sb, int32_t k)
192{
193 setsbufP(sb, lj_strfmt_wint(lj_buf_more(sb, STRFMT_MAXBUF_INT), k));
194 return sb;
195}
196
197#if LJ_HASJIT
198/* Add number to buffer. */
199SBuf * LJ_FASTCALL lj_strfmt_putnum(SBuf *sb, cTValue *o)
200{
201 setsbufP(sb, lj_strfmt_wnum(lj_buf_more(sb, STRFMT_MAXBUF_NUM), o));
202 return sb;
203}
204#endif
205
206/* Add quoted string to buffer. */
207SBuf * LJ_FASTCALL lj_strfmt_putquoted(SBuf *sb, GCstr *str)
208{
209 const char *s = strdata(str);
210 MSize len = str->len;
211 lj_buf_putb(sb, '"');
212 while (len--) {
213 uint32_t c = (uint32_t)(uint8_t)*s++;
214 char *p = lj_buf_more(sb, 4);
215 if (c == '"' || c == '\\' || c == '\n') {
216 *p++ = '\\';
217 } else if (lj_char_iscntrl(c)) { /* This can only be 0-31 or 127. */
218 uint32_t d;
219 *p++ = '\\';
220 if (c >= 100 || lj_char_isdigit((uint8_t)*s)) {
221 *p++ = (char)('0'+(c >= 100)); if (c >= 100) c -= 100;
222 goto tens;
223 } else if (c >= 10) {
224 tens:
225 d = (c * 205) >> 11; c -= d * 10; *p++ = (char)('0'+d);
226 }
227 c += '0';
228 }
229 *p++ = (char)c;
230 setsbufP(sb, p);
231 }
232 lj_buf_putb(sb, '"');
233 return sb;
234}
235
236/* -- Formatted conversions to buffer ------------------------------------- */
237
238/* Add formatted char to buffer. */
239SBuf *lj_strfmt_putfchar(SBuf *sb, SFormat sf, int32_t c)
240{
241 MSize width = STRFMT_WIDTH(sf);
242 char *p = lj_buf_more(sb, width > 1 ? width : 1);
243 if ((sf & STRFMT_F_LEFT)) *p++ = (char)c;
244 while (width-- > 1) *p++ = ' ';
245 if (!(sf & STRFMT_F_LEFT)) *p++ = (char)c;
246 setsbufP(sb, p);
247 return sb;
248}
249
250/* Add formatted string to buffer. */
251SBuf *lj_strfmt_putfstr(SBuf *sb, SFormat sf, GCstr *str)
252{
253 MSize len = str->len <= STRFMT_PREC(sf) ? str->len : STRFMT_PREC(sf);
254 MSize width = STRFMT_WIDTH(sf);
255 char *p = lj_buf_more(sb, width > len ? width : len);
256 if ((sf & STRFMT_F_LEFT)) p = lj_buf_wmem(p, strdata(str), len);
257 while (width-- > len) *p++ = ' ';
258 if (!(sf & STRFMT_F_LEFT)) p = lj_buf_wmem(p, strdata(str), len);
259 setsbufP(sb, p);
260 return sb;
261}
262
263/* Add formatted signed/unsigned integer to buffer. */
264SBuf *lj_strfmt_putfxint(SBuf *sb, SFormat sf, uint64_t k)
265{
266 char buf[STRFMT_MAXBUF_XINT], *q = buf + sizeof(buf), *p;
267#ifdef LUA_USE_ASSERT
268 char *ps;
269#endif
270 MSize prefix = 0, len, prec, pprec, width, need;
271
272 /* Figure out signed prefixes. */
273 if (STRFMT_TYPE(sf) == STRFMT_INT) {
274 if ((int64_t)k < 0) {
275 k = (uint64_t)-(int64_t)k;
276 prefix = 256 + '-';
277 } else if ((sf & STRFMT_F_PLUS)) {
278 prefix = 256 + '+';
279 } else if ((sf & STRFMT_F_SPACE)) {
280 prefix = 256 + ' ';
281 }
282 }
283
284 /* Convert number and store to fixed-size buffer in reverse order. */
285 prec = STRFMT_PREC(sf);
286 if ((int32_t)prec >= 0) sf &= ~STRFMT_F_ZERO;
