/* ** $Id: lstrlib.c,v 1.226 2015/02/09 18:05:46 roberto Exp roberto $ ** Standard library for string operations and pattern-matching ** See Copyright Notice in lua.h */ #define lstrlib_c #define LUA_LIB #include "lprefix.h" #include #include #include #include #include #include #include "lua.h" #include "lauxlib.h" #include "lualib.h" /* ** maximum number of captures that a pattern can do during ** pattern-matching. This limit is arbitrary. */ #if !defined(LUA_MAXCAPTURES) #define LUA_MAXCAPTURES 32 #endif /* macro to 'unsign' a character */ #define uchar(c) ((unsigned char)(c)) /* ** Some sizes are better limited to fit in 'int', but must also fit in ** 'size_t'. (We assume that 'lua_Integer' cannot be smaller than 'int'.) */ #define MAXSIZE \ (sizeof(size_t) < sizeof(int) ? (~(size_t)0) : (size_t)(INT_MAX)) static int str_len (lua_State *L) { size_t l; luaL_checklstring(L, 1, &l); lua_pushinteger(L, (lua_Integer)l); return 1; } /* translate a relative string position: negative means back from end */ static lua_Integer posrelat (lua_Integer pos, size_t len) { if (pos >= 0) return pos; else if (0u - (size_t)pos > len) return 0; else return (lua_Integer)len + pos + 1; } static int str_sub (lua_State *L) { size_t l; const char *s = luaL_checklstring(L, 1, &l); lua_Integer start = posrelat(luaL_checkinteger(L, 2), l); lua_Integer end = posrelat(luaL_optinteger(L, 3, -1), l); if (start < 1) start = 1; if (end > (lua_Integer)l) end = l; if (start <= end) lua_pushlstring(L, s + start - 1, (size_t)(end - start) + 1); else lua_pushliteral(L, ""); return 1; } static int str_reverse (lua_State *L) { size_t l, i; luaL_Buffer b; const char *s = luaL_checklstring(L, 1, &l); char *p = luaL_buffinitsize(L, &b, l); for (i = 0; i < l; i++) p[i] = s[l - i - 1]; luaL_pushresultsize(&b, l); return 1; } static int str_lower (lua_State *L) { size_t l; size_t i; luaL_Buffer b; const char *s = luaL_checklstring(L, 1, &l); char *p = luaL_buffinitsize(L, &b, l); for (i=0; i MAXSIZE / n) /* may overflow? */ return luaL_error(L, "resulting string too large"); else { size_t totallen = (size_t)n * l + (size_t)(n - 1) * lsep; luaL_Buffer b; char *p = luaL_buffinitsize(L, &b, totallen); while (n-- > 1) { /* first n-1 copies (followed by separator) */ memcpy(p, s, l * sizeof(char)); p += l; if (lsep > 0) { /* empty 'memcpy' is not that cheap */ memcpy(p, sep, lsep * sizeof(char)); p += lsep; } } memcpy(p, s, l * sizeof(char)); /* last copy (not followed by separator) */ luaL_pushresultsize(&b, totallen); } return 1; } static int str_byte (lua_State *L) { size_t l; const char *s = luaL_checklstring(L, 1, &l); lua_Integer posi = posrelat(luaL_optinteger(L, 2, 1), l); lua_Integer pose = posrelat(luaL_optinteger(L, 3, posi), l); int n, i; if (posi < 1) posi = 1; if (pose > (lua_Integer)l) pose = l; if (posi > pose) return 0; /* empty interval; return no values */ if (pose - posi >= INT_MAX) /* arithmetic overflow? */ return luaL_error(L, "string slice too long"); n = (int)(pose - posi) + 1; luaL_checkstack(L, n, "string slice too long"); for (i=0; i= ms->level || ms->capture[l].len == CAP_UNFINISHED) return luaL_error(ms->L, "invalid capture index %%%d", l + 1); return l; } static int capture_to_close (MatchState *ms) { int level = ms->level; for (level--; level>=0; level--) if (ms->capture[level].len == CAP_UNFINISHED) return level; return luaL_error(ms->L, "invalid pattern capture"); } static const char *classend (MatchState *ms, const char *p) { switch (*p++) { case L_ESC: { if (p == ms->p_end) luaL_error(ms->L, "malformed pattern (ends with '%%')"); return p+1; } case '[': { if (*p == '^') p++; do { /* look for a ']' */ if (p == ms->p_end) luaL_error(ms->L, "malformed pattern (missing ']')"); if (*(p++) == L_ESC && p < ms->p_end) p++; /* skip escapes (e.g. '%]') */ } while (*p != ']'); return p+1; } default: { return p; } } } static int match_class (int c, int cl) { int res; switch (tolower(cl)) { case 'a' : res = isalpha(c); break; case 'c' : res = iscntrl(c); break; case 'd' : res = isdigit(c); break; case 'g' : res = isgraph(c); break; case 'l' : res = islower(c); break; case 'p' : res = ispunct(c); break; case 's' : res = isspace(c); break; case 'u' : res = isupper(c); break; case 'w' : res = isalnum(c); break; case 'x' : res = isxdigit(c); break; case 'z' : res = (c == 0); break; /* deprecated option */ default: return (cl == c); } return (islower(cl) ? res : !res); } static int matchbracketclass (int c, const char *p, const char *ec) { int