ltable.c 31 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983
  1. /*
  2. ** $Id: ltable.c $
  3. ** Lua tables (hash)
  4. ** See Copyright Notice in lua.h
  5. */
  6. #define ltable_c
  7. #define LUA_CORE
  8. #include "lprefix.h"
  9. /*
  10. ** Implementation of tables (aka arrays, objects, or hash tables).
  11. ** Tables keep its elements in two parts: an array part and a hash part.
  12. ** Non-negative integer keys are all candidates to be kept in the array
  13. ** part. The actual size of the array is the largest 'n' such that
  14. ** more than half the slots between 1 and n are in use.
  15. ** Hash uses a mix of chained scatter table with Brent's variation.
  16. ** A main invariant of these tables is that, if an element is not
  17. ** in its main position (i.e. the 'original' position that its hash gives
  18. ** to it), then the colliding element is in its own main position.
  19. ** Hence even when the load factor reaches 100%, performance remains good.
  20. */
  21. #include <math.h>
  22. #include <limits.h>
  23. #include "lua.h"
  24. #include "ldebug.h"
  25. #include "ldo.h"
  26. #include "lgc.h"
  27. #include "lmem.h"
  28. #include "lobject.h"
  29. #include "lstate.h"
  30. #include "lstring.h"
  31. #include "ltable.h"
  32. #include "lvm.h"
  33. /*
  34. ** MAXABITS is the largest integer such that MAXASIZE fits in an
  35. ** unsigned int.
  36. */
  37. #define MAXABITS cast_int(sizeof(int) * CHAR_BIT - 1)
  38. /*
  39. ** MAXASIZE is the maximum size of the array part. It is the minimum
  40. ** between 2^MAXABITS and the maximum size that, measured in bytes,
  41. ** fits in a 'size_t'.
  42. */
  43. #define MAXASIZE luaM_limitN(1u << MAXABITS, TValue)
  44. /*
  45. ** MAXHBITS is the largest integer such that 2^MAXHBITS fits in a
  46. ** signed int.
  47. */
  48. #define MAXHBITS (MAXABITS - 1)
  49. /*
  50. ** MAXHSIZE is the maximum size of the hash part. It is the minimum
  51. ** between 2^MAXHBITS and the maximum size such that, measured in bytes,
  52. ** it fits in a 'size_t'.
  53. */
  54. #define MAXHSIZE luaM_limitN(1u << MAXHBITS, Node)
  55. /*
  56. ** When the original hash value is good, hashing by a power of 2
  57. ** avoids the cost of '%'.
  58. */
  59. #define hashpow2(t,n) (gnode(t, lmod((n), sizenode(t))))
  60. /*
  61. ** for other types, it is better to avoid modulo by power of 2, as
  62. ** they can have many 2 factors.
  63. */
  64. #define hashmod(t,n) (gnode(t, ((n) % ((sizenode(t)-1)|1))))
  65. #define hashstr(t,str) hashpow2(t, (str)->hash)
  66. #define hashboolean(t,p) hashpow2(t, p)
  67. #define hashpointer(t,p) hashmod(t, point2uint(p))
  68. #define dummynode (&dummynode_)
  69. static const Node dummynode_ = {
  70. {{NULL}, LUA_VEMPTY, /* value's value and type */
  71. LUA_VNIL, 0, {NULL}} /* key type, next, and key value */
  72. };
  73. static const TValue absentkey = {ABSTKEYCONSTANT};
  74. /*
  75. ** Hash for integers. To allow a good hash, use the remainder operator
  76. ** ('%'). If integer fits as a non-negative int, compute an int
  77. ** remainder, which is faster. Otherwise, use an unsigned-integer
  78. ** remainder, which uses all bits and ensures a non-negative result.
  79. */
  80. static Node *hashint (const Table *t, lua_Integer i) {
  81. lua_Unsigned ui = l_castS2U(i);
  82. if (ui <= (unsigned int)INT_MAX)
  83. return hashmod(t, cast_int(ui));
  84. else
  85. return hashmod(t, ui);
  86. }
  87. /*
  88. ** Hash for floating-point numbers.
  89. ** The main computation should be just
  90. ** n = frexp(n, &i); return (n * INT_MAX) + i
  91. ** but there are some numerical subtleties.
  92. ** In a two-complement representation, INT_MAX does not has an exact
  93. ** representation as a float, but INT_MIN does; because the absolute
  94. ** value of 'frexp' is smaller than 1 (unless 'n' is inf/NaN), the
  95. ** absolute value of the product 'frexp * -INT_MIN' is smaller or equal
  96. ** to INT_MAX. Next, the use of 'unsigned int' avoids overflows when
  97. ** adding 'i'; the use of '~u' (instead of '-u') avoids problems with
  98. ** INT_MIN.
