1*f0dad708Snikita /* $NetBSD: lstring.c,v 1.10 2023/04/16 20:46:17 nikita Exp $ */
2dbec5304Smbalmer
3dbec5304Smbalmer /*
4*f0dad708Snikita ** Id: lstring.c
5dbec5304Smbalmer ** String table (keeps all strings handled by Lua)
6dbec5304Smbalmer ** See Copyright Notice in lua.h
7dbec5304Smbalmer */
8dbec5304Smbalmer
973008250Slneto #define lstring_c
1073008250Slneto #define LUA_CORE
1173008250Slneto
1273008250Slneto #include "lprefix.h"
1373008250Slneto
14dbec5304Smbalmer
154ab4902eSlneto #ifndef _KERNEL
16dbec5304Smbalmer #include <string.h>
172d6cb6c2Slneto #endif /* _KERNEL */
18dbec5304Smbalmer
19dbec5304Smbalmer #include "lua.h"
20dbec5304Smbalmer
214ab4902eSlneto #include "ldebug.h"
224ab4902eSlneto #include "ldo.h"
23dbec5304Smbalmer #include "lmem.h"
24dbec5304Smbalmer #include "lobject.h"
25dbec5304Smbalmer #include "lstate.h"
26dbec5304Smbalmer #include "lstring.h"
27dbec5304Smbalmer
28dbec5304Smbalmer
294ab4902eSlneto /*
30*f0dad708Snikita ** Maximum size for string table.
314ab4902eSlneto */
32*f0dad708Snikita #define MAXSTRTB cast_int(luaM_limitN(MAX_INT, TString*))
334ab4902eSlneto
344ab4902eSlneto
354ab4902eSlneto /*
364ab4902eSlneto ** equality for long strings
374ab4902eSlneto */
luaS_eqlngstr(TString * a,TString * b)384ab4902eSlneto int luaS_eqlngstr (TString *a, TString *b) {
39bee09862Smbalmer size_t len = a->u.lnglen;
40*f0dad708Snikita lua_assert(a->tt == LUA_VLNGSTR && b->tt == LUA_VLNGSTR);
414ab4902eSlneto return (a == b) || /* same instance or... */
42bee09862Smbalmer ((len == b->u.lnglen) && /* equal length and ... */
434ab4902eSlneto (memcmp(getstr(a), getstr(b), len) == 0)); /* equal contents */
444ab4902eSlneto }
454ab4902eSlneto
464ab4902eSlneto
luaS_hash(const char * str,size_t l,unsigned int seed)474ab4902eSlneto unsigned int luaS_hash (const char *str, size_t l, unsigned int seed) {
48*f0dad708Snikita unsigned int h = seed ^ cast_uint(l);
49*f0dad708Snikita for (; l > 0; l--)
502d6cb6c2Slneto h ^= ((h<<5) + (h>>2) + cast_byte(str[l - 1]));
514ab4902eSlneto return h;
524ab4902eSlneto }
534ab4902eSlneto
544ab4902eSlneto
luaS_hashlongstr(TString * ts)552d6cb6c2Slneto unsigned int luaS_hashlongstr (TString *ts) {
56*f0dad708Snikita lua_assert(ts->tt == LUA_VLNGSTR);
572d6cb6c2Slneto if (ts->extra == 0) { /* no hash? */
58*f0dad708Snikita size_t len = ts->u.lnglen;
59*f0dad708Snikita ts->hash = luaS_hash(getstr(ts), len, ts->hash);
602d6cb6c2Slneto ts->extra = 1; /* now it has its hash */
612d6cb6c2Slneto }
622d6cb6c2Slneto return ts->hash;
632d6cb6c2Slneto }
642d6cb6c2Slneto
652d6cb6c2Slneto
tablerehash(TString ** vect,int osize,int nsize)66*f0dad708Snikita static void tablerehash (TString **vect, int osize, int nsize) {
67dbec5304Smbalmer int i;
68*f0dad708Snikita for (i = osize; i < nsize; i++) /* clear new elements */
69*f0dad708Snikita vect[i] = NULL;
70*f0dad708Snikita for (i = 0; i < osize; i++) { /* rehash old part of the array */
71*f0dad708Snikita TString *p = vect[i];
72*f0dad708Snikita vect[i] = NULL;
73*f0dad708Snikita while (p) { /* for each string in the list */
74bee09862Smbalmer TString *hnext = p->u.hnext; /* save next */
75*f0dad708Snikita unsigned int h = lmod(p->hash, nsize); /* new position */
76*f0dad708Snikita p->u.hnext = vect[h]; /* chain it into array */
77*f0dad708Snikita vect[h] = p;
784ab4902eSlneto p = hnext;
79dbec5304Smbalmer }
80dbec5304Smbalmer }
814ab4902eSlneto }
82*f0dad708Snikita
83*f0dad708Snikita
84*f0dad708Snikita /*
85*f0dad708Snikita ** Resize the string table. If allocation fails, keep the current size.
