Added lpeg for narrator

src/lua/lpeg-1.1.0/HISTORY

@@ -0,0 +1,109 @@
+HISTORY for LPeg 1.1.0
+
+* Changes from version 1.0.2 to 1.1.0
+  ---------------------------------
+  + accumulator capture
+  + UTF-8 ranges
+  + Larger limit for number of rules in a grammar
+  + Larger limit for number of captures in a match
+  + bug fixes
+  + other small improvements
+
+* Changes from version 1.0.1 to 1.0.2
+  ---------------------------------
+  + some bugs fixed
+
+* Changes from version 0.12 to 1.0.1
+  ---------------------------------
+  + group "names" can be any Lua value
+  + some bugs fixed
+  + other small improvements
+
+* Changes from version 0.11 to 0.12
+  ---------------------------------
+  + no "unsigned short" limit for pattern sizes
+  + mathtime captures considered nullable
+  + some bugs fixed
+
+* Changes from version 0.10 to 0.11
+  -------------------------------
+  + complete reimplementation of the code generator
+  + new syntax for table captures
+  + new functions in module 're'
+  + other small improvements
+
+* Changes from version 0.9 to 0.10
+  -------------------------------
+  + backtrack stack has configurable size
+  + better error messages
+  + Notation for non-terminals in 're' back to A instead o <A>
+  + experimental look-behind pattern
+  + support for external extensions
+  + works with Lua 5.2
+  + consumes less C stack
+
+  - "and" predicates do not keep captures
+
+* Changes from version 0.8 to 0.9
+  -------------------------------
+  + The accumulator capture was replaced by a fold capture;
+    programs that used the old 'lpeg.Ca' will need small changes.
+  + Some support for character classes from old C locales.
+  + A new named-group capture.
+
+* Changes from version 0.7 to 0.8
+  -------------------------------
+  + New "match-time" capture.
+  + New "argument capture" that allows passing arguments into the pattern.
+  + Better documentation for 're'.
+  + Several small improvements for 're'.
+  + The 're' module has an incompatibility with previous versions:
+    now, any use of a non-terminal must be enclosed in angle brackets
+    (like <B>).
+
+* Changes from version 0.6 to 0.7
+  -------------------------------
+  + Several improvements in module 're':
+    - better documentation;
+    - support for most captures (all but accumulator);
+    - limited repetitions p{n,m}.
+  + Small improvements in efficiency.
+  + Several small bugs corrected (special thanks to Hans Hagen
+    and Taco Hoekwater).
+
+* Changes from version 0.5 to 0.6
+  -------------------------------
+  + Support for non-numeric indices in grammars.
+  + Some bug fixes (thanks to the luatex team).
+  + Some new optimizations; (thanks to Mike Pall).
+  + A new page layout (thanks to Andre Carregal).
+  + Minimal documentation for module 're'.
+
+* Changes from version 0.4 to 0.5
+  -------------------------------
+  + Several optimizations.
+  + lpeg.P now accepts booleans.
+  + Some new examples.
+  + A proper license.
+  + Several small improvements.
+
+* Changes from version 0.3 to 0.4
+  -------------------------------
+  + Static check for loops in repetitions and grammars.
+  + Removed label option in captures.
+  + The implementation of captures uses less memory.
+
+* Changes from version 0.2 to 0.3
+  -------------------------------
+  + User-defined patterns in Lua.
+  + Several new captures.
+
+* Changes from version 0.1 to 0.2
+  -------------------------------
+  + Several small corrections.
+  + Handles embedded zeros like any other character.
+  + Capture "name" can be any Lua value.
+  + Unlimited number of captures.
+  + Match gets an optional initial position.
+
+(end of HISTORY)

src/lua/lpeg-1.1.0/README.md

@@ -0,0 +1,4 @@
+# LPeg - Parsing Expression Grammars For Lua
+
+For more information,
+see [Lpeg](//www.inf.puc-rio.br/~roberto/lpeg/).