  99. */
  100. #if !defined(l_hashfloat)
  101. static int l_hashfloat (lua_Number n) {
  102. int i;
  103. lua_Integer ni;
  104. n = l_mathop(frexp)(n, &i) * -cast_num(INT_MIN);
  105. if (!lua_numbertointeger(n, &ni)) { /* is 'n' inf/-inf/NaN? */
  106. lua_assert(luai_numisnan(n) || l_mathop(fabs)(n) == cast_num(HUGE_VAL));
  107. return 0;
  108. }
  109. else { /* normal case */
  110. unsigned int u = cast_uint(i) + cast_uint(ni);
  111. return cast_int(u <= cast_uint(INT_MAX) ? u : ~u);
  112. }
  113. }
  114. #endif
  115. /*
  116. ** returns the 'main' position of an element in a table (that is,
  117. ** the index of its hash value). The key comes broken (tag in 'ktt'
  118. ** and value in 'vkl') so that we can call it on keys inserted into
  119. ** nodes.
  120. */
  121. static Node *mainposition (const Table *t, int ktt, const Value kvl) {
  122. switch (withvariant(ktt)) {
  123. case LUA_VNUMINT: {
  124. lua_Integer key = ivalueraw(kvl);
  125. return hashint(t, key);
  126. }
  127. case LUA_VNUMFLT: {
  128. lua_Number n = fltvalueraw(kvl);
  129. return hashmod(t, l_hashfloat(n));
  130. }
  131. case LUA_VSHRSTR: {
  132. TString *ts = tsvalueraw(kvl);
  133. return hashstr(t, ts);
  134. }
  135. case LUA_VLNGSTR: {
  136. TString *ts = tsvalueraw(kvl);
  137. return hashpow2(t, luaS_hashlongstr(ts));
  138. }
  139. case LUA_VFALSE:
  140. return hashboolean(t, 0);
  141. case LUA_VTRUE:
  142. return hashboolean(t, 1);
  143. case LUA_VLIGHTUSERDATA: {
  144. void *p = pvalueraw(kvl);
  145. return hashpointer(t, p);
  146. }
  147. case LUA_VLCF: {
  148. lua_CFunction f = fvalueraw(kvl);
  149. return hashpointer(t, f);
  150. }
  151. default: {
  152. GCObject *o = gcvalueraw(kvl);
  153. return hashpointer(t, o);
  154. }
  155. }
  156. }
  157. /*
  158. ** Returns the main position of an element given as a 'TValue'
  159. */
  160. static Node *mainpositionTV (const Table *t, const TValue *key) {
  161. return mainposition(t, rawtt(key), valraw(key));
  162. }
  163. /*
  164. ** Check whether key 'k1' is equal to the key in node 'n2'. This
  165. ** equality is raw, so there are no metamethods. Floats with integer
  166. ** values have been normalized, so integers cannot be equal to
  167. ** floats. It is assumed that 'eqshrstr' is simply pointer equality, so
  168. ** that short strings are handled in the default case.
  169. ** A true 'deadok' means to accept dead keys as equal to their original
  170. ** values. All dead keys are compared in the default case, by pointer
  171. ** identity. (Only collectable objects can produce dead keys.) Note that
  172. ** dead long strings are also compared by identity.
  173. ** Once a key is dead, its corresponding value may be collected, and
  174. ** then another value can be created with the same address. If this
  175. ** other value is given to 'next', 'equalkey' will signal a false
  176. ** positive. In a regular traversal, this situation should never happen,
  177. ** as all keys given to 'next' came from the table itself, and therefore
  178. ** could not have been collected. Outside a regular traversal, we
  179. ** have garbage in, garbage out. What is relevant is that this false
  180. ** positive does not break anything. (In particular, 'next' will return
  181. ** some other valid item on the table or nil.)
  182. */
  183. static int equalkey (const TValue *k1, const Node *n2, int deadok) {
  184. if ((rawtt(k1) != keytt(n2)) && /* not the same variants? */
  185. !(deadok && keyisdead(n2) && iscollectable(k1)))
  186. return 0; /* cannot be same key */
  187. switch (keytt(n2)) {
  188. case LUA_VNIL: case LUA_VFALSE: case LUA_VTRUE:
  189. return 1;
  190. case LUA_VNUMINT:
  191. return (ivalue(k1) == keyival(n2));
  192. case LUA_VNUMFLT:
  193. return luai_numeq(fltvalue(k1), fltvalueraw(keyval(n2)));
  194. case LUA_VLIGHTUSERDATA:
  195. return pvalue(k1) == pvalueraw(keyval(n2));
  196. case LUA_VLCF:
  197. return fvalue(k1) == fvalueraw(keyval(n2));
  198. case ctb(LUA_VLNGSTR):
  199. return luaS_eqlngstr(tsvalue(k1), keystrval(n2));
  200. default:
  201. return gcvalue(k1) == gcvalueraw(keyval(n2));
  202. }
  203. }
  204. /*
  205. ** True if value of 'alimit' is equal to the real size of the array
  206. ** part of table 't'. (Otherwise, the array part must be larger than
  207. ** 'alimit'.)