86*f0dad708Snikita ** (This can degrade performance, but any non-zero size should work
87*f0dad708Snikita ** correctly.)
88*f0dad708Snikita */
luaS_resize(lua_State * L,int nsize)89*f0dad708Snikita void luaS_resize (lua_State *L, int nsize) {
90*f0dad708Snikita stringtable *tb = &G(L)->strt;
91*f0dad708Snikita int osize = tb->size;
92*f0dad708Snikita TString **newvect;
93*f0dad708Snikita if (nsize < osize) /* shrinking table? */
94*f0dad708Snikita tablerehash(tb->hash, osize, nsize); /* depopulate shrinking part */
95*f0dad708Snikita newvect = luaM_reallocvector(L, tb->hash, osize, nsize, TString*);
96*f0dad708Snikita if (l_unlikely(newvect == NULL)) { /* reallocation failed? */
97*f0dad708Snikita if (nsize < osize) /* was it shrinking table? */
98*f0dad708Snikita tablerehash(tb->hash, nsize, osize); /* restore to original size */
99*f0dad708Snikita /* leave table as it was */
100*f0dad708Snikita }
101*f0dad708Snikita else { /* allocation succeeded */
102*f0dad708Snikita tb->hash = newvect;
103*f0dad708Snikita tb->size = nsize;
104*f0dad708Snikita if (nsize > osize)
105*f0dad708Snikita tablerehash(newvect, osize, nsize); /* rehash for new size */
106*f0dad708Snikita }
107dbec5304Smbalmer }
108dbec5304Smbalmer
109dbec5304Smbalmer
110bee09862Smbalmer /*
111bee09862Smbalmer ** Clear API string cache. (Entries cannot be empty, so fill them with
112bee09862Smbalmer ** a non-collectable string.)
113bee09862Smbalmer */
luaS_clearcache(global_State * g)114bee09862Smbalmer void luaS_clearcache (global_State *g) {
1152d6cb6c2Slneto int i, j;
1162d6cb6c2Slneto for (i = 0; i < STRCACHE_N; i++)
1172d6cb6c2Slneto for (j = 0; j < STRCACHE_M; j++) {
1182d6cb6c2Slneto if (iswhite(g->strcache[i][j])) /* will entry be collected? */
1192d6cb6c2Slneto g->strcache[i][j] = g->memerrmsg; /* replace it with something fixed */
120bee09862Smbalmer }
121bee09862Smbalmer }
122bee09862Smbalmer
123bee09862Smbalmer
124bee09862Smbalmer /*
125bee09862Smbalmer ** Initialize the string table and the string cache
126bee09862Smbalmer */
luaS_init(lua_State * L)127bee09862Smbalmer void luaS_init (lua_State *L) {
128bee09862Smbalmer global_State *g = G(L);
1292d6cb6c2Slneto int i, j;
130*f0dad708Snikita stringtable *tb = &G(L)->strt;
131*f0dad708Snikita tb->hash = luaM_newvector(L, MINSTRTABSIZE, TString*);
132*f0dad708Snikita tablerehash(tb->hash, 0, MINSTRTABSIZE); /* clear array */
133*f0dad708Snikita tb->size = MINSTRTABSIZE;
134bee09862Smbalmer /* pre-create memory-error message */
135bee09862Smbalmer g->memerrmsg = luaS_newliteral(L, MEMERRMSG);
136bee09862Smbalmer luaC_fix(L, obj2gco(g->memerrmsg)); /* it should never be collected */
1372d6cb6c2Slneto for (i = 0; i < STRCACHE_N; i++) /* fill cache with valid strings */
1382d6cb6c2Slneto for (j = 0; j < STRCACHE_M; j++)
1392d6cb6c2Slneto g->strcache[i][j] = g->memerrmsg;
140bee09862Smbalmer }
141bee09862Smbalmer
142bee09862Smbalmer
1434ab4902eSlneto
1444ab4902eSlneto /*
1454ab4902eSlneto ** creates a new string object
1464ab4902eSlneto */
createstrobj(lua_State * L,size_t l,int tag,unsigned int h)1472d6cb6c2Slneto static TString *createstrobj (lua_State *L, size_t l, int tag, unsigned int h) {