src/lua/lpeg-1.1.0/lpcap.c

@@ -0,0 +1,612 @@
+
+#include "lua.h"
+#include "lauxlib.h"
+
+#include "lpcap.h"
+#include "lpprint.h"
+#include "lptypes.h"
+
+
+#define getfromktable(cs,v)	lua_rawgeti((cs)->L, ktableidx((cs)->ptop), v)
+
+#define pushluaval(cs)		getfromktable(cs, (cs)->cap->idx)
+
+
+
+#define skipclose(cs,head)  \
+	if (isopencap(head)) { assert(isclosecap(cs->cap)); cs->cap++; }
+
+
+/*
+** Return the size of capture 'cap'. If it is an open capture, 'close'
+** must be its corresponding close.
+*/
+static Index_t capsize (Capture *cap, Capture *close) {
+  if (isopencap(cap)) {
+    assert(isclosecap(close));
+    return close->index - cap->index;
+  }
+  else
+    return cap->siz - 1;
+}
+
+
+static Index_t closesize (CapState *cs, Capture *head) {
+  return capsize(head, cs->cap);
+}
+
+
+/*
+** Put at the cache for Lua values the value indexed by 'v' in ktable
+** of the running pattern (if it is not there yet); returns its index.
+*/
+static int updatecache (CapState *cs, int v) {
+  int idx = cs->ptop + 1;  /* stack index of cache for Lua values */
+  if (v != cs->valuecached) {  /* not there? */
+    getfromktable(cs, v);  /* get value from 'ktable' */
+    lua_replace(cs->L, idx);  /* put it at reserved stack position */
+    cs->valuecached = v;  /* keep track of what is there */
+  }
+  return idx;
+}
+
+
+static int pushcapture (CapState *cs);
+
+
+/*
+** Goes back in a list of captures looking for an open capture
+** corresponding to a close one.
+*/
+static Capture *findopen (Capture *cap) {
+  int n = 0;  /* number of closes waiting an open */
+  for (;;) {
+    cap--;
+    if (isclosecap(cap)) n++;  /* one more open to skip */
+    else if (isopencap(cap))
+      if (n-- == 0) return cap;
+  }
+}
+
+
+/*
+** Go to the next capture at the same level.
+*/
+static void nextcap (CapState *cs) {
+  Capture *cap = cs->cap;
+  if (isopencap(cap)) {  /* must look for a close? */
+    int n = 0;  /* number of opens waiting a close */
+    for (;;) {  /* look for corresponding close */
+      cap++;
+      if (isopencap(cap)) n++;
+      else if (isclosecap(cap))
+        if (n-- == 0) break;
+    }
+    cs->cap = cap + 1;  /* + 1 to skip last close */
+  }
+  else {
+    Capture *next;
+    for (next = cap + 1; capinside(cap, next); next++)
+      ;  /* skip captures inside current one */
+    cs->cap = next;
+  }
+}
+
+
+/*
+** Push on the Lua stack all values generated by nested captures inside
+** the current capture. Returns number of values pushed. 'addextra'
+** makes it push the entire match after all captured values. The
+** entire match is pushed also if there are no other nested values,
+** so the function never returns zero.
+*/
+static int pushnestedvalues (CapState *cs, int addextra) {
+  Capture *head = cs->cap++;  /* original capture */
+  int n = 0;  /* number of pushed subvalues */
+  /* repeat for all nested patterns */
+  while (capinside(head, cs->cap))
+    n += pushcapture(cs);
+  if (addextra || n == 0) {  /* need extra? */
+    lua_pushlstring(cs->L, cs->s + head->index, closesize(cs, head));
+    n++;
+  }
+  skipclose(cs, head);
+  return n;
+}
+
+
+/*
+** Push only the first value generated by nested captures
+*/
+static void pushonenestedvalue (CapState *cs) {
+  int n = pushnestedvalues(cs, 0);
+  if (n > 1)
+    lua_pop(cs->L, n - 1);  /* pop extra values */
+}
+
+
+/*
+** Checks whether group 'grp' is visible to 'ref', that is, 'grp' is
+** not nested inside a full capture that does not contain 'ref'.  (We
+** only need to care for full captures because the search at 'findback'
+** skips open-end blocks; so, if 'grp' is nested in a non-full capture,
+** 'ref' is also inside it.)  To check this, we search backward for the
+** inner full capture enclosing 'grp'.  A full capture cannot contain
+** non-full captures, so a close capture means we cannot be inside a
+** full capture anymore.
+*/
+static int capvisible (CapState *cs, Capture *grp, Capture *ref) {
+  Capture *cap = grp;
+  int i = MAXLOP;  /* maximum distance for an 'open' */
+  while (i-- > 0 && cap-- > cs->ocap) {
+    if (isclosecap(cap))
+      return 1;  /* can stop the search */
+    else if (grp->index - cap->index >= UCHAR_MAX)
+      return 1;  /* can stop the search */
+    else if (capinside(cap, grp))  /* is 'grp' inside cap? */
+      return capinside(cap, ref);  /* ok iff cap also contains 'ref' */
+  }
+  return 1;  /* 'grp' is not inside any capture */
+}
+
+
+/*
+** Try to find a named group capture with the name given at the top of
+** the stack; goes backward from 'ref'.
+*/
+static Capture *findback (CapState *cs, Capture *ref) {
+  lua_State *L = cs->L;
+  Capture *cap = ref;
+  while (cap-- > cs->ocap) {  /* repeat until end of list */
+    if (isclosecap(cap))
+      cap = findopen(cap);  /* skip nested captures */
+    else if (capinside(cap, ref))
+      continue;  /* enclosing captures are not visible to 'ref' */
+    if (captype(cap) == Cgroup && capvisible(cs, cap, ref)) {
+      getfromktable(cs, cap->idx);  /* get group name */
+      if (lp_equal(L, -2, -1)) {  /* right group? */
+        lua_pop(L, 2);  /* remove reference name and group name */
+        return cap;
+      }
+      else lua_pop(L, 1);  /* remove group name */
+    }
+  }
+  luaL_error(L, "back reference '%s' not found", lua_tostring(L, -1));
+  return NULL;  /* to avoid warnings */
+}
+
+
+/*
+** Back-reference capture. Return number of values pushed.
+*/
+static int backrefcap (CapState *cs) {
+  int n;
+  Capture *curr = cs->cap;
+  pushluaval(cs);  /* reference name */
+  cs->cap = findback(cs, curr);  /* find corresponding group */
+  n = pushnestedvalues(cs, 0);  /* push group's values */
+  cs->cap = curr + 1;
+  return n;
+}
+
+
+/*
+** Table capture: creates a new table and populates it with nested
+** captures.
+*/
+static int tablecap (CapState *cs) {
+  lua_State *L = cs->L;
+  Capture *head = cs->cap++;
+  int n = 0;
+  lua_newtable(L);
+  while (capinside(head, cs->cap)) {
+    if (captype(cs->cap) == Cgroup && cs->cap->idx != 0) {  /* named group? */
+      pushluaval(cs);  /* push group name */
+      pushonenestedvalue(cs);
+      lua_settable(L, -3);
+    }
+    else {  /* not a named group */
+      int i;
+      int k = pushcapture(cs);