  208. */
  209. #define limitequalsasize(t) (isrealasize(t) || ispow2((t)->alimit))
  210. /*
  211. ** Returns the real size of the 'array' array
  212. */
  213. LUAI_FUNC unsigned int luaH_realasize (const Table *t) {
  214. if (limitequalsasize(t))
  215. return t->alimit; /* this is the size */
  216. else {
  217. unsigned int size = t->alimit;
  218. /* compute the smallest power of 2 not smaller than 'n' */
  219. size |= (size >> 1);
  220. size |= (size >> 2);
  221. size |= (size >> 4);
  222. size |= (size >> 8);
  223. size |= (size >> 16);
  224. #if (UINT_MAX >> 30) > 3
  225. size |= (size >> 32); /* unsigned int has more than 32 bits */
  226. #endif
  227. size++;
  228. lua_assert(ispow2(size) && size/2 < t->alimit && t->alimit < size);
  229. return size;
  230. }
  231. }
  232. /*
  233. ** Check whether real size of the array is a power of 2.
  234. ** (If it is not, 'alimit' cannot be changed to any other value
  235. ** without changing the real size.)
  236. */
  237. static int ispow2realasize (const Table *t) {
  238. return (!isrealasize(t) || ispow2(t->alimit));
  239. }
  240. static unsigned int setlimittosize (Table *t) {
  241. t->alimit = luaH_realasize(t);
  242. setrealasize(t);
  243. return t->alimit;
  244. }
  245. #define limitasasize(t) check_exp(isrealasize(t), t->alimit)
  246. /*
  247. ** "Generic" get version. (Not that generic: not valid for integers,
  248. ** which may be in array part, nor for floats with integral values.)
  249. ** See explanation about 'deadok' in function 'equalkey'.
  250. */
  251. static const TValue *getgeneric (Table *t, const TValue *key, int deadok) {
  252. Node *n = mainpositionTV(t, key);
  253. for (;;) { /* check whether 'key' is somewhere in the chain */
  254. if (equalkey(key, n, deadok))
  255. return gval(n); /* that's it */
  256. else {
  257. int nx = gnext(n);
  258. if (nx == 0)
  259. return &absentkey; /* not found */
  260. n += nx;
  261. }
  262. }
  263. }
  264. /*
  265. ** returns the index for 'k' if 'k' is an appropriate key to live in
  266. ** the array part of a table, 0 otherwise.
  267. */
  268. static unsigned int arrayindex (lua_Integer k) {
  269. if (l_castS2U(k) - 1u < MAXASIZE) /* 'k' in [1, MAXASIZE]? */
  270. return cast_uint(k); /* 'key' is an appropriate array index */
  271. else
  272. return 0;
  273. }
  274. /*
  275. ** returns the index of a 'key' for table traversals. First goes all
  276. ** elements in the array part, then elements in the hash part. The
  277. ** beginning of a traversal is signaled by 0.
  278. */
  279. static unsigned int findindex (lua_State *L, Table *t, TValue *key,
  280. unsigned int asize) {
  281. unsigned int i;
  282. if (ttisnil(key)) return 0; /* first iteration */
  283. i = ttisinteger(key) ? arrayindex(ivalue(key)) : 0;
  284. if (i - 1u < asize) /* is 'key' inside array part? */
  285. return i; /* yes; that's the index */
  286. else {
  287. const TValue *n = getgeneric(t, key, 1);
  288. if (l_unlikely(isabstkey(n)))
  289. luaG_runerror(L, "invalid key to 'next'"); /* key not found */
  290. i = cast_int(nodefromval(n) - gnode(t, 0)); /* key index in hash table */
  291. /* hash elements are numbered after array ones */
  292. return (i + 1) + asize;
  293. }
  294. }
  295. int luaH_next (lua_State *L, Table *t, StkId key) {
  296. unsigned int asize = luaH_realasize(t);
  297. unsigned int i = findindex(L, t, s2v(key), asize); /* find original key */
  298. for (; i < asize; i++) { /* try first array part */
  299. if (!isempty(&t->array[i])) { /* a non-empty entry? */
  300. setivalue(s2v(key), i + 1);
  301. setobj2s(L, key + 1, &t->array[i]);
  302. return 1;
  303. }
  304. }
  305. for (i -= asize; cast_int(i) < sizenode(t); i++) { /* hash part */
  306. if (!isempty(gval(gnode(t, i)))) { /* a non-empty entry? */
  307. Node *n = gnode(t, i);
  308. getnodekey(L, s2v(key), n);
  309. setobj2s(L, key + 1, gval(n));
  310. return 1;
  311. }
  312. }
  313. return 0; /* no more elements */
  314. }
  315. static void freehash (lua_State *L, Table *t) {
  316. if (!isdummy(t))
  317. luaM_freearray(L, t->node, cast_sizet(sizenode(t)));
  318. }
  319. /*
  320. ** {=============================================================
  321. ** Rehash
  322. ** ==============================================================
  323. */
  324. /*
  325. ** Compute the optimal size for the array part of table 't'. 'nums' is a
  326. ** "count array" where 'nums[i]' is the number of integers in the table
  327. ** between 2^(i - 1) + 1 and 2^i. 'pna' enters with the total number of
  328. ** integer keys in the table and leaves with the number of keys that
  329. ** will go to the array part; return the optimal size. (The condition
  330. ** 'twotoi > 0' in the for loop stops the loop if 'twotoi' overflows.)