148dbec5304Smbalmer TString *ts;
1494ab4902eSlneto GCObject *o;
1504ab4902eSlneto size_t totalsize; /* total size of TString object */
15173008250Slneto totalsize = sizelstring(l);
1524ab4902eSlneto o = luaC_newobj(L, tag, totalsize);
15373008250Slneto ts = gco2ts(o);
15473008250Slneto ts->hash = h;
15573008250Slneto ts->extra = 0;
1562d6cb6c2Slneto getstr(ts)[l] = '\0'; /* ending 0 */
1572d6cb6c2Slneto return ts;
1582d6cb6c2Slneto }
1592d6cb6c2Slneto
1602d6cb6c2Slneto
luaS_createlngstrobj(lua_State * L,size_t l)1612d6cb6c2Slneto TString *luaS_createlngstrobj (lua_State *L, size_t l) {
162*f0dad708Snikita TString *ts = createstrobj(L, l, LUA_VLNGSTR, G(L)->seed);
1632d6cb6c2Slneto ts->u.lnglen = l;
164dbec5304Smbalmer return ts;
165dbec5304Smbalmer }
166dbec5304Smbalmer
167dbec5304Smbalmer
luaS_remove(lua_State * L,TString * ts)1684ab4902eSlneto void luaS_remove (lua_State *L, TString *ts) {
1694ab4902eSlneto stringtable *tb = &G(L)->strt;
17073008250Slneto TString **p = &tb->hash[lmod(ts->hash, tb->size)];
1714ab4902eSlneto while (*p != ts) /* find previous element */
172bee09862Smbalmer p = &(*p)->u.hnext;
173bee09862Smbalmer *p = (*p)->u.hnext; /* remove element from its list */
1744ab4902eSlneto tb->nuse--;
1754ab4902eSlneto }
1764ab4902eSlneto
1774ab4902eSlneto
growstrtab(lua_State * L,stringtable * tb)178*f0dad708Snikita static void growstrtab (lua_State *L, stringtable *tb) {
179*f0dad708Snikita if (l_unlikely(tb->nuse == MAX_INT)) { /* too many strings? */
180*f0dad708Snikita luaC_fullgc(L, 1); /* try to free some... */
181*f0dad708Snikita if (tb->nuse == MAX_INT) /* still too many? */
182*f0dad708Snikita luaM_error(L); /* cannot even create a message... */
183*f0dad708Snikita }
184*f0dad708Snikita if (tb->size <= MAXSTRTB / 2) /* can grow string table? */
185*f0dad708Snikita luaS_resize(L, tb->size * 2);
186*f0dad708Snikita }
187*f0dad708Snikita
188*f0dad708Snikita
1894ab4902eSlneto /*
190*f0dad708Snikita ** Checks whether short string exists and reuses it or creates a new one.
1914ab4902eSlneto */
internshrstr(lua_State * L,const char * str,size_t l)1924ab4902eSlneto static TString *internshrstr (lua_State *L, const char *str, size_t l) {
1934ab4902eSlneto TString *ts;
1944ab4902eSlneto global_State *g = G(L);
195*f0dad708Snikita stringtable *tb = &g->strt;
1964ab4902eSlneto unsigned int h = luaS_hash(str, l, g->seed);
197*f0dad708Snikita TString **list = &tb->hash[lmod(h, tb->size)];
1982d6cb6c2Slneto lua_assert(str != NULL); /* otherwise 'memcmp'/'memcpy' are undefined */
199bee09862Smbalmer for (ts = *list; ts != NULL; ts = ts->u.hnext) {
200*f0dad708Snikita if (l == ts->shrlen && (memcmp(str, getstr(ts), l * sizeof(char)) == 0)) {
2014ab4902eSlneto /* found! */
20273008250Slneto if (isdead(g, ts)) /* dead (but not collected yet)? */
20373008250Slneto changewhite(ts); /* resurrect it */
2044ab4902eSlneto return ts;
2054ab4902eSlneto }
2064ab4902eSlneto }
207*f0dad708Snikita /* else must create a new string */
208*f0dad708Snikita if (tb->nuse >= tb->size) { /* need to grow string table? */