+      for (i = k; i > 0; i--)  /* store all values into table */
+        lua_rawseti(L, -(i + 1), n + i);
+      n += k;
+    }
+  }
+  skipclose(cs, head);
+  return 1;  /* number of values pushed (only the table) */
+}
+
+
+/*
+** Table-query capture
+*/
+static int querycap (CapState *cs) {
+  int idx = cs->cap->idx;
+  pushonenestedvalue(cs);  /* get nested capture */
+  lua_gettable(cs->L, updatecache(cs, idx));  /* query cap. value at table */
+  if (!lua_isnil(cs->L, -1))
+    return 1;
+  else {  /* no value */
+    lua_pop(cs->L, 1);  /* remove nil */
+    return 0;
+  }
+}
+
+
+/*
+** Fold capture
+*/
+static int foldcap (CapState *cs) {
+  int n;
+  lua_State *L = cs->L;
+  Capture *head = cs->cap++;
+  int idx = head->idx;
+  if (isclosecap(cs->cap) ||  /* no nested captures (large subject)? */
+      (n = pushcapture(cs)) == 0)  /* nested captures with no values? */
+    return luaL_error(L, "no initial value for fold capture");
+  if (n > 1)
+    lua_pop(L, n - 1);  /* leave only one result for accumulator */
+  while (capinside(head, cs->cap)) {
+    lua_pushvalue(L, updatecache(cs, idx));  /* get folding function */
+    lua_insert(L, -2);  /* put it before accumulator */
+    n = pushcapture(cs);  /* get next capture's values */
+    lua_call(L, n + 1, 1);  /* call folding function */
+  }
+  skipclose(cs, head);
+  return 1;  /* only accumulator left on the stack */
+}
+
+
+/*
+** Function capture
+*/
+static int functioncap (CapState *cs) {
+  int n;
+  int top = lua_gettop(cs->L);
+  pushluaval(cs);  /* push function */
+  n = pushnestedvalues(cs, 0);  /* push nested captures */
+  lua_call(cs->L, n, LUA_MULTRET);  /* call function */
+  return lua_gettop(cs->L) - top;  /* return function's results */
+}
+
+
+/*
+** Accumulator capture
+*/
+static int accumulatorcap (CapState *cs) {
+  lua_State *L = cs->L;
+  int n;
+  if (lua_gettop(L) < cs->firstcap)
+    luaL_error(L, "no previous value for accumulator capture");
+  pushluaval(cs);  /* push function */
+  lua_insert(L, -2);  /* previous value becomes first argument */
+  n = pushnestedvalues(cs, 0);  /* push nested captures */
+  lua_call(L, n + 1, 1);  /* call function */
+  return 0;  /* did not add any extra value */
+}
+
+
+/*
+** Select capture
+*/
+static int numcap (CapState *cs) {
+  int idx = cs->cap->idx;  /* value to select */
+  if (idx == 0) {  /* no values? */
+    nextcap(cs);  /* skip entire capture */
+    return 0;  /* no value produced */
+  }
+  else {
+    int n = pushnestedvalues(cs, 0);
+    if (n < idx)  /* invalid index? */
+      return luaL_error(cs->L, "no capture '%d'", idx);
+    else {
+      lua_pushvalue(cs->L, -(n - idx + 1));  /* get selected capture */
+      lua_replace(cs->L, -(n + 1));  /* put it in place of 1st capture */
+      lua_pop(cs->L, n - 1);  /* remove other captures */
+      return 1;
+    }
+  }
+}
+
+
+/*
+** Return the stack index of the first runtime capture in the given
+** list of captures (or zero if no runtime captures)
+*/
+int finddyncap (Capture *cap, Capture *last) {
+  for (; cap < last; cap++) {
+    if (cap->kind == Cruntime)
+      return cap->idx;  /* stack position of first capture */
+  }
+  return 0;  /* no dynamic captures in this segment */
+}
+
+
+/*
+** Calls a runtime capture. Returns number of captures "removed" by the
+** call, that is, those inside the group capture. Captures to be added
+** are on the Lua stack.
+*/
+int runtimecap (CapState *cs, Capture *close, const char *s, int *rem) {
+  int n, id;
+  lua_State *L = cs->L;
+  int otop = lua_gettop(L);
+  Capture *open = findopen(close);  /* get open group capture */
+  assert(captype(open) == Cgroup);
+  id = finddyncap(open, close);  /* get first dynamic capture argument */
+  close->kind = Cclose;  /* closes the group */
+  close->index = s - cs->s;
+  cs->cap = open; cs->valuecached = 0;  /* prepare capture state */
+  luaL_checkstack(L, 4, "too many runtime captures");
+  pushluaval(cs);  /* push function to be called */
+  lua_pushvalue(L, SUBJIDX);  /* push original subject */
+  lua_pushinteger(L, s - cs->s + 1);  /* push current position */
+  n = pushnestedvalues(cs, 0);  /* push nested captures */
+  lua_call(L, n + 2, LUA_MULTRET);  /* call dynamic function */
+  if (id > 0) {  /* are there old dynamic captures to be removed? */
+    int i;
+    for (i = id; i <= otop; i++)
+      lua_remove(L, id);  /* remove old dynamic captures */
+    *rem = otop - id + 1;  /* total number of dynamic captures removed */
+  }
+  else
+    *rem = 0;  /* no dynamic captures removed */
+  return close - open - 1;  /* number of captures to be removed */
+}
+
+
+/*
+** Auxiliary structure for substitution and string captures: keep
+** information about nested captures for future use, avoiding to push
+** string results into Lua
+*/
+typedef struct StrAux {
+  int isstring;  /* whether capture is a string */
+  union {
+    Capture *cp;  /* if not a string, respective capture */
+    struct {  /* if it is a string... */
+      Index_t idx;  /* starts here */
+      Index_t siz;  /* with this size */
+    } s;
+  } u;
+} StrAux;
+
+#define MAXSTRCAPS	10
+
+/*
+** Collect values from current capture into array 'cps'. Current
+** capture must be Cstring (first call) or Csimple (recursive calls).
+** (In first call, fills %0 with whole match for Cstring.)
+** Returns number of elements in the array that were filled.
+*/
+static int getstrcaps (CapState *cs, StrAux *cps, int n) {
+  int k = n++;
+  Capture *head = cs->cap++;
+  cps[k].isstring = 1;  /* get string value */
+  cps[k].u.s.idx = head->index;  /* starts here */
+  while (capinside(head, cs->cap)) {
+    if (n >= MAXSTRCAPS)  /* too many captures? */
+      nextcap(cs);  /* skip extra captures (will not need them) */
+    else if (captype(cs->cap) == Csimple)  /* string? */
+      n = getstrcaps(cs, cps, n);  /* put info. into array */
+    else {
+      cps[n].isstring = 0;  /* not a string */
+      cps[n].u.cp = cs->cap;  /* keep original capture */
+      nextcap(cs);
+      n++;
+    }
+  }
+  cps[k].u.s.siz = closesize(cs, head);
+  skipclose(cs, head);
+  return n;
+}
+
+
+/*
+** add next capture value (which should be a string) to buffer 'b'
+*/
+static int addonestring (luaL_Buffer *b, CapState *cs, const char *what);
