Struct library patch WAS: Roberto's struct library missing "size" function?

c0 replaced with c<nothing>. Means you can build the format string programatically without treating zero length strings as a special case.
Can handle 'long long' integers (i8 / I8)
Can insert/specify padding anywhere in a struct. ('X' eg. when a string is following a union)
Can report current offset in both pack and unpack ('=')
Can mask out return values when you only want to calculate sizes or unmarshal pascal-style strings. '(' & ')'
Handles doubles with swapped hi/lo words (Seen on ARM platform)
Can unpack arbitrary userdata,length datums (like alien version of struct)

** Valid formats:

** > - big endian

** < - little endian

** ![num] - alignment

** x[num] - pad num bytes, default 1

** X[num] - pad to num align, default MAXALIGN

** b/B - signed/unsigned byte

** h/H - signed/unsigned short

** l/L - signed/unsigned long

** i/I[num] - signed/unsigned integer with size `n' (default is size of int)

** c[num] - sequence of `num' chars (from/to a string); when packing, num

absent means the whole string; when unpacking, num absent means

use the previous read number as the string length.

** s - zero terminated string

** f - float

** d - double

** ' ' - ignored

** '(' ')' - stop assigning items. ')' start assigning (padding when packing)

** '=' - return current position / offset

/Flemming

--- struct.orig 2009-10-15 08:30:59.000000000 +0200 +++ struct.c 2009-10-15 09:45:31.000000000 +0200 @@ -1,37 +1,43 @@ - #include <assert.h> #include <ctype.h> #include <limits.h> #include <string.h> +#include <stdbool.h> #include "lua.h" #include "lauxlib.h" +#ifndef LLONG_MAX +#define LLONG_MAX 9223372036854775807LL +#endif /* ** {====================================================== ** Library for packing/unpacking structures. -** $Id: struct.c,v 1.2 2008/04/18 20:06:01 roberto Exp $ +** $Id: struct.c 20405 2009-10-15 07:06:45Z fm $ ** ======================================================= */ /* ** Valid formats: -** > - big endian -** < - little endian -** ![num] - alignment -** x - pading -** b/B - signed/unsigned byte -** h/H - signed/unsigned short -** l/L - signed/unsigned long -** i/In - signed/unsigned integer with size `n' (default is size of int) -** cn - sequence of `n' chars (from/to a string); when packing, n==0 means - the whole string; when unpacking, n==0 means use the previous - read number as the string length -** s - zero-terminated string -** f - float -** d - doulbe -** ' ' - ignored +** > - big endian +** < - little endian +** ![num] - alignment +** x[num] - pad num bytes, default 1 +** X[num] - pad to num align, default MAXALIGN +** b/B - signed/unsigned byte +** h/H - signed/unsigned short +** l/L - signed/unsigned long +** i/I[num] - signed/unsigned integer with size `n' (default is size of int) +** c[num] - sequence of `num' chars (from/to a string); when packing, num + absent means the whole string; when unpacking, num absent means + use the previous read number as the string length. +** s - zero terminated string +** f - float +** d - double +** ' ' - ignored +** '(' ')' - stop assigning items. ')' start assigning (padding when packing) +** '=' - return current position / offset */ @@ -59,10 +65,18 @@ char endian; } const native = {1}; +static union dblswap { + long long dummy; + double dbl; + long l[2]; +} const swaptest = {0x00000000c0000000}; + typedef struct Header { int endian; int align; + bool noassign; + bool dblswap; } Header; @@ -79,7 +93,8 @@ } -#define defaultoptions(h) ((h)->endian = native.endian, (h)->align = 1) +#define defaultoptions(h) ((h)->endian = native.endian, (h)->align = 1, (h)->noassign = 0, \ + (h)->dblswap = false) @@ -90,9 +105,18 @@ case 'L': case 'l': return sizeof(long); case 'f': return sizeof(float); case 'd': return sizeof(double); - case 'x': return 1; - case 'c': return getnum(fmt, 1); - case 's': case ' ': case '<': case '>': case '!': return 0; + case 'x': return getnum(fmt, 1); + case 'X': return getnum(fmt, MAXALIGN); + case 'c': return getnum(fmt, 0); + case 's': + case ' ': + case '<': + case '>': + case '(': + case ')': + case '!': + case '=': + return 0; case 'i': case 'I': { int sz = getnum(fmt, sizeof(int)); if (!isp2(sz)) @@ -108,7 +132,7 @@ static int gettoalign (size_t len, Header *h, int opt, size_t size) { - if (size == 0 || opt == 'c') return 0; + if (size == 0 || opt == 'c' || opt == 's') return 0; if (size > (size_t)h->align) size = h->align; /* respect max. alignment */ return (size - (len & (size - 1))) & (size - 1); } @@ -117,8 +141,10 @@ static void commoncases (lua_State *L, int opt, const char **fmt, Header *h) { switch (opt) { case ' ': return; /* ignore