  331. */
  332. static unsigned int computesizes (unsigned int nums[], unsigned int *pna) {
  333. int i;
  334. unsigned int twotoi; /* 2^i (candidate for optimal size) */
  335. unsigned int a = 0; /* number of elements smaller than 2^i */
  336. unsigned int na = 0; /* number of elements to go to array part */
  337. unsigned int optimal = 0; /* optimal size for array part */
  338. /* loop while keys can fill more than half of total size */
  339. for (i = 0, twotoi = 1;
  340. twotoi > 0 && *pna > twotoi / 2;
  341. i++, twotoi *= 2) {
  342. a += nums[i];
  343. if (a > twotoi/2) { /* more than half elements present? */
  344. optimal = twotoi; /* optimal size (till now) */
  345. na = a; /* all elements up to 'optimal' will go to array part */
  346. }
  347. }
  348. lua_assert((optimal == 0 || optimal / 2 < na) && na <= optimal);
  349. *pna = na;
  350. return optimal;
  351. }
  352. static int countint (lua_Integer key, unsigned int *nums) {
  353. unsigned int k = arrayindex(key);
  354. if (k != 0) { /* is 'key' an appropriate array index? */
  355. nums[luaO_ceillog2(k)]++; /* count as such */
  356. return 1;
  357. }
  358. else
  359. return 0;
  360. }
  361. /*
  362. ** Count keys in array part of table 't': Fill 'nums[i]' with
  363. ** number of keys that will go into corresponding slice and return
  364. ** total number of non-nil keys.
  365. */
  366. static unsigned int numusearray (const Table *t, unsigned int *nums) {
  367. int lg;
  368. unsigned int ttlg; /* 2^lg */
  369. unsigned int ause = 0; /* summation of 'nums' */
  370. unsigned int i = 1; /* count to traverse all array keys */
  371. unsigned int asize = limitasasize(t); /* real array size */
  372. /* traverse each slice */
  373. for (lg = 0, ttlg = 1; lg <= MAXABITS; lg++, ttlg *= 2) {
  374. unsigned int lc = 0; /* counter */
  375. unsigned int lim = ttlg;
  376. if (lim > asize) {
  377. lim = asize; /* adjust upper limit */
  378. if (i > lim)
  379. break; /* no more elements to count */
  380. }
  381. /* count elements in range (2^(lg - 1), 2^lg] */
  382. for (; i <= lim; i++) {
  383. if (!isempty(&t->array[i-1]))
  384. lc++;
  385. }
  386. nums[lg] += lc;
  387. ause += lc;
  388. }
  389. return ause;
  390. }
  391. static int numusehash (const Table *t, unsigned int *nums, unsigned int *pna) {
  392. int totaluse = 0; /* total number of elements */
  393. int ause = 0; /* elements added to 'nums' (can go to array part) */
  394. int i = sizenode(t);
  395. while (i--) {
  396. Node *n = &t->node[i];
  397. if (!isempty(gval(n))) {
  398. if (keyisinteger(n))
  399. ause += countint(keyival(n), nums);
  400. totaluse++;
  401. }
  402. }
  403. *pna += ause;
  404. return totaluse;
  405. }
  406. /*
  407. ** Creates an array for the hash part of a table with the given
  408. ** size, or reuses the dummy node if size is zero.
  409. ** The computation for size overflow is in two steps: the first
  410. ** comparison ensures that the shift in the second one does not
  411. ** overflow.
  412. */
  413. static void setnodevector (lua_State *L, Table *t, unsigned int size) {
  414. if (size == 0) { /* no elements to hash part? */
  415. t->node = cast(Node *, dummynode); /* use common 'dummynode' */
  416. t->lsizenode = 0;
  417. t->lastfree = NULL; /* signal that it is using dummy node */
  418. }
  419. else {
  420. int i;
  421. int lsize = luaO_ceillog2(size);
  422. if (lsize > MAXHBITS || (1u << lsize) > MAXHSIZE)
  423. luaG_runerror(L, "table overflow");
  424. size = twoto(lsize);
  425. t->node = luaM_newvector(L, size, Node);
  426. for (i = 0; i < (int)size; i++) {
  427. Node *n = gnode(t, i);
  428. gnext(n) = 0;
  429. setnilkey(n);
  430. setempty(gval(n));
  431. }
  432. t->lsizenode = cast_byte(lsize);
  433. t->lastfree = gnode(t, size); /* all positions are free */
  434. }
  435. }
  436. /*
  437. ** (Re)insert all elements from the hash part of 'ot' into table 't'.