+
+
+/*
+** String capture: add result to buffer 'b' (instead of pushing
+** it into the stack)
+*/
+static void stringcap (luaL_Buffer *b, CapState *cs) {
+  StrAux cps[MAXSTRCAPS];
+  int n;
+  size_t len, i;
+  const char *fmt;  /* format string */
+  fmt = lua_tolstring(cs->L, updatecache(cs, cs->cap->idx), &len);
+  n = getstrcaps(cs, cps, 0) - 1;  /* collect nested captures */
+  for (i = 0; i < len; i++) {  /* traverse format string */
+    if (fmt[i] != '%')  /* not an escape? */
+      luaL_addchar(b, fmt[i]);  /* add it to buffer */
+    else if (fmt[++i] < '0' || fmt[i] > '9')  /* not followed by a digit? */
+      luaL_addchar(b, fmt[i]);  /* add to buffer */
+    else {
+      int l = fmt[i] - '0';  /* capture index */
+      if (l > n)
+        luaL_error(cs->L, "invalid capture index (%d)", l);
+      else if (cps[l].isstring)
+        luaL_addlstring(b, cs->s + cps[l].u.s.idx, cps[l].u.s.siz);
+      else {
+        Capture *curr = cs->cap;
+        cs->cap = cps[l].u.cp;  /* go back to evaluate that nested capture */
+        if (!addonestring(b, cs, "capture"))
+          luaL_error(cs->L, "no values in capture index %d", l);
+        cs->cap = curr;  /* continue from where it stopped */
+      }
+    }
+  }
+}
+
+
+/*
+** Substitution capture: add result to buffer 'b'
+*/
+static void substcap (luaL_Buffer *b, CapState *cs) {
+  const char *curr = cs->s + cs->cap->index;
+  Capture *head = cs->cap++;
+  while (capinside(head, cs->cap)) {
+    Capture *cap = cs->cap;
+    const char *caps = cs->s + cap->index;
+    luaL_addlstring(b, curr, caps - curr);  /* add text up to capture */
+    if (addonestring(b, cs, "replacement"))
+      curr = caps + capsize(cap, cs->cap - 1);  /* continue after match */
+    else  /* no capture value */
+      curr = caps;  /* keep original text in final result */
+  }
+  /* add last piece of text */
+  luaL_addlstring(b, curr, cs->s + head->index + closesize(cs, head) - curr);
+  skipclose(cs, head);
+}
+
+
+/*
+** Evaluates a capture and adds its first value to buffer 'b'; returns
+** whether there was a value
+*/
+static int addonestring (luaL_Buffer *b, CapState *cs, const char *what) {
+  switch (captype(cs->cap)) {
+    case Cstring:
+      stringcap(b, cs);  /* add capture directly to buffer */
+      return 1;
+    case Csubst:
+      substcap(b, cs);  /* add capture directly to buffer */
+      return 1;
+    case Cacc:  /* accumulator capture? */
+      return luaL_error(cs->L, "invalid context for an accumulator capture");
+    default: {
+      lua_State *L = cs->L;
+      int n = pushcapture(cs);
+      if (n > 0) {
+        if (n > 1) lua_pop(L, n - 1);  /* only one result */
+        if (!lua_isstring(L, -1))
+          return luaL_error(L, "invalid %s value (a %s)",
+                               what, luaL_typename(L, -1));
+        luaL_addvalue(b);
+      }
+      return n;
+    }
+  }
+}
+
+
+#if !defined(MAXRECLEVEL)
+#define MAXRECLEVEL	200
+#endif
+
+
+/*
+** Push all values of the current capture into the stack; returns
+** number of values pushed
+*/
+static int pushcapture (CapState *cs) {
+  lua_State *L = cs->L;
+  int res;
+  luaL_checkstack(L, 4, "too many captures");
+  if (cs->reclevel++ > MAXRECLEVEL)
+    return luaL_error(L, "subcapture nesting too deep");
+  switch (captype(cs->cap)) {
+    case Cposition: {
+      lua_pushinteger(L, cs->cap->index + 1);
+      cs->cap++;
+      res = 1;
+      break;
+    }
+    case Cconst: {
+      pushluaval(cs);
+      cs->cap++;
+      res = 1;
+      break;
+    }
+    case Carg: {
+      int arg = (cs->cap++)->idx;
+      if (arg + FIXEDARGS > cs->ptop)
+        return luaL_error(L, "reference to absent extra argument #%d", arg);
+      lua_pushvalue(L, arg + FIXEDARGS);
+      res = 1;
+      break;
+    }
+    case Csimple: {
+      int k = pushnestedvalues(cs, 1);
+      lua_insert(L, -k);  /* make whole match be first result */
+      res = k;
+      break;
+    }
+    case Cruntime: {
+      lua_pushvalue(L, (cs->cap++)->idx);  /* value is in the stack */
+      res = 1;
+      break;
+    }
+    case Cstring: {
+      luaL_Buffer b;
+      luaL_buffinit(L, &b);
+      stringcap(&b, cs);
+      luaL_pushresult(&b);
+      res = 1;
+      break;
+    }
+    case Csubst: {
+      luaL_Buffer b;
+      luaL_buffinit(L, &b);
+      substcap(&b, cs);
+      luaL_pushresult(&b);
+      res = 1;
+      break;
+    }
+    case Cgroup: {
+      if (cs->cap->idx == 0)  /* anonymous group? */
+        res = pushnestedvalues(cs, 0);  /* add all nested values */
+      else {  /* named group: add no values */
+        nextcap(cs);  /* skip capture */
+        res = 0;
+      }
+      break;
+    }
+    case Cbackref: res = backrefcap(cs); break;
+    case Ctable: res = tablecap(cs); break;
+    case Cfunction: res = functioncap(cs); break;
+    case Cacc: res = accumulatorcap(cs); break;
+    case Cnum: res = numcap(cs); break;
+    case Cquery: res = querycap(cs); break;
+    case Cfold: res = foldcap(cs); break;
+    default: assert(0); res = 0;
+  }
+  cs->reclevel--;
+  return res;
+}
+
+
+/*
+** Prepare a CapState structure and traverse the entire list of
+** captures in the stack pushing its results. 's' is the subject
+** string, 'r' is the final position of the match, and 'ptop'
+** the index in the stack where some useful values were pushed.
+** Returns the number of results pushed. (If the list produces no
+** results, push the final position of the match.)
+*/
+int getcaptures (lua_State *L, const char *s, const char *r, int ptop) {
+  Capture *capture = (Capture *)lua_touserdata(L, caplistidx(ptop));
+  int n = 0;
+  /* printcaplist(capture); */
+  if (!isclosecap(capture)) {  /* is there any capture? */
+    CapState cs;
+    cs.ocap = cs.cap = capture; cs.L = L; cs.reclevel = 0;
+    cs.s = s; cs.valuecached = 0; cs.ptop = ptop;
+    cs.firstcap = lua_gettop(L) + 1;  /* where first value (if any) will go */
+    do {  /* collect their values */
+      n += pushcapture(&cs);
+    } while (!isclosecap(cs.cap));
+    assert(lua_gettop(L) - cs.firstcap == n - 1);
+  }
+  if (n == 0) {  /* no capture values? */
+    lua_pushinteger(L, r - s + 1);  /* return only end position */
+    n = 1;
+  }
+  return n;
+}
+
+