white spaces */ - case '>': h->endian = BIG; return; - case '<': h->endian = LITTLE; return; + case '>': h->endian = BIG; (h)->dblswap = (swaptest.dbl == -2); return; + case '<': h->endian = LITTLE; (h)->dblswap = (swaptest.dbl == -2); return; + case '(': h->noassign = true; return; + case ')': h->noassign = false; return; case '!': { int a = getnum(fmt, MAXALIGN); if (!isp2(a)) @@ -134,11 +160,11 @@ static void putinteger (lua_State *L, luaL_Buffer *b, int arg, int endian, int size) { lua_Number n = luaL_checknumber(L, arg); - unsigned long value; - if (n < (lua_Number)LONG_MAX) - value = (long)n; + unsigned long long value; + if (n < (lua_Number)LLONG_MAX) + value = (long long)n; else - value = (unsigned long)n; + value = (unsigned long long)n; if (endian == LITTLE) { int i; for (i = 0; i < size; i++) @@ -168,6 +194,8 @@ luaL_Buffer b; const char *fmt = luaL_checkstring(L, 1); Header h; + int poscnt = 0; + int posBuf[10]; int arg = 2; size_t totalsize = 0; defaultoptions(&h); @@ -179,14 +207,19 @@ int toalign = gettoalign(totalsize, &h, opt, size); totalsize += toalign; while (toalign-- > 0) luaL_putchar(&b, '\0'); + if (opt == 'X') + size = 0; + if (h.noassign && size) + opt = 'x'; switch (opt) { case 'b': case 'B': case 'h': case 'H': case 'l': case 'L': case 'i': case 'I': { /* integer types */ putinteger(L, &b, arg++, h.endian, size); break; } - case 'x': { - luaL_putchar(&b, '\0'); + case 'x': case 'X': { + size_t l = size; + while (l-- > 0) luaL_putchar(&b, '\0'); break; } case 'f': { @@ -196,15 +229,21 @@ break; } case 'd': { - double d = luaL_checknumber(L, arg++); + union dblswap d; + d.dbl = luaL_checknumber(L, arg++); correctbytes((char *)&d, size, h.endian); + if (h.dblswap) { + long tmp = d.l[0]; + d.l[0] = d.l[1]; + d.l[1] = tmp; + } luaL_addlstring(&b, (char *)&d, size); break; } case 'c': case 's': { size_t l; const char *s = luaL_checklstring(L, arg++, &l); - if (size == 0) size = l; + if (opt == 's' || (opt == 'c' && fmt[-1] == 'c')) size = l; luaL_argcheck(L, l >= (size_t)size, arg, "string too short"); luaL_addlstring(&b, s, size); if (opt == 's') { @@ -213,12 +252,19 @@ } break; } + case '=': { + if (poscnt < sizeof(posBuf)/sizeof(posBuf[0])) + posBuf[poscnt++] = totalsize + 1; + break; + } default: commoncases(L, opt, &fmt, &h); } totalsize += size; } luaL_pushresult(&b); - return 1; + for (arg = 0; arg < poscnt; arg++) + lua_pushinteger(L, posBuf[arg]); + return poscnt + 1; } @@ -250,8 +296,24 @@ Header h; const char *fmt = luaL_checkstring(L, 1); size_t ld; - const char *data = luaL_checklstring(L, 2, &ld); - size_t pos = luaL_optinteger(L, 3, 1) - 1; + const char *data; + size_t pos; + lua_Number lastnum = 0; + + void pushnumber(lua_Number n) { + lastnum = n; + if (h.noassign) return; + lua_pushnumber(L, n); + } + + if (lua_isuserdata(L, 2)) { + data = (const char*)lua_touserdata(L, 2); + ld = (size_t)luaL_checkinteger(L, 3); + pos = luaL_optinteger(L, 4, 1) - 1; + } else { + data = luaL_checklstring(L, 2, &ld); + pos = luaL_optinteger(L, 3, 1) - 1; + } defaultoptions(&h); lua_settop(L, 2); while (*fmt) { @@ -259,40 +321,45 @@ size_t size = optsize(L, opt, &fmt); pos += gettoalign(pos, &h, opt, size); luaL_argcheck(L, pos+size <= ld, 2, "data string too short"); + if (opt == 'X') + size = 0; switch (opt) { case 'b': case 'B': case 'h': case 'H': case 'l': case 'L': case 'i': case 'I': { /* integer types */ int issigned = islower(opt); lua_Number res = getinteger(data+pos, h.endian, issigned, size); - lua_pushnumber(L, res); + pushnumber(res); break; } - case 'x': { + case 'x': case 'X': { break; } case 'f': { float f; memcpy(&f, data+pos, size); correctbytes((char *)&f, sizeof(f), h.endian); - lua_pushnumber(L, f); + pushnumber(f); break; } case 'd': { - double d; + union dblswap d; memcpy(&d, data+pos, size); correctbytes((char *)&d, sizeof(d), h.endian); - lua_pushnumber(L, d); + if (h.dblswap) { + long tmp = d.l[0]; + d.l[0] = d.l[1]; + d.l[1] = tmp; + } + pushnumber(d.dbl); break; } case 'c': { - if (size == 0) { - if (!lua_isnumber(L, -1)) - luaL_error(L, "format `c0' needs a previous size"); - size = lua_tonumber(L, -1); - lua_pop(L, 1); + if (fmt[-1] == 'c') { + size = lastnum; luaL_argcheck(L, pos+size <= ld, 2, "data string too short"); } - lua_pushlstring(L, data+pos, size); + if (!h.noassign) + lua_pushlstring(L, data+pos, size); break; } case 's': { @@ -300,7 +367,12 @@ if (e == NULL) luaL_error(L, "unfinished string in data"); size = (e - (data+pos)) + 1; - lua_pushlstring(L, data+pos, size - 1); + if (!h.noassign) + lua_pushlstring(L, data+pos, size - 1); + break; + } + case '=': { + lua_pushinteger(L, pos + 1); break; } default: commoncases(L, opt, &fmt, &h);