  438. */
  439. static void reinsert (lua_State *L, Table *ot, Table *t) {
  440. int j;
  441. int size = sizenode(ot);
  442. for (j = 0; j < size; j++) {
  443. Node *old = gnode(ot, j);
  444. if (!isempty(gval(old))) {
  445. /* doesn't need barrier/invalidate cache, as entry was
  446. already present in the table */
  447. TValue k;
  448. getnodekey(L, &k, old);
  449. luaH_set(L, t, &k, gval(old));
  450. }
  451. }
  452. }
  453. /*
  454. ** Exchange the hash part of 't1' and 't2'.
  455. */
  456. static void exchangehashpart (Table *t1, Table *t2) {
  457. lu_byte lsizenode = t1->lsizenode;
  458. Node *node = t1->node;
  459. Node *lastfree = t1->lastfree;
  460. t1->lsizenode = t2->lsizenode;
  461. t1->node = t2->node;
  462. t1->lastfree = t2->lastfree;
  463. t2->lsizenode = lsizenode;
  464. t2->node = node;
  465. t2->lastfree = lastfree;
  466. }
  467. /*
  468. ** Resize table 't' for the new given sizes. Both allocations (for
  469. ** the hash part and for the array part) can fail, which creates some
  470. ** subtleties. If the first allocation, for the hash part, fails, an
  471. ** error is raised and that is it. Otherwise, it copies the elements from
  472. ** the shrinking part of the array (if it is shrinking) into the new
  473. ** hash. Then it reallocates the array part. If that fails, the table
  474. ** is in its original state; the function frees the new hash part and then
  475. ** raises the allocation error. Otherwise, it sets the new hash part
  476. ** into the table, initializes the new part of the array (if any) with
  477. ** nils and reinserts the elements of the old hash back into the new
  478. ** parts of the table.
  479. */
  480. void luaH_resize (lua_State *L, Table *t, unsigned int newasize,
  481. unsigned int nhsize) {
  482. unsigned int i;
  483. Table newt; /* to keep the new hash part */
  484. unsigned int oldasize = setlimittosize(t);
  485. TValue *newarray;
  486. /* create new hash part with appropriate size into 'newt' */
  487. setnodevector(L, &newt, nhsize);
  488. if (newasize < oldasize) { /* will array shrink? */
  489. t->alimit = newasize; /* pretend array has new size... */
  490. exchangehashpart(t, &newt); /* and new hash */
  491. /* re-insert into the new hash the elements from vanishing slice */
  492. for (i = newasize; i < oldasize; i++) {
  493. if (!isempty(&t->array[i]))
  494. luaH_setint(L, t, i + 1, &t->array[i]);
  495. }
  496. t->alimit = oldasize; /* restore current size... */
  497. exchangehashpart(t, &newt); /* and hash (in case of errors) */
  498. }
  499. /* allocate new array */
  500. newarray = luaM_reallocvector(L, t->array, oldasize, newasize, TValue);
  501. if (l_unlikely(newarray == NULL && newasize > 0)) { /* allocation failed? */
  502. freehash(L, &newt); /* release new hash part */
  503. luaM_error(L); /* raise error (with array unchanged) */
  504. }
  505. /* allocation ok; initialize new part of the array */
  506. exchangehashpart(t, &newt); /* 't' has the new hash ('newt' has the old) */
  507. t->array = newarray; /* set new array part */
  508. t->alimit = newasize;
  509. for (i = oldasize; i < newasize; i++) /* clear new slice of the array */
  510. setempty(&t->array[i]);
  511. /* re-insert elements from old hash part into new parts */
  512. reinsert(L, &newt, t); /* 'newt' now has the old hash */
  513. freehash(L, &newt); /* free old hash part */
  514. }
  515. void luaH_resizearray (lua_State *L, Table *t, unsigned int nasize) {
  516. int nsize = allocsizenode(t);
  517. luaH_resize(L, t, nasize, nsize);
  518. }
  519. /*
  520. ** nums[i] = number of keys 'k' where 2^(i - 1) < k <= 2^i
  521. */
  522. static void rehash (lua_State *L, Table *t, const TValue *ek) {
  523. unsigned int asize; /* optimal size for array part */
  524. unsigned int na; /* number of keys in the array part */
  525. unsigned int nums[MAXABITS + 1];
  526. int i;
  527. int totaluse;
  528. for (i = 0; i <= MAXABITS; i++) nums[i] = 0; /* reset counts */
  529. setlimittosize(t);
  530. na = numusearray(t, nums); /* count keys in array part */
  531. totaluse = na; /* all those keys are integer keys */
  532. totaluse += numusehash(t, nums, &na); /* count keys in hash part */
  533. /* count extra key */
  534. if (ttisinteger(ek))
  535. na += countint(ivalue(ek), nums);
  536. totaluse++;
  537. /* compute new size for array part */
  538. asize = computesizes(nums, &na);
  539. /* resize the table to new computed sizes */
  540. luaH_resize(L, t, asize, totaluse - na);
  541. }
  542. /*
  543. ** }=============================================================
  544. */
  545. Table *luaH_new (lua_State *L) {
  546. GCObject *o = luaC_newobj(L, LUA_VTABLE, sizeof(Table));
  547. Table *t = gco2t(o);
  548. t->metatable = NULL;
  549. t->flags = cast_byte(maskflags); /* table has no metamethod fields */
  550. t->array = NULL;
  551. t->alimit = 0;
  552. setnodevector(L, t, 0);
  553. return t;
  554. }
  555. void luaH_free (lua_State *L, Table *t) {
  556. freehash(L, t);
  557. luaM_freearray(L, t->array, luaH_realasize(t));
  558. luaM_free(L, t);
  559. }
  560. static Node *getfreepos (Table *t) {
  561. if (!isdummy(t)) {
  562. while (t->lastfree > t->node) {
  563. t->lastfree--;
  564. if (keyisnil(t->lastfree))
  565. return t->lastfree;
  566. }
  567. }
  568. return NULL; /* could not find a free place */
  569. }
  570. /*
  571. ** inserts a new key into a hash table; first, check whether key's main
  572. ** position is free. If not, check whether colliding node is in its main
  573. ** position or not: if it is not, move colliding node to an empty place and
  574. ** put new key in its main position; otherwise (colliding node is in its main
  575. ** position), new key goes to an empty position.