src/lua/lpeg-1.1.0/lpcap.h

@@ -0,0 +1,86 @@
+
+#if !defined(lpcap_h)
+#define lpcap_h
+
+
+#include "lptypes.h"
+
+
+/* kinds of captures */
+typedef enum CapKind {
+  Cclose,  /* not used in trees */
+  Cposition,
+  Cconst,  /* ktable[key] is Lua constant */
+  Cbackref,  /* ktable[key] is "name" of group to get capture */
+  Carg,  /* 'key' is arg's number */
+  Csimple,  /* next node is pattern */
+  Ctable,  /* next node is pattern */
+  Cfunction,  /* ktable[key] is function; next node is pattern */
+  Cacc,  /* ktable[key] is function; next node is pattern */
+  Cquery,  /* ktable[key] is table; next node is pattern */
+  Cstring,  /* ktable[key] is string; next node is pattern */
+  Cnum,  /* numbered capture; 'key' is number of value to return */
+  Csubst,  /* substitution capture; next node is pattern */
+  Cfold,  /* ktable[key] is function; next node is pattern */
+  Cruntime,  /* not used in trees (is uses another type for tree) */
+  Cgroup  /* ktable[key] is group's "name" */
+} CapKind;
+
+
+/*
+** An unsigned integer large enough to index any subject entirely.
+** It can be size_t, but that will double the size of the array
+** of captures in a 64-bit machine.
+*/
+#if !defined(Index_t)
+typedef uint Index_t;
+#endif
+
+#define MAXINDT		(~(Index_t)0)
+
+
+typedef struct Capture {
+  Index_t index;  /* subject position */
+  unsigned short idx;  /* extra info (group name, arg index, etc.) */
+  byte kind;  /* kind of capture */
+  byte siz;  /* size of full capture + 1 (0 = not a full capture) */
+} Capture;
+
+
+typedef struct CapState {
+  Capture *cap;  /* current capture */
+  Capture *ocap;  /* (original) capture list */
+  lua_State *L;
+  int ptop;  /* stack index of last argument to 'match' */
+  int firstcap;  /* stack index of first capture pushed in the stack */
+  const char *s;  /* original string */
+  int valuecached;  /* value stored in cache slot */
+  int reclevel;  /* recursion level */
+} CapState;
+
+
+#define captype(cap)    ((cap)->kind)
+
+#define isclosecap(cap) (captype(cap) == Cclose)
+#define isopencap(cap)  ((cap)->siz == 0)
+
+/* true if c2 is (any number of levels) inside c1 */
+#define capinside(c1,c2)  \
+	(isopencap(c1) ? !isclosecap(c2) \
+                       : (c2)->index < (c1)->index + (c1)->siz - 1)
+
+
+/**
+** Maximum number of captures to visit when looking for an 'open'.
+*/
+#define MAXLOP		20
+
+
+
+int runtimecap (CapState *cs, Capture *close, const char *s, int *rem);
+int getcaptures (lua_State *L, const char *s, const char *r, int ptop);
+int finddyncap (Capture *cap, Capture *last);
+
+#endif
+
+

src/lua/lpeg-1.1.0/lpcode.h

@@ -0,0 +1,36 @@
+
+#if !defined(lpcode_h)
+#define lpcode_h
+
+#include "lua.h"
+
+#include "lptypes.h"
+#include "lptree.h"
+#include "lpvm.h"
+
+int checkaux (TTree *tree, int pred);
+int fixedlen (TTree *tree);
+int hascaptures (TTree *tree);
+int lp_gc (lua_State *L);
+Instruction *compile (lua_State *L, Pattern *p, uint size);
+void freecode (lua_State *L, Pattern *p);
+int sizei (const Instruction *i);
+
+
+#define PEnullable      0
+#define PEnofail        1
+
+/*
+** nofail(t) implies that 't' cannot fail with any input
+*/
+#define nofail(t)	checkaux(t, PEnofail)
+
+/*
+** (not nullable(t)) implies 't' cannot match without consuming
+** something
+*/
+#define nullable(t)	checkaux(t, PEnullable)
+
+
+
+#endif

