Optional 'init' argument to 'string.gmatch'

The function 'string.gmatch' now has an optional 'init' argument,
similar to 'string.find' and 'string.match'. Moreover, there was
some reorganization in the manipulation of indices in the string
library.
This commit also includes small janitorial work in the manual
and in comments in the interpreter loop.

lstrlib.c

@@ -60,23 +60,50 @@ static int str_len (lua_State *L) {
 }
 
 
-/* translate a relative string position: negative means back from end */
+/*
-static lua_Integer posrelat (lua_Integer pos, size_t len) {
+** translate a relative initial string position
-  if (pos >= 0) return pos;
+** (negative means back from end): clip result to [1, inf).
-  else if (0u - (size_t)pos > len) return 0;
+** The length of any string in Lua must fit in a lua_Integer,
-  else return (lua_Integer)len + pos + 1;
+** so there are no overflows in the casts.
+** The inverted comparison avoids a possible overflow
+** computing '-pos'.
+*/
+static size_t posrelatI (lua_Integer pos, size_t len) {
+  if (pos > 0)
+    return (size_t)pos;
+  else if (pos == 0)
+    return 1;
+  else if (pos < -(lua_Integer)len)  /* inverted comparison */
+    return 1;  /* clip to 1 */
+  else return len + (size_t)pos + 1;
+}
+
+
+/*
+** Gets an optional ending string position from argument 'arg',
+** with default value 'def'.
+** Negative means back from end: clip result to [0, len]
+*/
+static size_t getendpos (lua_State *L, int arg, lua_Integer def,
+                         size_t len) {
+  lua_Integer pos = luaL_optinteger(L, arg, def);
+  if (pos > (lua_Integer)len)
+    return len;
+  else if (pos >= 0)
+    return (size_t)pos;
+  else if (pos < -(lua_Integer)len)
+    return 0;
+  else return len + (size_t)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);
+  size_t start = posrelatI(luaL_checkinteger(L, 2), l);
-  lua_Integer end = posrelat(luaL_optinteger(L, 3, -1), l);
+  size_t end = getendpos(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);
+    lua_pushlstring(L, s + start - 1, (end - start) + 1);
   else lua_pushliteral(L, "");
   return 1;
 }
@@ -149,11 +176,10 @@ static int str_rep (lua_State *L) {
 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 pi = luaL_optinteger(L, 2, 1);
-  lua_Integer pose = posrelat(luaL_optinteger(L, 3, posi), l);
+  size_t posi = posrelatI(pi, l);
+  size_t pose = getendpos(L, 3, pi, 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");
@@ -171,8 +197,8 @@ static int str_char (lua_State *L) {
   luaL_Buffer b;
   char *p = luaL_buffinitsize(L, &b, n);
   for (i=1; i<=n; i++) {
-    lua_Integer c = luaL_checkinteger(L, i);
+    lua_Unsigned c = (lua_Unsigned)luaL_checkinteger(L, i);
-    luaL_argcheck(L, uchar(c) == c, i, "value out of range");
+    luaL_argcheck(L, c <= (lua_Unsigned)UCHAR_MAX, i, "value out of range");
     p[i - 1] = uchar(c);
   }
   luaL_pushresultsize(&b, n);
@@ -695,16 +721,15 @@ 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);
+  size_t init = posrelatI(luaL_optinteger(L, 3, 1), ls) - 1;
-  if (init < 1) init = 1;
+  if (init > ls) {  /* start after string's end? */
-  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);
+    const char *s2 = lmemfind(s + init, ls - init, p, lp);
     if (s2) {
       lua_pushinteger(L, (s2 - s) + 1);
       lua_pushinteger(L, (s2 - s) + lp);
@@ -713,7 +738,7 @@ static int str_find_aux (lua_State *L, int find) {
   }
   else {
     MatchState ms;
-    const char *s1 = s + init - 1;
+    const char *s1 = s + init;
     int anchor = (*p == '^');
     if (anchor) {
       p++; lp--;  /* skip anchor character */
@@ -777,11 +802,14 @@ static int gmatch (lua_State *L) {
   size_t ls, lp;
   const char *s = luaL_checklstring(L, 1, &ls);
   const char *p = luaL_checklstring(L, 2, &lp);
+  size_t init = posrelatI(luaL_optinteger(L, 3, 1), ls) - 1;
   GMatchState *gm;
-  lua_settop(L, 2);  /* keep them on closure to avoid being collected */
+  lua_settop(L, 2);  /* keep strings on closure to avoid being collected */
   gm = (GMatchState *)lua_newuserdatauv(L, sizeof(GMatchState), 0);
+  if (init > ls)  /* start after string's end? */
+    init = ls + 1;  /* avoid overflows in 's + init' */
   prepstate(&gm->ms, L, s, ls, p, lp);
-  gm->src = s; gm->p = p; gm->lastmatch = NULL;
+  gm->src = s + init; gm->p = p; gm->lastmatch = NULL;
   lua_pushcclosure(L, gmatch_aux, 3);
   return 1;
 }
@@ -1572,7 +1600,7 @@ static int str_unpack (lua_State *L) {
   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;
+  size_t pos = posrelatI(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);