209*f0dad708Snikita growstrtab(L, tb);
210*f0dad708Snikita list = &tb->hash[lmod(h, tb->size)]; /* rehash with new size */
2114ab4902eSlneto }
212*f0dad708Snikita ts = createstrobj(L, l, LUA_VSHRSTR, h);
2132d6cb6c2Slneto memcpy(getstr(ts), str, l * sizeof(char));
214bee09862Smbalmer ts->shrlen = cast_byte(l);
215bee09862Smbalmer ts->u.hnext = *list;
2164ab4902eSlneto *list = ts;
217*f0dad708Snikita tb->nuse++;
2184ab4902eSlneto return ts;
2194ab4902eSlneto }
2204ab4902eSlneto
2214ab4902eSlneto
2224ab4902eSlneto /*
2234ab4902eSlneto ** new string (with explicit length)
2244ab4902eSlneto */
luaS_newlstr(lua_State * L,const char * str,size_t l)225dbec5304Smbalmer TString *luaS_newlstr (lua_State *L, const char *str, size_t l) {
2264ab4902eSlneto if (l <= LUAI_MAXSHORTLEN) /* short string? */
2274ab4902eSlneto return internshrstr(L, str, l);
2284ab4902eSlneto else {
229bee09862Smbalmer TString *ts;
230*f0dad708Snikita if (l_unlikely(l >= (MAX_SIZE - sizeof(TString))/sizeof(char)))
231dbec5304Smbalmer luaM_toobig(L);
2322d6cb6c2Slneto ts = luaS_createlngstrobj(L, l);
2332d6cb6c2Slneto memcpy(getstr(ts), str, l * sizeof(char));
234bee09862Smbalmer return ts;
2354ab4902eSlneto }
2364ab4902eSlneto }
2374ab4902eSlneto
2384ab4902eSlneto
2394ab4902eSlneto /*
240bee09862Smbalmer ** Create or reuse a zero-terminated string, first checking in the
241bee09862Smbalmer ** cache (using the string address as a key). The cache can contain
242bee09862Smbalmer ** only zero-terminated strings, so it is safe to use 'strcmp' to
243bee09862Smbalmer ** check hits.
2444ab4902eSlneto */
luaS_new(lua_State * L,const char * str)2454ab4902eSlneto TString *luaS_new (lua_State *L, const char *str) {
2462d6cb6c2Slneto unsigned int i = point2uint(str) % STRCACHE_N; /* hash */
2472d6cb6c2Slneto int j;
248bee09862Smbalmer TString **p = G(L)->strcache[i];
2492d6cb6c2Slneto for (j = 0; j < STRCACHE_M; j++) {
2502d6cb6c2Slneto if (strcmp(str, getstr(p[j])) == 0) /* hit? */
2512d6cb6c2Slneto return p[j]; /* that is it */
252bee09862Smbalmer }
2532d6cb6c2Slneto /* normal route */
2542d6cb6c2Slneto for (j = STRCACHE_M - 1; j > 0; j--)
2552d6cb6c2Slneto p[j] = p[j - 1]; /* move out last element */
2562d6cb6c2Slneto /* new element is first in the list */
2572d6cb6c2Slneto p[0] = luaS_newlstr(L, str, strlen(str));
2582d6cb6c2Slneto return p[0];
2594ab4902eSlneto }
2604ab4902eSlneto
2614ab4902eSlneto
luaS_newudata(lua_State * L,size_t s,int nuvalue)262*f0dad708Snikita Udata *luaS_newudata (lua_State *L, size_t s, int nuvalue) {
2634ab4902eSlneto Udata *u;
264*f0dad708Snikita int i;
2654ab4902eSlneto GCObject *o;
266*f0dad708Snikita if (l_unlikely(s > MAX_SIZE - udatamemoffset(nuvalue)))
2674ab4902eSlneto luaM_toobig(L);
268*f0dad708Snikita o = luaC_newobj(L, LUA_VUSERDATA, sizeudata(nuvalue, s));
26973008250Slneto u = gco2u(o);
27073008250Slneto u->len = s;
271*f0dad708Snikita u->nuvalue = nuvalue;
27273008250Slneto u->metatable = NULL;
273*f0dad708Snikita for (i = 0; i < nuvalue; i++)
274*f0dad708Snikita setnilvalue(&u->uv[i].uv);
275dbec5304Smbalmer return u;
276dbec5304Smbalmer }
277dbec5304Smbalmer
278