sig = 1; if (*(p+1) == '^') { sig = 0; p++; /* skip the '^' */ } while (++p < ec) { if (*p == L_ESC) { p++; if (match_class(c, uchar(*p))) return sig; } else if ((*(p+1) == '-') && (p+2 < ec)) { p+=2; if (uchar(*(p-2)) <= c && c <= uchar(*p)) return sig; } else if (uchar(*p) == c) return sig; } return !sig; } static int singlematch (MatchState *ms, const char *s, const char *p, const char *ep) { if (s >= ms->src_end) return 0; else { int c = uchar(*s); switch (*p) { case '.': return 1; /* matches any char */ case L_ESC: return match_class(c, uchar(*(p+1))); case '[': return matchbracketclass(c, p, ep-1); default: return (uchar(*p) == c); } } } static const char *matchbalance (MatchState *ms, const char *s, const char *p) { if (p >= ms->p_end - 1) luaL_error(ms->L, "malformed pattern (missing arguments to '%%b')"); if (*s != *p) return NULL; else { int b = *p; int e = *(p+1); int cont = 1; while (++s < ms->src_end) { if (*s == e) { if (--cont == 0) return s+1; } else if (*s == b) cont++; } } return NULL; /* string ends out of balance */ } static const char *max_expand (MatchState *ms, const char *s, const char *p, const char *ep) { ptrdiff_t i = 0; /* counts maximum expand for item */ while (singlematch(ms, s + i, p, ep)) i++; /* keeps trying to match with the maximum repetitions */ while (i>=0) { const char *res = match(ms, (s+i), ep+1); if (res) return res; i--; /* else didn't match; reduce 1 repetition to try again */ } return NULL; } static const char *min_expand (MatchState *ms, const char *s, const char *p, const char *ep) { for (;;) { const char *res = match(ms, s, ep+1); if (res != NULL) return res; else if (singlematch(ms, s, p, ep)) s++; /* try with one more repetition */ else return NULL; } } static const char *start_capture (MatchState *ms, const char *s, const char *p, int what) { const char *res; int level = ms->level; if (level >= LUA_MAXCAPTURES) luaL_error(ms->L, "too many captures"); ms->capture[level].init = s; ms->capture[level].len = what; ms->level = level+1; if ((res=match(ms, s, p)) == NULL) /* match failed? */ ms->level--; /* undo capture */ return res; } static const char *end_capture (MatchState *ms, const char *s, const char *p) { int l = capture_to_close(ms); const char *res; ms->capture[l].len = s - ms->capture[l].init; /* close capture */ if ((res = match(ms, s, p)) == NULL) /* match failed? */ ms->capture[l].len = CAP_UNFINISHED; /* undo capture */ return res; } static const char *match_capture (MatchState *ms, const char *s, int l) { size_t len; l = check_capture(ms, l); len = ms->capture[l].len; if ((size_t)(ms->src_end-s) >= len && memcmp(ms->capture[l].init, s, len) == 0) return s+len; else return NULL; } static const char *match (MatchState *ms, const char *s, const char *p) { if (ms->matchdepth-- == 0) luaL_error(ms->L, "pattern too complex"); init: /* using goto's to optimize tail recursion */ if (p != ms->p_end) { /* end of pattern? */ switch (*p) { case '(': { /* start capture */ if (*(p + 1) == ')') /* position capture? */ s = start_capture(ms, s, p + 2, CAP_POSITION); else s = start_capture(ms, s, p + 1, CAP_UNFINISHED); break; } case ')': { /* end capture */ s = end_capture(ms, s, p + 1); break; } case '$': { if ((p + 1) != ms->p_end) /* is the '$' the last char in pattern? */ goto dflt; /* no; go to default */ s = (s == ms->src_end) ? s : NULL; /* check end of string */ break; } case L_ESC: { /* escaped sequences not in the format class[*+?-]? */ switch (*(p + 1)) { case 'b': { /* balanced string? */ s = matchbalance(ms, s, p + 2); if (s != NULL) { p += 4; goto init; /* return match(ms, s, p + 4); */ } /* else fail (s == NULL) */ break; } case 'f': { /* frontier? */ const char *ep; char previous; p += 2; if (*p != '[') luaL_error(ms->L, "missing '[' after '%%f' in pattern"); ep = classend(ms, p); /* points to what is next */ previous = (s == ms->src_init) ? '\0' : *(s - 1); if (!matchbracketclass(uchar(previous), p, ep - 1) && matchbracketclass(uchar(*s), p, ep - 1)) { p = ep; goto init; /* return match(ms, s, ep); */ } s = NULL; /* match failed */ break; } case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': { /* capture results (%0-%9)? */ s = match_capture(ms, s, uchar(*(p + 1))); if (s != NULL) { p += 2; goto init; /* return match(ms, s, p + 2) */ } break; } default: goto dflt; } break; } default: dflt: { /* pattern class plus optional suffix */ const char *ep = classend(ms, p); /* points to optional suffix */ /* does not match at least once? */ if (!singlematch(ms, s, p, ep)) { if (*ep == '*' || *ep == '?' || *ep == '-') { /* accept empty? */ p = ep + 1; goto init; /* return match(ms, s, ep + 1); */ } else /* '+' or no suffix */ s = NULL; /* fail */ } else { /* matched once */ switch (*ep) { /* handle optional suffix */ case '?': { /* optional */ const char *res; if ((res = match(ms, s + 1, ep + 1)) != NULL) s = res; else { p = ep + 1; goto init; /* else return match(ms, s, ep + 1); */ } break; } case '+': /* 1 or more repetitions */ s++; /* 1 match already done */ /* FALLTHROUGH */ case '*': /* 0 or more repetitions */ s = max_expand(ms, s, p, ep); break; case '-': /* 0 or more repetitions (minimum) */ s = min_expand(ms, s, p, ep); break; default: /* no suffix */ s++; p = ep; goto init; /* return match(ms, s + 1, ep); */ } } break; } } } ms->matchdepth++; return s; } static const char *lmemfind (const char *s1, size_t l1, const char *s2, size_t l2) { if (l2 == 0) return s1; /* empty strings are everywhere */ else if (l2 > l1) return NULL; /* avoids a negative 'l1' */ else { const char *init; /* to search for a '*s2' inside 's1' */ l2--; /* 1st char will be checked by 'memchr' */ l1 = l1-l2; /* 's2' cannot be found after that */ while (l1 > 0 && (init = (const char *)memchr(s1, *s2, l1)) != NULL) { init++; /* 1st char is already checked */ if (memcmp(init, s2+1, l2) == 0) return init-1; else { /* correct 'l1' and 's1' to try again */ l1 -= init-s1; s1 = init; } } return NULL; /* not found */ } } static void push_onecapture (MatchState *ms, int i, const char *s, const char *e) { if (i >= ms->level) { if (i == 0) /* ms->level == 0, too */ lua_pushlstring(ms->L, s, e - s); /* add whole match */ else luaL_error(ms->L, "invalid capture index %%%d", i + 1); } else { ptrdiff_t l = ms->capture[i].len; if (l == CAP_UNFINISHED) luaL_error(ms->L, "unfinished capture"); if (l == CAP_POSITION) lua_pushinteger(ms->L, (ms->capture[i].init - ms->src_init) + 1); else lua_pushlstring(ms->L, ms->capture[i].init, l); } } static int push_captures (MatchState *ms, const char *s, const char *e) { int i; int nlevels = (ms->level == 0 && s) ? 1 : ms->level; luaL_checkstack(ms->L, nlevels, "too many captures"); for (i = 0; i < nlevels; i++) push_onecapture(ms, i, s, e); return nlevels; /* number of strings pushed */ } /* check whether pattern has no special characters */ static int nospecials (const char *p, size_t l) { size_t upto = 0; do { if (strpbrk(p + upto, SPECIALS)) return 0; /* pattern has a special character */ upto += strlen(p + upto) + 1; /* may have more after \0 */ } while (upto <= l); return 1; /* no special chars found */ } static int str_find_aux (lua_State *L, int find) { size_t ls, lp; const char *s = luaL_checklstring(L, 1, &ls); const char *p = luaL_checklstring(L, 2, &lp); lua_Integer init = posrelat(luaL_optinteger(L, 3, 1), ls); if (init < 1) init = 1; else if (init > (lua_Integer)ls + 1) { /* start after string's end? */ lua_pushnil(L); /* cannot find anything */ return 1; } /* explicit request or no special characters? */ if (find && (lua_toboolean(L, 4) || nospecials(p, lp))) { /* do a plain search */ const char *s2 = lmemfind(s + init - 1, ls - (size_t)init + 1, p, lp); if (s2) { lua_pushinteger(L, (s2 - s) + 1); lua_pushinteger(L, (s2 - s) + lp); return 2; } } else { MatchState ms; const char *s1 = s + init - 1; int anchor = (*p == '^'); if (anchor) { p++; lp--; /* skip anchor character */ } ms.L = L; ms.matchdepth = MAXCCALLS; ms.src_init = s; ms.src_end = s + ls; ms.p_end = p + lp; do { const char *res; ms.level = 0; lua_assert(ms.matchdepth == MAXCCALLS); if ((res=match(&ms, s1, p)) != NULL) { if (find) { lua_pushinteger(L, (s1 - s) + 1); /* start */ lua_pushinteger(L, res - s); /* end */ return push_captures(&ms, NULL, 0) + 2; } else return push_captures(&ms, s1, res); } } while (s1++ < ms.src_end && !anchor); } lua_pushnil(L); /* not found */ return 1; } static int str_find (lua_State *L) { return str_find_aux(L, 1); } static int str_match (lua_State *L) { return str_find_aux(L, 0); } static int gmatch_aux (lua_State *L) { MatchState ms; size_t ls, lp; const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls); const char *p = lua_tolstring(L, lua_upvalueindex(2), &lp); const char *src; ms.L = L; ms.matchdepth = MAXCCALLS; ms.src_init = s; ms.src_end = s+ls; ms.p_end = p + lp; for (src = s + (size_t)lua_tointeger(L, lua_upvalueindex(3)); src <= ms.src_end; src++) { const char *e; ms.level = 0; lua_assert(ms.matchdepth == MAXCCALLS); if ((e = match(&ms, src, p)) != NULL) { lua_Integer newstart = e-s; if (e == src) newstart++; /* empty match? go at least one position */ lua_pushinteger(L, newstart); lua_replace(L, lua_upvalueindex(3)); return push_captures(&ms, src, e); } } return 0; /* not found */ } static int gmatch (lua_State *L) { luaL_checkstring(L, 1); luaL_checkstring(L, 2); lua_settop(L, 2); lua_pushinteger(L, 0); lua_pushcclosure(L, gmatch_aux, 3); return 1; } static void add_s (MatchState *ms, luaL_Buffer *b, const char *s, const char *e) { size_t l, i; lua_State *L = ms->L; const char *news = lua_tolstring(L, 3, &l); for (i = 0; i < l; i++) { if (news[i] != L_ESC) luaL_addchar(b, news[i]); else { i++; /* skip ESC */ if (!isdigit(uchar(news[i]))) { if (news[i] != L_ESC) luaL_error(L, "invalid use of '%c' in replacement string", L_ESC); luaL_addchar(b, news[i]); } else if (news[i] == '0') luaL_addlstring(b, s, e - s); else { push_onecapture(ms, news[i] - '1', s, e); luaL_tolstring(L, -1, NULL); /* if number, convert it to string */ lua_remove(L, -2); /* remove original value */ luaL_addvalue(b); /* add capture to accumulated result */ } } } } static void add_value (MatchState *ms, luaL_Buffer *b, const char *s, const char *e, int tr) { lua_State *L = ms->L; switch (tr) { case LUA_TFUNCTION: { int n; lua_pushvalue(L, 3); n = push_captures(ms, s, e); lua_call(L, n, 1); break; } case LUA_TTABLE: { push_onecapture(ms, 0, s, e); lua_gettable(L, 3); break; } default: { /* LUA_TNUMBER or LUA_TSTRING */ add_s(ms, b, s, e); return; } } if (!lua_toboolean(L, -1)) { /* nil or false? */ lua_pop(L, 1); lua_pushlstring(L, s, e - s); /* keep original text */ } else if (!lua_isstring(L, -1)) luaL_error(L, "invalid replacement value (a %s)", luaL_typename(L, -1)); luaL_addvalue(b); /* add result to accumulator */ } static int str_gsub (lua_State *L) { size_t srcl, lp; const char *src = luaL_checklstring(L, 1, &srcl); const char *p = luaL_checklstring(L, 2, &lp); int tr = lua_type(L, 3); lua_Integer max_s = luaL_optinteger(L, 4, srcl + 1); int anchor = (*p == '^'); lua_Integer n = 0; MatchState ms; luaL_Buffer b; luaL_argcheck(L, tr == LUA_TNUMBER || tr == LUA_TSTRING || tr == LUA_TFUNCTION || tr == LUA_TTABLE, 3, "string/function/table expected"); luaL_buffinit(L, &b); if (anchor) { p++; lp--; /* skip anchor character */ } ms.L = L; ms.matchdepth = MAXCCALLS; ms.src_init = src; ms.src_end = src+srcl; ms.p_end = p + lp; while (n < max_s) { const char *e; ms.level = 0; lua_assert(ms.matchdepth == MAXCCALLS); e = match(&ms, src, p); if (e) { n++; add_value(&ms, &b, src, e, tr); } if (e && e>src) /* non empty match? */ src = e; /* skip it */ else if (src < ms.src_end) luaL_addchar(&b, *src++); else break; if (anchor) break; } luaL_addlstring(&b, src, ms.src_end-src); luaL_pushresult(&b); lua_pushinteger(L, n); /* number of substitutions */ return 2; } /* }====================================================== */ /* ** {====================================================== ** STRING FORMAT ** ======================================================= */ #if !defined(lua_number2strx) /* { */ /* ** Hexadecimal floating-point formatter */ #include #define SIZELENMOD (sizeof(LUA_NUMBER_FRMLEN)/sizeof(char)) /* ** Number of bits that goes into the first digit. It can be any value ** between 1 and 4; the following definition tries to align the number ** to nibble boundaries. The default is 1 bit, that aligns double ** (1+52-bit mantissa) and quad precision (1+112-bit mantissa). For ** float (24-bit mantissa) and 80-bit long double (64-bit mantissa), 4 ** does the alignment. */ #define L_NBFD ((sizeof(lua_Number) == 4 || sizeof(lua_Number) == 12) ? 4 : 1) /* ** Add integer part of 'x' to buffer and return new 'x' */ static lua_Number adddigit (char *buff, int n, lua_Number x) { lua_Number dd = l_mathop(floor)(x); /* get integer part from 'x' */ int d = (int)dd; buff[n] = (d < 10 ? d + '0' : d - 10 + 'a'); /* add to buffer */ return x - dd; /* return what is left */ } static int num2straux (char *buff, lua_Number x) { if (x != x || x == HUGE_VAL || x == -HUGE_VAL) /* inf or NaN? */ return sprintf(buff, LUA_NUMBER_FMT, x); /* equal to '%g' */ else if (x == 0) { /* can be -0... */ sprintf(buff, LUA_NUMBER_FMT, x); strcat(buff, "x0p+0"); /* reuses '0/-0' from 'sprintf'... */ return strlen(buff); } else { int e; lua_Number m = l_mathop(frexp)(x, &e); /* 'x' fraction and exponent */ int n = 0; /* character count */ if (m < 0) { /* is number negative? */ buff[n++] = '-'; /* add signal */ m = -m; /* make it positive */ } buff[n++] = '0'; buff[n++] = 'x'; /* add "0x" */ m = adddigit(buff, n++, m * (1 << L_NBFD)); /* add first digit */ e -= L_NBFD; /* this digit goes before the radix point */ if (m > 0) { /* more digits? */ buff[n++] = '.'; /* add radix point */ do { /* add as many digits as needed */ m = adddigit(buff, n++, m * 16); } while (m > 0); } n += sprintf(buff + n, "p%+d", e); /* add exponent */ return n; } } static int lua_number2strx (lua_State *L, char *buff, const char *fmt, lua_Number x) { int n = num2straux(buff, x); if (fmt[SIZELENMOD] == 'A') { int i; for (i = 0; i < n; i++) buff[i] = toupper(uchar(buff[i])); } else if (fmt[SIZELENMOD] != 'a') luaL_error(L, "modifiers for format '%%a'/'%%A' not implemented"); return n; } #endif /* } */ /* ** Maximum size of each formatted item. This maximum size is produced ** by format('%.99f', minfloat), and is equal to 99 + 2 ('-' and '.') + ** number of decimal digits to represent minfloat (which is ~308 for ** a double and ~4932 for long double). */ #define MAX_ITEM \ (sizeof(lua_Number) <= 4 ? 150 : sizeof(lua_Number) <= 8 ? 450 : 5050) /* valid flags in a format specification */ #define FLAGS "-+ #0" /* ** maximum size of each format specification (such as "%-099.99d") */ #define MAX_FORMAT 32 static void addquoted (lua_State *L, luaL_Buffer *b, int arg) { size_t l; const char *s = luaL_checklstring(L, arg, &l); luaL_addchar(b, '"'); while (l--) { if (*s == '"' || *s == '\\' || *s == '\n') { luaL_addchar(b, '\\'); luaL_addchar(b, *s); } else if (*s == '\0' || iscntrl(uchar(*s))) { char buff[10]; if (!isdigit(uchar(*(s+1)))) sprintf(buff, "\\%d", (int)uchar(*s)); else sprintf(buff, "\\%03d", (int)uchar(*s)); luaL_addstring(b, buff); } else luaL_addchar(b, *s); s++; } luaL_addchar(b, '"'); } static const char *scanformat (lua_State *L, const char *strfrmt, char *form) { const char *p = strfrmt; while (*p != '\0' && strchr(FLAGS, *p) != NULL) p++; /* skip flags */ if ((size_t)(p - strfrmt) >= sizeof(FLAGS)/sizeof(char)) luaL_error(L, "invalid format (repeated flags)"); if (isdigit(uchar(*p))) p++; /* skip width */ if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ if (*p == '.') { p++; if (isdigit(uchar(*p))) p++; /* skip precision */ if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ } if (isdigit(uchar(*p))) luaL_error(L, "invalid format (width or precision too long)"); *(form++) = '%'; memcpy(form, strfrmt, ((p - strfrmt) + 1) * sizeof(char)); form += (p - strfrmt) + 1; *form = '\0'; return p; } /* ** add length modifier into formats */ static void addlenmod (char *form, const char *lenmod) { size_t l = strlen(form); size_t lm = strlen(lenmod); char spec = form[l - 1]; strcpy(form + l - 1, lenmod); form[l + lm - 1] = spec; form[l + lm] = '\0'; } static int str_format (lua_State *L) { int top = lua_gettop(L); int arg = 1; size_t sfl; const char *strfrmt = luaL_checklstring(L, arg, &sfl); const char *strfrmt_end = strfrmt+sfl; luaL_Buffer b; luaL_buffinit(L, &b); while (strfrmt < strfrmt_end) { if (*strfrmt != L_ESC) luaL_addchar(&b, *strfrmt++); else if (*++strfrmt == L_ESC) luaL_addchar(&b, *strfrmt++); /* %% */ else { /* format item */ char form[MAX_FORMAT]; /* to store the format ('%...') */ char *buff = luaL_prepbuffsize(&b, MAX_ITEM); /* to put formatted item */ int nb = 0; /* number of bytes in added item */ if (++arg > top) luaL_argerror(L, arg, "no value"); strfrmt = scanformat(L, strfrmt, form); switch (*strfrmt++) { case 'c': { nb = sprintf(buff, form, (int)luaL_checkinteger(L, arg)); break; } case 'd': case 'i': case 'o': case 'u': case 'x': case 'X': { lua_Integer n = luaL_checkinteger(L, arg); addlenmod(form, LUA_INTEGER_FRMLEN); nb = sprintf(buff, form, n); break; } case 'a': case 'A': addlenmod(form, LUA_NUMBER_FRMLEN); nb = lua_number2strx(L, buff, form, luaL_checknumber(L, arg)); break; case 'e': case 'E': case 'f': case 'g': case 'G': { addlenmod(form, LUA_NUMBER_FRMLEN); nb = sprintf(buff, form, luaL_checknumber(L, arg)); break; } case 'q': { addquoted(L, &b, arg); break; } case 's': { size_t l; const char *s = luaL_tolstring(L, arg, &l); if (!strchr(form, '.') && l >= 100) { /* no precision and string is too long to be formatted; keep original string */ luaL_addvalue(&b); } else { nb = sprintf(buff, form, s); lua_pop(L, 1); /* remove result from 'luaL_tolstring' */ } break; } default: { /* also treat cases 'pnLlh' */ return luaL_error(L, "invalid option '%%%c' to 'format'", *(strfrmt - 1)); } } luaL_addsize(&b, nb); } } luaL_pushresult(&b); return 