src/lua/lpeg-1.1.0/lpcset.c

@@ -0,0 +1,110 @@
+
+#include "lptypes.h"
+#include "lpcset.h"
+
+
+/*
+** Add to 'c' the index of the (only) bit set in byte 'b'
+*/
+static int onlybit (int c, int b) {
+  if ((b & 0xF0) != 0) { c += 4; b >>= 4; }
+  if ((b & 0x0C) != 0) { c += 2; b >>= 2; }
+  if ((b & 0x02) != 0) { c += 1; }
+  return c;
+}
+
+
+/*
+** Check whether a charset is empty (returns IFail), singleton (IChar),
+** full (IAny), or none of those (ISet). When singleton, 'info.offset'
+** returns which character it is. When generic set, 'info' returns
+** information about its range.
+*/
+Opcode charsettype (const byte *cs, charsetinfo *info) {
+  int low0, low1, high0, high1;
+  for (low1 = 0; low1 < CHARSETSIZE && cs[low1] == 0; low1++)
+    /* find lowest byte with a 1-bit */;
+  if (low1 == CHARSETSIZE)
+    return IFail;  /* no characters in set */
+  for (high1 = CHARSETSIZE - 1; cs[high1] == 0; high1--)
+    /* find highest byte with a 1-bit; low1 is a sentinel */;
+  if (low1 == high1) {  /* only one byte with 1-bits? */
+    int b = cs[low1];
+    if ((b & (b - 1)) == 0) {  /* does byte has only one 1-bit? */
+      info->offset = onlybit(low1 * BITSPERCHAR, b);  /* get that bit */
+      return IChar;  /* single character */
+    }
+  }
+  for (low0 = 0; low0 < CHARSETSIZE && cs[low0] == 0xFF; low0++)
+    /* find lowest byte with a 0-bit */;
+  if (low0 == CHARSETSIZE)
+    return IAny;  /* set has all bits set */
+  for (high0 = CHARSETSIZE - 1; cs[high0] == 0xFF; high0--)
+    /* find highest byte with a 0-bit; low0 is a sentinel */;
+  if (high1 - low1 <= high0 - low0) {  /* range of 1s smaller than of 0s? */
+    info->offset = low1;
+    info->size = high1 - low1 + 1;
+    info->deflt = 0;  /* all discharged bits were 0 */
+  }
+  else {
+    info->offset = low0;
+    info->size = high0 - low0 + 1;
+    info->deflt = 0xFF;  /* all discharged bits were 1 */
+  }
+  info->cs = cs + info->offset;
+  return ISet;
+}
+
+
+/*
+** Get a byte from a compact charset. If index is inside the charset
+** range, get the byte from the supporting charset (correcting it
+** by the offset). Otherwise, return the default for the set.
+*/
+byte getbytefromcharset (const charsetinfo *info, int index) {
+  if (index < info->size)
+    return info->cs[index];
+  else return info->deflt;
+}
+
+
+/*
+** If 'tree' is a 'char' pattern (TSet, TChar, TAny, TFalse), convert it
+** into a charset and return 1; else return 0.
+*/
+int tocharset (TTree *tree, Charset *cs) {
+  switch (tree->tag) {
+    case TChar: {  /* only one char */
+      assert(0 <= tree->u.n && tree->u.n <= UCHAR_MAX);
+      clearset(cs->cs);  /* erase all chars */
+      setchar(cs->cs, tree->u.n);  /* add that one */
+      return 1;
+    }
+    case TAny: {
+      fillset(cs->cs, 0xFF);  /* add all characters to the set */
+      return 1;
+    }
+    case TFalse: {
+      clearset(cs->cs);  /* empty set */
+      return 1;
+    }
+    case TSet: {  /* fill set */
+      int i;
+      fillset(cs->cs, tree->u.set.deflt);
+      for (i = 0; i < tree->u.set.size; i++)
+        cs->cs[tree->u.set.offset + i] = treebuffer(tree)[i];
+      return 1;
+    }
+    default: return 0;
+  }
+}
+
+
+void tree2cset (TTree *tree, charsetinfo *info) {
+  assert(tree->tag == TSet);
+  info->offset = tree->u.set.offset;
+  info->size = tree->u.set.size;
+  info->deflt = tree->u.set.deflt;
+  info->cs = treebuffer(tree);
+}
+

src/lua/lpeg-1.1.0/lpcset.h

@@ -0,0 +1,30 @@
+
+#if !defined(lpset_h)
+#define lpset_h
+
+#include "lpcset.h"
+#include "lpcode.h"
+#include "lptree.h"
+
+
+/*
+** Extra information for the result of 'charsettype'.  When result is
+** IChar, 'offset' is the character.  When result is ISet, 'cs' is the
+** supporting bit array (with offset included), 'offset' is the offset
+** (in bytes), 'size' is the size (in bytes), and 'delt' is the default
+** value for bytes outside the set.
+*/
+typedef struct {
+  const byte *cs;
+  int offset;
+  int size;
+  int deflt;
+} charsetinfo;
+
+
+int tocharset (TTree *tree, Charset *cs);
+Opcode charsettype (const byte *cs, charsetinfo *info);
+byte getbytefromcharset (const charsetinfo *info, int index);
+void tree2cset (TTree *tree, charsetinfo *info);
+
+#endif