  576. */
  577. void luaH_newkey (lua_State *L, Table *t, const TValue *key, TValue *value) {
  578. Node *mp;
  579. TValue aux;
  580. if (l_unlikely(ttisnil(key)))
  581. luaG_runerror(L, "table index is nil");
  582. else if (ttisfloat(key)) {
  583. lua_Number f = fltvalue(key);
  584. lua_Integer k;
  585. if (luaV_flttointeger(f, &k, F2Ieq)) { /* does key fit in an integer? */
  586. setivalue(&aux, k);
  587. key = &aux; /* insert it as an integer */
  588. }
  589. else if (l_unlikely(luai_numisnan(f)))
  590. luaG_runerror(L, "table index is NaN");
  591. }
  592. if (ttisnil(value))
  593. return; /* do not insert nil values */
  594. mp = mainpositionTV(t, key);
  595. if (!isempty(gval(mp)) || isdummy(t)) { /* main position is taken? */
  596. Node *othern;
  597. Node *f = getfreepos(t); /* get a free place */
  598. if (f == NULL) { /* cannot find a free place? */
  599. rehash(L, t, key); /* grow table */
  600. /* whatever called 'newkey' takes care of TM cache */
  601. luaH_set(L, t, key, value); /* insert key into grown table */
  602. return;
  603. }
  604. lua_assert(!isdummy(t));
  605. othern = mainposition(t, keytt(mp), keyval(mp));
  606. if (othern != mp) { /* is colliding node out of its main position? */
  607. /* yes; move colliding node into free position */
  608. while (othern + gnext(othern) != mp) /* find previous */
  609. othern += gnext(othern);
  610. gnext(othern) = cast_int(f - othern); /* rechain to point to 'f' */
  611. *f = *mp; /* copy colliding node into free pos. (mp->next also goes) */
  612. if (gnext(mp) != 0) {
  613. gnext(f) += cast_int(mp - f); /* correct 'next' */
  614. gnext(mp) = 0; /* now 'mp' is free */
  615. }
  616. setempty(gval(mp));
  617. }
  618. else { /* colliding node is in its own main position */
  619. /* new node will go into free position */
  620. if (gnext(mp) != 0)
  621. gnext(f) = cast_int((mp + gnext(mp)) - f); /* chain new position */
  622. else lua_assert(gnext(f) == 0);
  623. gnext(mp) = cast_int(f - mp);
  624. mp = f;
  625. }
  626. }
  627. setnodekey(L, mp, key);
  628. luaC_barrierback(L, obj2gco(t), key);
  629. lua_assert(isempty(gval(mp)));
  630. setobj2t(L, gval(mp), value);
  631. }
  632. /*
  633. ** Search function for integers. If integer is inside 'alimit', get it
  634. ** directly from the array part. Otherwise, if 'alimit' is not equal to
  635. ** the real size of the array, key still can be in the array part. In
  636. ** this case, try to avoid a call to 'luaH_realasize' when key is just
  637. ** one more than the limit (so that it can be incremented without
  638. ** changing the real size of the array).