287 if (k == 0) { /* Special-case zero argument. */
288 if (prec != 0 ||
289 (sf & (STRFMT_T_OCT|STRFMT_F_ALT)) == (STRFMT_T_OCT|STRFMT_F_ALT))
290 *--q = '0';
291 } else if (!(sf & (STRFMT_T_HEX|STRFMT_T_OCT))) { /* Decimal. */
292 uint32_t k2;
293 while ((k >> 32)) { *--q = (char)('0' + k % 10); k /= 10; }
294 k2 = (uint32_t)k;
295 do { *--q = (char)('0' + k2 % 10); k2 /= 10; } while (k2);
296 } else if ((sf & STRFMT_T_HEX)) { /* Hex. */
297 const char *hexdig = (sf & STRFMT_F_UPPER) ? "0123456789ABCDEF" :
298 "0123456789abcdef";
299 do { *--q = hexdig[(k & 15)]; k >>= 4; } while (k);
300 if ((sf & STRFMT_F_ALT)) prefix = 512 + 'x';
301 } else { /* Octal. */
302 do { *--q = (char)('0' + (uint32_t)(k & 7)); k >>= 3; } while (k);
303 if ((sf & STRFMT_F_ALT)) *--q = '0';
304 }
305
306 /* Calculate sizes. */
307 len = (MSize)(buf + sizeof(buf) - q);
308 if ((int32_t)len >= (int32_t)prec) prec = len;
309 width = STRFMT_WIDTH(sf);
310 pprec = prec + (prefix >> 8);
311 need = width > pprec ? width : pprec;
312 p = lj_buf_more(sb, need);
313#ifdef LUA_USE_ASSERT
314 ps = p;
315#endif
316
317 /* Format number with leading/trailing whitespace and zeros. */
318 if ((sf & (STRFMT_F_LEFT|STRFMT_F_ZERO)) == 0)
319 while (width-- > pprec) *p++ = ' ';
320 if (prefix) {
321 if ((char)prefix == 'x') *p++ = '0';
322 *p++ = (char)prefix;
323 }
324 if ((sf & (STRFMT_F_LEFT|STRFMT_F_ZERO)) == STRFMT_F_ZERO)
325 while (width-- > pprec) *p++ = '0';
326 while (prec-- > len) *p++ = '0';
327 while (q < buf + sizeof(buf)) *p++ = *q++; /* Add number itself. */
328 if ((sf & STRFMT_F_LEFT))
329 while (width-- > pprec) *p++ = ' ';
330
331 lua_assert(need == (MSize)(p - ps));
332 setsbufP(sb, p);
333 return sb;
334}
335
336/* Add number formatted as signed integer to buffer. */
337SBuf *lj_strfmt_putfnum_int(SBuf *sb, SFormat sf, lua_Number n)
338{
339 int64_t k = (int64_t)n;
340 if (checki32(k) && sf == STRFMT_INT)
341 return lj_strfmt_putint(sb, (int32_t)k); /* Shortcut for plain %d. */
342 else
343 return lj_strfmt_putfxint(sb, sf, (uint64_t)k);
344}
345
346/* Add number formatted as unsigned integer to buffer. */
347SBuf *lj_strfmt_putfnum_uint(SBuf *sb, SFormat sf, lua_Number n)
348{
349 int64_t k;
350 if (n >= 9223372036854775808.0)
351 k = (int64_t)(n - 18446744073709551616.0);
352 else
353 k = (int64_t)n;
354 return lj_strfmt_putfxint(sb, sf, (uint64_t)k);
355}
356
357/* Max. sprintf buffer size needed. At least #string.format("%.99f", -1e308). */
358#define STRFMT_FMTNUMBUF 512
359
360/* Add formatted floating-point number to buffer. */
361SBuf *lj_strfmt_putfnum(SBuf *sb, SFormat sf, lua_Number n)
362{
363 TValue tv;
364 tv.n = n;
365 if (LJ_UNLIKELY((tv.u32.hi << 1) >= 0xffe00000)) {
366 /* Canonicalize output of non-finite values. */
367 MSize width = STRFMT_WIDTH(sf), len = 3;