src/lua/lpeg-1.1.0/lpprint.c

@@ -0,0 +1,298 @@
+
+#include <ctype.h>
+#include <limits.h>
+#include <stdio.h>
+
+
+#include "lptypes.h"
+#include "lpprint.h"
+#include "lpcode.h"
+
+
+#if defined(LPEG_DEBUG)
+
+/*
+** {======================================================
+** Printing patterns (for debugging)
+** =======================================================
+*/
+
+
+void printcharset (const byte *st) {
+  int i;
+  printf("[");
+  for (i = 0; i <= UCHAR_MAX; i++) {
+    int first = i;
+    while (i <= UCHAR_MAX && testchar(st, i)) i++;
+    if (i - 1 == first)  /* unary range? */
+      printf("(%02x)", first);
+    else if (i - 1 > first)  /* non-empty range? */
+      printf("(%02x-%02x)", first, i - 1);
+  }
+  printf("]");
+}
+
+
+static void printIcharset (const Instruction *inst, const byte *buff) {
+  byte cs[CHARSETSIZE];
+  int i;
+  printf("(%02x-%d) ", inst->i.aux2.set.offset, inst->i.aux2.set.size);
+  clearset(cs);
+  for (i = 0; i < CHARSETSIZE * 8; i++) {
+    if (charinset(inst, buff, i))
+      setchar(cs, i);
+  }
+  printcharset(cs);
+}
+
+
+static void printTcharset (TTree *tree) {
+  byte cs[CHARSETSIZE];
+  int i;
+  printf("(%02x-%d) ", tree->u.set.offset, tree->u.set.size);
+  fillset(cs, tree->u.set.deflt);
+  for (i = 0; i < tree->u.set.size; i++)
+    cs[tree->u.set.offset + i] = treebuffer(tree)[i];
+  printcharset(cs);
+}
+
+
+static const char *capkind (int kind) {
+  const char *const modes[] = {
+    "close", "position", "constant", "backref",
+    "argument", "simple", "table", "function", "accumulator",
+    "query", "string", "num", "substitution", "fold",
+    "runtime", "group"};
+  return modes[kind];
+}
+
+
+static void printjmp (const Instruction *op, const Instruction *p) {
+  printf("-> %d", (int)(p + (p + 1)->offset - op));
+}
+
+
+void printinst (const Instruction *op, const Instruction *p) {
+  const char *const names[] = {
+    "any", "char", "set",
+    "testany", "testchar", "testset",
+    "span", "utf-range", "behind",
+    "ret", "end",
+    "choice", "jmp", "call", "open_call",
+    "commit", "partial_commit", "back_commit", "failtwice", "fail", "giveup",
+     "fullcapture", "opencapture", "closecapture", "closeruntime",
+     "--"
+  };
+  printf("%02ld: %s ", (long)(p - op), names[p->i.code]);
+  switch ((Opcode)p->i.code) {
+    case IChar: {
+      printf("'%c' (%02x)", p->i.aux1, p->i.aux1);
+      break;
+    }
+    case ITestChar: {
+      printf("'%c' (%02x)", p->i.aux1, p->i.aux1); printjmp(op, p);
+      break;
+    }
+    case IUTFR: {
+      printf("%d - %d", p[1].offset, utf_to(p));
+      break;
+    }
+    case IFullCapture: {
+      printf("%s (size = %d)  (idx = %d)",
+             capkind(getkind(p)), getoff(p), p->i.aux2.key);
+      break;
+    }
+    case IOpenCapture: {
+      printf("%s (idx = %d)", capkind(getkind(p)), p->i.aux2.key);
+      break;
+    }
+    case ISet: {
+      printIcharset(p, (p+1)->buff);
+      break;
+    }
+    case ITestSet: {
+      printIcharset(p, (p+2)->buff); printjmp(op, p);
+      break;
+    }
+    case ISpan: {
+      printIcharset(p, (p+1)->buff);
+      break;
+    }
+    case IOpenCall: {
+      printf("-> %d", (p + 1)->offset);
+      break;
+    }
+    case IBehind: {
+      printf("%d", p->i.aux1);
+      break;
+    }
+    case IJmp: case ICall: case ICommit: case IChoice:
+    case IPartialCommit: case IBackCommit: case ITestAny: {
+      printjmp(op, p);
+      break;
+    }
+    default: break;
+  }
+  printf("\n");
+}
+
+
+void printpatt (Instruction *p) {
+  Instruction *op = p;
+  uint n = op[-1].codesize - 1;
+  while (p < op + n) {
+    printinst(op, p);
+    p += sizei(p);
+  }
+}
+
+
+static void printcap (Capture *cap, int ident) {
+  while (ident--) printf(" ");
+  printf("%s (idx: %d - size: %d) -> %lu  (%p)\n",
+         capkind(cap->kind), cap->idx, cap->siz, (long)cap->index, (void*)cap);
+}
+
+
+/*
+** Print a capture and its nested captures
+*/
+static Capture *printcap2close (Capture *cap, int ident) {
+  Capture *head = cap++;
+  printcap(head, ident);  /* print head capture */
+  while (capinside(head, cap))
+    cap = printcap2close(cap, ident + 2);  /* print nested captures */
+  if (isopencap(head)) {
+    assert(isclosecap(cap));
+    printcap(cap++, ident);  /* print and skip close capture */
+  }
+  return cap;
+}
+
+
+void printcaplist (Capture *cap) {
+  {  /* for debugging, print first a raw list of captures */
+    Capture *c = cap;
+    while (c->index != MAXINDT) { printcap(c, 0); c++; }
+  }
+  printf(">======\n");
+  while (!isclosecap(cap))
+    cap = printcap2close(cap, 0);
+  printf("=======\n");
+}
+
+/* }====================================================== */
+
+
+/*
+** {======================================================
+** Printing trees (for debugging)
+** =======================================================
+*/
+
+static const char *tagnames[] = {
+  "char", "set", "any",
+  "true", "false", "utf8.range",
+  "rep",
+  "seq", "choice",
+  "not", "and",
+  "call", "opencall", "rule", "xinfo", "grammar",
+  "behind",
+  "capture", "run-time"
+};
+
+
+void printtree (TTree *tree, int ident) {
+  int i;
+  int sibs = numsiblings[tree->tag];
+  for (i = 0; i < ident; i++) printf(" ");
+  printf("%s", tagnames[tree->tag]);
+  switch (tree->tag) {
+    case TChar: {
+      int c = tree->u.n;
+      if (isprint(c))
+        printf(" '%c'\n", c);
+      else
+        printf(" (%02X)\n", c);
+      break;
+    }
+    case TSet: {
+      printTcharset(tree);
+      printf("\n");
+      break;
+    }
+    case TUTFR: {
+      assert(sib1(tree)->tag == TXInfo);
+      printf(" %d (%02x %d) - %d (%02x %d) \n",
+        tree->u.n, tree->key, tree->cap,
+        sib1(tree)->u.n, sib1(tree)->key, sib1(tree)->cap);
+      break;
+    }
+    case TOpenCall: case TCall: {
+      assert(sib1(sib2(tree))->tag == TXInfo);
+      printf(" key: %d  (rule: %d)\n", tree->key, sib1(sib2(tree))->u.n);
+      break;
+    }
+    case TBehind: {
+      printf(" %d\n", tree->u.n);
+      break;
+    }
+    case TCapture: {
+      printf(" kind: '%s'  key: %d\n", capkind(tree->cap), tree->key);
+      break;
+    }
+    case TRule: {
+      printf(" key: %d\n", tree->key);
+      sibs = 1;  /* do not print 'sib2' (next rule) as a sibling */
+      break;
+    }
+    case TXInfo: {
+      printf(" n: %d\n", tree->u.n);
+      break;
+    }
+    case TGrammar: {
+      TTree *rule = sib1(tree);
+      printf(" %d\n", tree->u.n);  /* number of rules */
+      for (i = 0; i < tree->u.n; i++) {
+        printtree(rule, ident + 2);
+        rule = sib2(rule);
+      }
+      assert(rule->tag == TTrue);  /* sentinel */
+      sibs = 0;  /* siblings already handled */
+      break;
+    }
+    default:
+      printf("\n");
+      break;
+  }
+  if (sibs >= 1) {
+    printtree(sib1(tree), ident + 2);
+    if (sibs >= 2)
+      printtree(sib2(tree), ident + 2);
+  }
+}
+
+
+void printktable (lua_State *L, int idx) {
+  int n, i;
+  lua_getuservalue(L, idx);
+  if (lua_isnil(L, -1))  /* no ktable? */
+    return;
+  n = lua_rawlen(L, -1);
+  printf("[");
+  for (i = 1; i <= n; i++) {
+    printf("%d = ", i);
+    lua_rawgeti(L, -1, i);
+    if (lua_isstring(L, -1))
+      printf("%s  ", lua_tostring(L, -1));
+    else
+      printf("%s  ", lua_typename(L, lua_type(L, -1)));
+    lua_pop(L, 1);
+  }
+  printf("]\n");
+  /* leave ktable at the stack */
+}
+
+/* }====================================================== */
+
+#endif