lvm.c

@@ -991,7 +991,8 @@ void luaV_finishOp (lua_State *L) {
 
 /*
 ** Protect code that will finish the loop (returns) or can only raise
-** errors.
+** errors. (That is, it will not return to the interpreter main loop
+** after changing the stack or hooks.)
 */
 #define halfProtect(exp)  (savepc(L), (exp))
 
@@ -1607,7 +1608,7 @@ void luaV_execute (lua_State *L, CallInfo *ci) {
           L->top = ra;
           halfProtect(luaD_poscall(L, ci, 0));  /* no hurry... */
         }
-        else {
+        else {  /* do the 'poscall' here */
           int nres = ci->nresults;
           L->ci = ci->previous;  /* back to caller */
           L->top = base - 1;
@@ -1621,7 +1622,7 @@ void luaV_execute (lua_State *L, CallInfo *ci) {
           L->top = ra + 1;
           halfProtect(luaD_poscall(L, ci, 1));  /* no hurry... */
         }
-        else {
+        else {  /* do the 'poscall' here */
           int nres = ci->nresults;
           L->ci = ci->previous;  /* back to caller */
           if (nres == 0)
@@ -1717,8 +1718,7 @@ void luaV_execute (lua_State *L, CallInfo *ci) {
         vmbreak;
       }
       vmcase(OP_TFORPREP) {
-        /* is 'toclose' not nil? */
+        if (!ttisnil(s2v(ra + 3))) {  /* is 'toclose' not nil? */
-        if (!ttisnil(s2v(ra + 3))) {
           /* create to-be-closed upvalue for it */
           halfProtect(luaF_newtbcupval(L, ra + 3));
         }

manual/manual.of

@@ -83,25 +83,10 @@ it usually represents the absence of a useful value.
 The type @emph{boolean} has two values, @false and @true.
 Both @nil and @false make a condition false;
 any other value makes it true.
-The type @emph{number} represents both
-integer numbers and real (floating-point) numbers.
-The type @emph{string} represents immutable sequences of bytes.
-@index{eight-bit clean}
-Lua is 8-bit clean:
-strings can contain any 8-bit value,
-including @x{embedded zeros} (@Char{\0}).
-Lua is also encoding-agnostic;
-it makes no assumptions about the contents of a string.
 
-The type @emph{number} uses two internal representations,
+The type @emph{number} represents both
-or two @x{subtypes},
+integer numbers and real (floating-point) numbers,
-one called @def{integer} and the other called @def{float}.
+using two @x{subtypes}: @def{integer} and @def{float}.
-Lua has explicit rules about when each representation is used,
-but it also converts between them automatically as needed @see{coercion}.
-Therefore,
-the programmer may choose to mostly ignore the difference
-between integers and floats
-or to assume complete control over the representation of each number.
 Standard Lua uses 64-bit integers and double-precision (64-bit) floats,
 but you can also compile Lua so that it
 uses 32-bit integers and/or single-precision (32-bit) floats.
@@ -110,6 +95,22 @@ is particularly attractive
 for small machines and embedded systems.
 (See macro @id{LUA_32BITS} in file @id{luaconf.h}.)
 