368 int prefix = 0, ch = (sf & STRFMT_F_UPPER) ? 0x202020 : 0;
369 char *p;
370 if (((tv.u32.hi & 0x000fffff) | tv.u32.lo) != 0) {
371 ch ^= ('n' << 16) | ('a' << 8) | 'n';
372 if ((sf & STRFMT_F_SPACE)) prefix = ' ';
373 } else {
374 ch ^= ('i' << 16) | ('n' << 8) | 'f';
375 if ((tv.u32.hi & 0x80000000)) prefix = '-';
376 else if ((sf & STRFMT_F_PLUS)) prefix = '+';
377 else if ((sf & STRFMT_F_SPACE)) prefix = ' ';
378 }
379 if (prefix) len = 4;
380 p = lj_buf_more(sb, width > len ? width : len);
381 if (!(sf & STRFMT_F_LEFT)) while (width-- > len) *p++ = ' ';
382 if (prefix) *p++ = prefix;
383 *p++ = (char)(ch >> 16); *p++ = (char)(ch >> 8); *p++ = (char)ch;
384 if ((sf & STRFMT_F_LEFT)) while (width-- > len) *p++ = ' ';
385 setsbufP(sb, p);
386 } else { /* Delegate to sprintf() for now. */
387 uint8_t width = (uint8_t)STRFMT_WIDTH(sf), prec = (uint8_t)STRFMT_PREC(sf);
388 char fmt[1+5+2+3+1+1], *p = fmt;
389 *p++ = '%';
390 if ((sf & STRFMT_F_LEFT)) *p++ = '-';
391 if ((sf & STRFMT_F_PLUS)) *p++ = '+';
392 if ((sf & STRFMT_F_ZERO)) *p++ = '0';
393 if ((sf & STRFMT_F_SPACE)) *p++ = ' ';
394 if ((sf & STRFMT_F_ALT)) *p++ = '#';
395 if (width) {
396 uint8_t x = width / 10, y = width % 10;
397 if (x) *p++ = '0' + x;
398 *p++ = '0' + y;
399 }
400 if (prec != 255) {
401 uint8_t x = prec / 10, y = prec % 10;
402 *p++ = '.';
403 if (x) *p++ = '0' + x;
404 *p++ = '0' + y;
405 }
406 *p++ = (0x67666561 >> (STRFMT_FP(sf)<<3)) ^ ((sf & STRFMT_F_UPPER)?0x20:0);
407 *p = '\0';
408 p = lj_buf_more(sb, STRFMT_FMTNUMBUF);
409 setsbufP(sb, p + sprintf(p, fmt, n));
410 }
411 return sb;
412}
413
414/* -- Conversions to strings ---------------------------------------------- */
415
416/* Convert integer to string. */
417GCstr * LJ_FASTCALL lj_strfmt_int(lua_State *L, int32_t k)
418{
419 char buf[STRFMT_MAXBUF_INT];
420 MSize len = (MSize)(lj_strfmt_wint(buf, k) - buf);
421 return lj_str_new(L, buf, len);
422}
423
424/* Convert number to string. */
425GCstr * LJ_FASTCALL lj_strfmt_num(lua_State *L, cTValue *o)
426{
427 char buf[STRFMT_MAXBUF_NUM];
428 MSize len = (MSize)(lj_strfmt_wnum(buf, o) - buf);
429 return lj_str_new(L, buf, len);
430}
431
432/* Convert integer or number to string. */
433GCstr * LJ_FASTCALL lj_strfmt_number(lua_State *L, cTValue *o)
434{
435 return tvisint(o) ? lj_strfmt_int(L, intV(o)) : lj_strfmt_num(L, o);
436}
437
438#if LJ_HASJIT
439/* Convert char value to string. */
440GCstr * LJ_FASTCALL lj_strfmt_char(lua_State *L, int c)
441{
442 char buf[1];
443 buf[0] = c;
444 return lj_str_new(L, buf, 1);
445}
446#endif
447
448/* Raw conversion of object to string. */
449GCstr * LJ_FASTCALL lj_strfmt_obj(lua_State *L, cTValue *o)
450{
451 if (tvisstr(o)) {
452 return strV(o);
453 } else if (tvisnumber(o)) {
454 return lj_strfmt_number(L, o);
455 } else if (tvisnil(o)) {