  639. */
  640. const TValue *luaH_getint (Table *t, lua_Integer key) {
  641. if (l_castS2U(key) - 1u < t->alimit) /* 'key' in [1, t->alimit]? */
  642. return &t->array[key - 1];
  643. else if (!limitequalsasize(t) && /* key still may be in the array part? */
  644. (l_castS2U(key) == t->alimit + 1 ||
  645. l_castS2U(key) - 1u < luaH_realasize(t))) {
  646. t->alimit = cast_uint(key); /* probably '#t' is here now */
  647. return &t->array[key - 1];
  648. }
  649. else {
  650. Node *n = hashint(t, key);
  651. for (;;) { /* check whether 'key' is somewhere in the chain */
  652. if (keyisinteger(n) && keyival(n) == key)
  653. return gval(n); /* that's it */
  654. else {
  655. int nx = gnext(n);
  656. if (nx == 0) break;
  657. n += nx;
  658. }
  659. }
  660. return &absentkey;
  661. }
  662. }
  663. /*
  664. ** search function for short strings
  665. */
  666. const TValue *luaH_getshortstr (Table *t, TString *key) {
  667. Node *n = hashstr(t, key);
  668. lua_assert(key->tt == LUA_VSHRSTR);
  669. for (;;) { /* check whether 'key' is somewhere in the chain */
  670. if (keyisshrstr(n) && eqshrstr(keystrval(n), key))
  671. return gval(n); /* that's it */
  672. else {
  673. int nx = gnext(n);
  674. if (nx == 0)
  675. return &absentkey; /* not found */
  676. n += nx;
  677. }
  678. }
  679. }
  680. const TValue *luaH_getstr (Table *t, TString *key) {
  681. if (key->tt == LUA_VSHRSTR)
  682. return luaH_getshortstr(t, key);
  683. else { /* for long strings, use generic case */
  684. TValue ko;
  685. setsvalue(cast(lua_State *, NULL), &ko, key);
  686. return getgeneric(t, &ko, 0);
  687. }
  688. }
  689. /*
  690. ** main search function
  691. */
  692. const TValue *luaH_get (Table *t, const TValue *key) {
  693. switch (ttypetag(key)) {
  694. case LUA_VSHRSTR: return luaH_getshortstr(t, tsvalue(key));
  695. case LUA_VNUMINT: return luaH_getint(t, ivalue(key));
  696. case LUA_VNIL: return &absentkey;
  697. case LUA_VNUMFLT: {
  698. lua_Integer k;
  699. if (luaV_flttointeger(fltvalue(key), &k, F2Ieq)) /* integral index? */
  700. return luaH_getint(t, k); /* use specialized version */
  701. /* else... */
  702. } /* FALLTHROUGH */
  703. default:
  704. return getgeneric(t, key, 0);
  705. }
  706. }
  707. /*
  708. ** Finish a raw "set table" operation, where 'slot' is where the value
  709. ** should have been (the result of a previous "get table").
  710. ** Beware: when using this function you probably need to check a GC
  711. ** barrier and invalidate the TM cache.
  712. */
  713. void luaH_finishset (lua_State *L, Table *t, const TValue *key,
  714. const TValue *slot, TValue *value) {
  715. if (isabstkey(slot))
  716. luaH_newkey(L, t, key, value);
  717. else
  718. setobj2t(L, cast(TValue *, slot), value);
  719. }
  720. /*
  721. ** beware: when using this function you probably need to check a GC
  722. ** barrier and invalidate the TM cache.
  723. */
  724. void luaH_set (lua_State *L, Table *t, const TValue *key, TValue *value) {
  725. const TValue *slot = luaH_get(t, key);
  726. luaH_finishset(L, t, key, slot, value);
  727. }
  728. void luaH_setint (lua_State *L, Table *t, lua_Integer key, TValue *value) {
  729. const TValue *p = luaH_getint(t, key);
  730. if (isabstkey(p)) {
  731. TValue k;
  732. setivalue(&k, key);
  733. luaH_newkey(L, t, &k, value);
  734. }
  735. else
  736. setobj2t(L, cast(TValue *, p), value);
  737. }
  738. /*
  739. ** Try to find a boundary in the hash part of table 't'. From the
  740. ** caller, we know that 'j' is zero or present and that 'j + 1' is
  741. ** present. We want to find a larger key that is absent from the
  742. ** table, so that we can do a binary search between the two keys to
  743. ** find a boundary. We keep doubling 'j' until we get an absent index.
  744. ** If the doubling would overflow, we try LUA_MAXINTEGER. If it is
  745. ** absent, we are ready for the binary search. ('j', being max integer,
  746. ** is larger or equal to 'i', but it cannot be equal because it is
  747. ** absent while 'i' is present; so 'j > i'.) Otherwise, 'j' is a
  748. ** boundary. ('j + 1' cannot be a present integer key because it is
  749. ** not a valid integer in Lua.)
  750. */
  751. static lua_Unsigned hash_search (Table *t, lua_Unsigned j) {
  752. lua_Unsigned i;
  753. if (j == 0) j++; /* the caller ensures 'j + 1' is present */
  754. do {
  755. i = j; /* 'i' is a present index */
  756. if (j <= l_castS2U(LUA_MAXINTEGER) / 2)
  757. j *= 2;
  758. else {
  759. j = LUA_MAXINTEGER;
  760. if (isempty(luaH_getint(t, j))) /* t[j] not present? */
  761. break; /* 'j' now is an absent index */
  762. else /* weird case */
  763. return j; /* well, max integer is a boundary... */
  764. }
  765. } while (!isempty(luaH_getint(t, j))); /* repeat until an absent t[j] */
  766. /* i < j && t[i] present && t[j] absent */
  767. while (j - i > 1u) { /* do a binary search between them */
  768. lua_Unsigned m = (i + j) / 2;
  769. if (isempty(luaH_getint(t, m))) j = m;
  770. else i = m;
  771. }
  772. return i;
  773. }
  774. static unsigned int binsearch (const TValue *array, unsigned int i,
  775. unsigned int j) {
  776. while (j - i > 1u) { /* binary search */
  777. unsigned int m = (i + j) / 2;
  778. if (isempty(&array[m - 1])) j = m;
  779. else i = m;
  780. }
  781. return i;
  782. }
  783. /*
  784. ** Try to find a boundary in table 't'. (A 'boundary' is an integer index
  785. ** such that t[i] is present and t[i+1] is absent, or 0 if t[1] is absent
  786. ** and 'maxinteger' if t[maxinteger] is present.)