+Lua has explicit rules about when each subtype is used,
+but it also converts between them automatically as needed @see{coercion}.
+Therefore,
+the programmer may choose to mostly ignore the difference
+between integers and floats
+or to assume complete control over the representation of each number.
+
+The type @emph{string} represents immutable sequences of bytes.
+@index{eight-bit clean}
+Lua is 8-bit clean:
+strings can contain any 8-bit value,
+including @x{embedded zeros} (@Char{\0}).
+Lua is also encoding-agnostic;
+it makes no assumptions about the contents of a string.
+The length of any string in Lua must fit in a Lua integer.
+
 Lua can call (and manipulate) functions written in Lua and
 functions written in C @see{functioncall}.
 Both are represented by the type @emph{function}.
@@ -6788,13 +6789,16 @@ the string argument should not contain @x{embedded zeros}.
 
 }
 
-@LibEntry{string.gmatch (s, pattern)|
+@LibEntry{string.gmatch (s, pattern [, init])|
 Returns an iterator function that,
 each time it is called,
 returns the next captures from @id{pattern} @see{pm}
 over the string @id{s}.
 If @id{pattern} specifies no captures,
 then the whole match is produced in each call.
+A third, optional numeric argument @id{init} specifies
+where to start the search;
+its default value @N{is 1} and can be negative.
 
 As an example, the following loop
 will iterate over all the words from string @id{s},

testes/pm.lua

@@ -297,6 +297,35 @@ for k,v in pairs(t) do assert(k+1 == v+0); a=a+1 end
 assert(a == 3)
 
 
+do   -- init parameter in gmatch
+  local s = 0
+  for k in string.gmatch("10 20 30", "%d+", 3) do
+    s = s + tonumber(k)
+  end
+  assert(s == 50)
+
+  s = 0
+  for k in string.gmatch("11 21 31", "%d+", -4) do
+    s = s + tonumber(k)
+  end
+  assert(s == 32)
+
+  -- there is an empty string at the end of the subject
+  s = 0
+  for k in string.gmatch("11 21 31", "%w*", 9) do
+    s = s + 1
+  end
+  assert(s == 1)
+
+  -- there are no empty strings after the end of the subject
+  s = 0
+  for k in string.gmatch("11 21 31", "%w*", 10) do
+    s = s + 1
+  end
+  assert(s == 0)
+end
+
+
 -- tests for `%f' (`frontiers')
 
 assert(string.gsub("aaa aa a aaa a", "%f[%w]a", "x") == "xaa xa x xaa x")

testes/strings.lua

@@ -94,6 +94,11 @@ assert(string.char(string.byte("\xe4l\0
 assert(string.char(string.byte("\xe4l\0óu", 1, 0)) == "")
 assert(string.char(string.byte("\xe4l\0óu", -10, 100)) == "\xe4l\0óu")
 
+checkerror("out of range", string.char, 256)
+checkerror("out of range", string.char, -1)
+checkerror("out of range", string.char, math.maxinteger)
+checkerror("out of range", string.char, math.mininteger)
+
 assert(string.upper("ab\0c") == "AB\0C")
 assert(string.lower("\0ABCc%$") == "\0abcc%$")
 assert(string.rep('teste', 0) == '')

testes/tpack.lua

@@ -314,9 +314,7 @@ do    -- testing initial position
   for i = 1, #x + 1 do
     assert(unpack("c0", x, i) == "")
   end
-  checkerror("out of string", unpack, "c0", x, 0)
   checkerror("out of string", unpack, "c0", x, #x + 2)
-  checkerror("out of string", unpack, "c0", x, -(#x + 1))
  
 end