1; } /* }====================================================== */ /* ** {====================================================== ** PACK/UNPACK ** ======================================================= */ /* value used for padding */ #if !defined(LUA_PACKPADBYTE) #define LUA_PACKPADBYTE 0x00 #endif /* maximum size for the binary representation of an integer */ #define MAXINTSIZE 16 /* number of bits in a character */ #define NB CHAR_BIT /* mask for one character (NB 1's) */ #define MC ((1 << NB) - 1) /* size of a lua_Integer */ #define SZINT ((int)sizeof(lua_Integer)) /* dummy union to get native endianness */ static const union { int dummy; char little; /* true iff machine is little endian */ } nativeendian = {1}; /* dummy structure to get native alignment requirements */ struct cD { char c; union { double d; void *p; lua_Integer i; lua_Number n; } u; }; #define MAXALIGN (offsetof(struct cD, u)) /* ** Union for serializing floats */ typedef union Ftypes { float f; double d; lua_Number n; char buff[5 * sizeof(lua_Number)]; /* enough for any float type */ } Ftypes; /* ** information to pack/unpack stuff */ typedef struct Header { lua_State *L; int islittle; int maxalign; } Header; /* ** options for pack/unpack */ typedef enum KOption { Kint, /* signed integers */ Kuint, /* unsigned integers */ Kfloat, /* floating-point numbers */ Kchar, /* fixed-length strings */ Kstring, /* strings with prefixed length */ Kzstr, /* zero-terminated strings */ Kpadding, /* padding */ Kpaddalign, /* padding for alignment */ Knop /* no-op (configuration or spaces) */ } KOption; /* ** Read an integer numeral from string 'fmt' or return 'df' if ** there is no numeral */ static int digit (int c) { return '0' <= c && c <= '9'; } static int getnum (const char **fmt, int df) { if (!digit(**fmt)) /* no number? */ return df; /* return default value */ else { int a = 0; do { a = a*10 + (*((*fmt)++) - '0'); } while (digit(**fmt) && a <= ((int)MAXSIZE - 9)/10); return a; } } /* ** Read an integer numeral and raises an error if it is larger ** than the maximum size for integers. */ static int getnumlimit (Header *h, const char **fmt, int df) { int sz = getnum(fmt, df); if (sz > MAXINTSIZE || sz <= 0) luaL_error(h->L, "integral size (%d) out of limits [1,%d]", sz, MAXINTSIZE); return sz; } /* ** Initialize Header */ static void initheader (lua_State *L, Header *h) { h->L = L; h->islittle = nativeendian.little; h->maxalign = 1; } /* ** Read and classify next option. 'size' is filled with option's size. */ static KOption getoption (Header *h, const char **fmt, int *size) { int opt = *((*fmt)++); *size = 0; /* default */ switch (opt) { case 'b': *size = sizeof(char); return Kint; case 'B': *size = sizeof(char); return Kuint; case 'h': *size = sizeof(short); return Kint; case 'H': *size = sizeof(short); return Kuint; case 'l': *size = sizeof(long); return Kint; case 'L': *size = sizeof(long); return Kuint; case 'j': *size = sizeof(lua_Integer); return Kint; case 'J': *size = sizeof(lua_Integer); return Kuint; case 'T': *size = sizeof(size_t); return Kuint; case 'f': *size = sizeof(float); return Kfloat; case 'd': *size = sizeof(double); return Kfloat; case 'n': *size = sizeof(lua_Number); return Kfloat; case 'i': *size = getnumlimit(h, fmt, sizeof(int)); return Kint; case 'I': *size = getnumlimit(h, fmt, sizeof(int)); return Kuint; case 's': *size = getnumlimit(h, fmt, sizeof(size_t)); return Kstring; case 'c': *size = getnum(fmt, -1); if (*size == -1) luaL_error(h->L, "missing size for format option 'c'"); return Kchar; case 'z': return Kzstr; case 'x': *size = 1; return Kpadding; case 'X': return Kpaddalign; case ' ': break; case '<': h->islittle = 1; break; case '>': h->islittle = 0; break; case '=': h->islittle = nativeendian.little; break; case '!': h->maxalign = getnumlimit(h, fmt, MAXALIGN); break; default: luaL_error(h->L, "invalid format option '%c'", opt); } return Knop; } /* ** Read, classify, and fill other details about the next option. ** 'psize' is filled with option's size, 'notoalign' with its ** alignment requirements. ** Local variable 'size' gets the size to be aligned. (Kpadal option ** always gets its full alignment, other options are limited by ** the maximum alignment ('maxalign'). Kchar option needs no alignment ** despite its size. */ static KOption getdetails (Header *h, size_t totalsize, const char **fmt, int *psize, int *ntoalign) { KOption opt = getoption(h, fmt, psize); int align = *psize; /* usually, alignment follows size */ if (opt == Kpaddalign) { /* 'X' gets alignment from following option */ if (**fmt == '\0' || getoption(h, fmt, &align) == Kchar || align == 0) luaL_argerror(h->L, 1, "invalid next option for option 'X'"); } if (align <= 1 || opt == Kchar) /* need no alignment? */ *ntoalign = 0; else { if (align > h->maxalign) /* enforce maximum alignment */ align = h->maxalign; if ((align & (align - 1)) != 0) /* is 'align' not a power of 2? */ luaL_argerror(h->L, 1, "format asks for alignment not power of 2"); *ntoalign = (align - (int)(totalsize & (align - 1))) & (align - 1); } return opt; } /* ** Pack integer 'n' with 'size' bytes and 'islittle' endianness. ** The final 'if' handles the case when 'size' is larger than ** the size of a Lua integer, correcting the extra sign-extension ** bytes if necessary (by default they would be zeros). */ static void packint (luaL_Buffer *b, lua_Unsigned n, int islittle, int size, int neg) { char *buff = luaL_prepbuffsize(b, size); int i; buff[islittle ? 0 : size - 1] = (char)(n & MC); /* first byte */ for (i = 1; i < size; i++) { n >>= NB; buff[islittle ? i : size - 1 - i] = (char)(n & MC); } if (neg && size > SZINT) { /* negative number need sign extension? */ for (i = SZINT; i < size; i++) /* correct extra bytes */ buff[islittle ? i : size - 1 - i] = (char)MC; } luaL_addsize(b, size); /* add result to buffer */ } /* ** Copy 'size' bytes from 'src' to 'dest', correcting endianness if ** given 'islittle' is different from native endianness. */ static void copywithendian (volatile char *dest, volatile const char *src, int size, int islittle) { if (islittle == nativeendian.little) { while (size-- != 0) *(dest++) = *(src++); } else { dest += size - 1; while (size-- != 0) *(dest--) = *(src++); } } static int str_pack (lua_State *L) { luaL_Buffer b; Header h; const char *fmt = luaL_checkstring(L, 1); /* format string */ int arg = 1; /* current argument to pack */ size_t totalsize = 0; /* accumulate total size of result */ initheader(L, &h); lua_pushnil(L); /* mark to separate arguments from string buffer */ luaL_buffinit(L, &b); while (*fmt != '\0') { int size, ntoalign; KOption opt = getdetails(&h, totalsize, &fmt, &size, &ntoalign); totalsize += ntoalign + size; while (ntoalign-- > 0) luaL_addchar(&b, LUA_PACKPADBYTE); /* fill alignment */ arg++; switch (opt) { case Kint: { /* signed integers */ lua_Integer n = luaL_checkinteger(L, arg); if (size < SZINT) { /* need overflow check? */ lua_Integer lim = (lua_Integer)1 << ((size * NB) - 1); luaL_argcheck(L, -lim <= n && n < lim, arg, "integer overflow"); } packint(&b, (lua_Unsigned)n, h.islittle, size, (n < 0)); break; } case Kuint: { /* unsigned integers */ lua_Integer n = luaL_checkinteger(L, arg); if (size < SZINT) /* need overflow check? */ luaL_argcheck(L, (lua_Unsigned)n < ((lua_Unsigned)1 << (size * NB)), arg, "unsigned overflow"); packint(&b, (lua_Unsigned)n, h.islittle, size, 0); break; } case Kfloat: { /* floating-point options */ volatile Ftypes u; char *buff = luaL_prepbuffsize(&b, size); lua_Number n = luaL_checknumber(L, arg); /* get argument */ if (size == sizeof(u.f)) u.f = (float)n; /* copy it into 'u' */ else if (size == sizeof(u.d)) u.d = (double)n; else u.n = n; /* move 'u' to final result, correcting endianness if needed */ copywithendian(buff, u.buff, size, h.islittle); luaL_addsize(&b, size); break; } case Kchar: { /* fixed-size string */ size_t len; const char *s = luaL_checklstring(L, arg, &len); luaL_argcheck(L, len == (size_t)size, arg, "wrong length"); luaL_addlstring(&b, s, size); break; } case Kstring: { /* strings with length count */ size_t len; const char *s = luaL_checklstring(L, arg, &len); luaL_argcheck(L, size >= (int)sizeof(size_t) || len < ((size_t)1 << (size * NB)), arg, "string length does not fit in given size"); packint(&b, (lua_Unsigned)len, h.islittle, size, 0); /* pack length */ luaL_addlstring(&b, s, len); totalsize += len; break; } case Kzstr: { /* zero-terminated string */ size_t len; const char *s = luaL_checklstring(L, arg, &len); luaL_argcheck(L, strlen(s) == len, arg, "string contains zeros"); luaL_addlstring(&b, s, len); luaL_addchar(&b, '\0'); /* add zero at the end */ totalsize += len + 1; break; } case Kpadding: luaL_addchar(&b, LUA_PACKPADBYTE); /* FALLTHROUGH */ case Kpaddalign: case Knop: arg--; /* undo increment */ break; } } luaL_pushresult(&b); return 1; } static int str_packsize (lua_State *L) { Header h; const char *fmt = luaL_checkstring(L, 