  787. ** (In the next explanation, we use Lua indices, that is, with base 1.
  788. ** The code itself uses base 0 when indexing the array part of the table.)
  789. ** The code starts with 'limit = t->alimit', a position in the array
  790. ** part that may be a boundary.
  791. **
  792. ** (1) If 't[limit]' is empty, there must be a boundary before it.
  793. ** As a common case (e.g., after 't[#t]=nil'), check whether 'limit-1'
  794. ** is present. If so, it is a boundary. Otherwise, do a binary search
  795. ** between 0 and limit to find a boundary. In both cases, try to
  796. ** use this boundary as the new 'alimit', as a hint for the next call.
  797. **
  798. ** (2) If 't[limit]' is not empty and the array has more elements
  799. ** after 'limit', try to find a boundary there. Again, try first
  800. ** the special case (which should be quite frequent) where 'limit+1'
  801. ** is empty, so that 'limit' is a boundary. Otherwise, check the
  802. ** last element of the array part. If it is empty, there must be a
  803. ** boundary between the old limit (present) and the last element
  804. ** (absent), which is found with a binary search. (This boundary always
  805. ** can be a new limit.)
  806. **
  807. ** (3) The last case is when there are no elements in the array part
  808. ** (limit == 0) or its last element (the new limit) is present.
  809. ** In this case, must check the hash part. If there is no hash part
  810. ** or 'limit+1' is absent, 'limit' is a boundary. Otherwise, call
  811. ** 'hash_search' to find a boundary in the hash part of the table.
  812. ** (In those cases, the boundary is not inside the array part, and
  813. ** therefore cannot be used as a new limit.)
  814. */
  815. lua_Unsigned luaH_getn (Table *t) {
  816. unsigned int limit = t->alimit;
  817. if (limit > 0 && isempty(&t->array[limit - 1])) { /* (1)? */
  818. /* there must be a boundary before 'limit' */
  819. if (limit >= 2 && !isempty(&t->array[limit - 2])) {
  820. /* 'limit - 1' is a boundary; can it be a new limit? */
  821. if (ispow2realasize(t) && !ispow2(limit - 1)) {
  822. t->alimit = limit - 1;
  823. setnorealasize(t); /* now 'alimit' is not the real size */
  824. }
  825. return limit - 1;
  826. }
  827. else { /* must search for a boundary in [0, limit] */
  828. unsigned int boundary = binsearch(t->array, 0, limit);
  829. /* can this boundary represent the real size of the array? */
  830. if (ispow2realasize(t) && boundary > luaH_realasize(t) / 2) {
  831. t->alimit = boundary; /* use it as the new limit */
  832. setnorealasize(t);
  833. }
  834. return boundary;
  835. }
  836. }
  837. /* 'limit' is zero or present in table */
  838. if (!limitequalsasize(t)) { /* (2)? */
  839. /* 'limit' > 0 and array has more elements after 'limit' */
  840. if (isempty(&t->array[limit])) /* 'limit + 1' is empty? */
  841. return limit; /* this is the boundary */
  842. /* else, try last element in the array */
  843. limit = luaH_realasize(t);
  844. if (isempty(&t->array[limit - 1])) { /* empty? */
  845. /* there must be a boundary in the array after old limit,
  846. and it must be a valid new limit */
  847. unsigned int boundary = binsearch(t->array, t->alimit, limit);
  848. t->alimit = boundary;
  849. return boundary;
  850. }
  851. /* else, new limit is present in the table; check the hash part */
  852. }
  853. /* (3) 'limit' is the last element and either is zero or present in table */
  854. lua_assert(limit == luaH_realasize(t) &&
  855. (limit == 0 || !isempty(&t->array[limit - 1])));
  856. if (isdummy(t) || isempty(luaH_getint(t, cast(lua_Integer, limit + 1))))
  857. return limit; /* 'limit + 1' is absent */
  858. else /* 'limit + 1' is also present */
  859. return hash_search(t, limit);
  860. }
  861. #if defined(LUA_DEBUG)
  862. /* export these functions for the test library */
  863. Node *luaH_mainposition (const Table *t, const TValue *key) {
  864. return mainpositionTV(t, key);
  865. }
  866. int luaH_isdummy (const Table *t) { return isdummy(t); }
  867. #endif