456 return lj_str_newlit(L, "nil");
457 } else if (tvisfalse(o)) {
458 return lj_str_newlit(L, "false");
459 } else if (tvistrue(o)) {
460 return lj_str_newlit(L, "true");
461 } else {
462 char buf[8+2+2+16], *p = buf;
463 p = lj_buf_wmem(p, lj_typename(o), (MSize)strlen(lj_typename(o)));
464 *p++ = ':'; *p++ = ' ';
465 if (tvisfunc(o) && isffunc(funcV(o))) {
466 p = lj_buf_wmem(p, "builtin#", 8);
467 p = lj_strfmt_wint(p, funcV(o)->c.ffid);
468 } else {
469 p = lj_strfmt_wptr(p, lj_obj_ptr(o));
470 }
471 return lj_str_new(L, buf, (size_t)(p - buf));
472 }
473}
474
475/* -- Internal string formatting ------------------------------------------ */
476
477/*
478** These functions are only used for lua_pushfstring(), lua_pushvfstring()
479** and for internal string formatting (e.g. error messages). Caveat: unlike
480** string.format(), only a limited subset of formats and flags are supported!
481**
482** LuaJIT has support for a couple more formats than Lua 5.1/5.2:
483** - %d %u %o %x with full formatting, 32 bit integers only.
484** - %f and other FP formats are really %.14g.
485** - %s %c %p without formatting.
486*/
487
488/* Push formatted message as a string object to Lua stack. va_list variant. */
489const char *lj_strfmt_pushvf(lua_State *L, const char *fmt, va_list argp)
490{
491 SBuf *sb = lj_buf_tmp_(L);
492 FormatState fs;
493 SFormat sf;
494 GCstr *str;
495 lj_strfmt_init(&fs, fmt, (MSize)strlen(fmt));
496 while ((sf = lj_strfmt_parse(&fs)) != STRFMT_EOF) {
497 switch (STRFMT_TYPE(sf)) {
498 case STRFMT_LIT:
499 lj_buf_putmem(sb, fs.str, fs.len);
500 break;
501 case STRFMT_INT:
502 lj_strfmt_putfxint(sb, sf, va_arg(argp, int32_t));
503 break;
504 case STRFMT_UINT:
505 lj_strfmt_putfxint(sb, sf, va_arg(argp, uint32_t));
506 break;
507 case STRFMT_NUM: {
508 TValue tv;
509 tv.n = va_arg(argp, lua_Number);
510 setsbufP(sb, lj_strfmt_wnum(lj_buf_more(sb, STRFMT_MAXBUF_NUM), &tv));
511 break;
512 }
513 case STRFMT_STR: {
514 const char *s = va_arg(argp, char *);
515 if (s == NULL) s = "(null)";
516 lj_buf_putmem(sb, s, (MSize)strlen(s));
517 break;
518 }
519 case STRFMT_CHAR:
520 lj_buf_putb(sb, va_arg(argp, int));
521 break;
522 case STRFMT_PTR:
523 setsbufP(sb, lj_strfmt_wptr(lj_buf_more(sb, STRFMT_MAXBUF_PTR),
524 va_arg(argp, void *)));
525 break;
526 case STRFMT_ERR:
527 default:
528 lj_buf_putb(sb, '?');
529 lua_assert(0);
530 break;
531 }
532 }
533 str = lj_buf_str(L, sb);
534 setstrV(L, L->top, str);
535 incr_top(L);
536 return strdata(str);
537}
538
539/* Push formatted message as a string object to Lua stack. Vararg variant. */
540const char *lj_strfmt_pushf(lua_State *L, const char *fmt, ...)
541{
542 const char *msg;
543 va_list argp;
544 va_start(argp, fmt);
545 msg = lj_strfmt_pushvf(L, fmt, argp);
546 va_end(argp);
547 return msg;
548}
549