1); /* format string */ size_t totalsize = 0; /* accumulate total size of result */ initheader(L, &h); while (*fmt != '\0') { int size, ntoalign; KOption opt = getdetails(&h, totalsize, &fmt, &size, &ntoalign); size += ntoalign; /* total space used by option */ luaL_argcheck(L, totalsize <= MAXSIZE - size, 1, "format result too large"); totalsize += size; switch (opt) { case Kstring: /* strings with length count */ case Kzstr: /* zero-terminated string */ luaL_argerror(L, 1, "variable-length format"); break; default: break; } } lua_pushinteger(L, (lua_Integer)totalsize); return 1; } /* ** Unpack an integer with 'size' bytes and 'islittle' endianness. ** If size is smaller than the size of a Lua integer and integer ** is signed, must do sign extension (propagating the sign to the ** higher bits); if size is larger than the size of a Lua integer, ** it must check the unread bytes to see whether they do not cause an ** overflow. */ static lua_Integer unpackint (lua_State *L, const char *str, int islittle, int size, int issigned) { lua_Unsigned res = 0; int i; int limit = (size <= SZINT) ? size : SZINT; for (i = limit - 1; i >= 0; i--) { res <<= NB; res |= (lua_Unsigned)(unsigned char)str[islittle ? i : size - 1 - i]; } if (size < SZINT) { /* real size smaller than lua_Integer? */ if (issigned) { /* needs sign extension? */ lua_Unsigned mask = (lua_Unsigned)1 << (size*NB - 1); res = ((res ^ mask) - mask); /* do sign extension */ } } else if (size > SZINT) { /* must check unread bytes */ int mask = (!issigned || (lua_Integer)res >= 0) ? 0 : MC; for (i = limit; i < size; i++) { if ((unsigned char)str[islittle ? i : size - 1 - i] != mask) luaL_error(L, "%d-byte integer does not fit into Lua Integer", size); } } return (lua_Integer)res; } static int str_unpack (lua_State *L) { Header h; const char *fmt = luaL_checkstring(L, 1); size_t ld; const char *data = luaL_checklstring(L, 2, &ld); size_t pos = (size_t)posrelat(luaL_optinteger(L, 3, 1), ld) - 1; int n = 0; /* number of results */ luaL_argcheck(L, pos <= ld, 3, "initial position out of string"); initheader(L, &h); while (*fmt != '\0') { int size, ntoalign; KOption opt = getdetails(&h, pos, &fmt, &size, &ntoalign); if ((size_t)ntoalign + size > ~pos || pos + ntoalign + size > ld) luaL_argerror(L, 2, "data string too short"); pos += ntoalign; /* skip alignment */ /* stack space for item + next position */ luaL_checkstack(L, 2, "too many results"); n++; switch (opt) { case Kint: case Kuint: { lua_Integer res = unpackint(L, data + pos, h.islittle, size, (opt == Kint)); lua_pushinteger(L, res); break; } case Kfloat: { volatile Ftypes u; lua_Number num; copywithendian(u.buff, data + pos, size, h.islittle); if (size == sizeof(u.f)) num = (lua_Number)u.f; else if (size == sizeof(u.d)) num = (lua_Number)u.d; else num = u.n; lua_pushnumber(L, num); break; } case Kchar: { lua_pushlstring(L, data + pos, size); break; } case Kstring: { size_t len = (size_t)unpackint(L, data + pos, h.islittle, size, 0); luaL_argcheck(L, pos + len + size <= ld, 2, "data string too short"); lua_pushlstring(L, data + pos + size, len); pos += len; /* skip string */ break; } case Kzstr: { size_t len = (int)strlen(data + pos); lua_pushlstring(L, data + pos, len); pos += len + 1; /* skip string plus final '\0' */ break; } case Kpaddalign: case Kpadding: case Knop: n--; /* undo increment */ break; } pos += size; } lua_pushinteger(L, pos + 1); /* next position */ return n + 1; } /* }====================================================== */ static const luaL_Reg strlib[] = { {"byte", str_byte}, {"char", str_char}, {"dump", str_dump}, {"find", str_find}, {"format", str_format}, {"gmatch", gmatch}, {"gsub", str_gsub}, {"len", str_len}, {"lower", str_lower}, {"match", str_match}, {"rep", str_rep}, {"reverse", str_reverse}, {"sub", str_sub}, {"upper", str_upper}, {"pack", str_pack}, {"packsize", str_packsize}, {"unpack", str_unpack}, {NULL, NULL} }; static void createmetatable (lua_State *L) { lua_createtable(L, 0, 1); /* table to be metatable for strings */ lua_pushliteral(L, ""); /* dummy string */ lua_pushvalue(L, -2); /* copy table */ lua_setmetatable(L, -2); /* set table as metatable for strings */ lua_pop(L, 1); /* pop dummy string */ lua_pushvalue(L, -2); /* get string library */ lua_setfield(L, -2, "__index"); /* metatable.__index = string */ lua_pop(L, 1); /* pop metatable */ } /* ** Open string library */ LUAMOD_API int luaopen_string (lua_State *L) { luaL_newlib(L, strlib); createmetatable(L); return 1; }