src/lua/lpeg-1.1.0/lpprint.h

@@ -0,0 +1,32 @@
+
+#if !defined(lpprint_h)
+#define lpprint_h
+
+
+#include "lptree.h"
+#include "lpvm.h"
+
+
+#if defined(LPEG_DEBUG)
+
+void printpatt (Instruction *p);
+void printtree (TTree *tree, int ident);
+void printktable (lua_State *L, int idx);
+void printcharset (const byte *st);
+void printcaplist (Capture *cap);
+void printinst (const Instruction *op, const Instruction *p);
+
+#else
+
+#define printktable(L,idx)  \
+	luaL_error(L, "function only implemented in debug mode")
+#define printtree(tree,i)  \
+	luaL_error(L, "function only implemented in debug mode")
+#define printpatt(p)  \
+	luaL_error(L, "function only implemented in debug mode")
+
+#endif
+
+
+#endif
+

src/lua/lpeg-1.1.0/lptree.h

@@ -0,0 +1,92 @@
+
+#if !defined(lptree_h)
+#define lptree_h
+
+
+#include "lptypes.h"
+
+
+/*
+** types of trees
+*/
+typedef enum TTag {
+  TChar = 0,  /* 'n' = char */
+  TSet,  /* the set is encoded in 'u.set' and the next 'u.set.size' bytes */
+  TAny,
+  TTrue,
+  TFalse,
+  TUTFR,  /* range of UTF-8 codepoints; 'n' has initial codepoint;
+             'cap' has length; 'key' has first byte;
+             extra info is similar for end codepoint */
+  TRep,  /* 'sib1'* */
+  TSeq,  /* 'sib1' 'sib2' */
+  TChoice,  /* 'sib1' / 'sib2' */
+  TNot,  /* !'sib1' */
+  TAnd,  /* &'sib1' */
+  TCall,  /* ktable[key] is rule's key; 'sib2' is rule being called */
+  TOpenCall,  /* ktable[key] is rule's key */
+  TRule,  /* ktable[key] is rule's key (but key == 0 for unused rules);
+             'sib1' is rule's pattern pre-rule; 'sib2' is next rule;
+             extra info 'n' is rule's sequential number */
+  TXInfo,  /* extra info */
+  TGrammar,  /* 'sib1' is initial (and first) rule */
+  TBehind,  /* 'sib1' is pattern, 'n' is how much to go back */
+  TCapture,  /* captures: 'cap' is kind of capture (enum 'CapKind');
+                ktable[key] is Lua value associated with capture;
+                'sib1' is capture body */
+  TRunTime  /* run-time capture: 'key' is Lua function;
+               'sib1' is capture body */
+} TTag;
+
+
+/*
+** Tree trees
+** The first child of a tree (if there is one) is immediately after
+** the tree.  A reference to a second child (ps) is its position
+** relative to the position of the tree itself.
+*/
+typedef struct TTree {
+  byte tag;
+  byte cap;  /* kind of capture (if it is a capture) */
+  unsigned short key;  /* key in ktable for Lua data (0 if no key) */
+  union {
+    int ps;  /* occasional second child */
+    int n;  /* occasional counter */
+    struct {
+      byte offset;  /* compact set offset (in bytes) */
+      byte size;  /* compact set size (in bytes) */
+      byte deflt;  /* default value */
+      byte bitmap[1];  /* bitmap (open array) */
+    } set;  /* for compact sets */
+  } u;
+} TTree;
+
+
+/* access to charset */
+#define treebuffer(t)      ((t)->u.set.bitmap)
+
+
+/*
+** A complete pattern has its tree plus, if already compiled,
+** its corresponding code
+*/
+typedef struct Pattern {
+  union Instruction *code;
+  TTree tree[1];
+} Pattern;
+
+
+/* number of children for each tree */
+extern const byte numsiblings[];
+
+/* access to children */
+#define sib1(t)         ((t) + 1)
+#define sib2(t)         ((t) + (t)->u.ps)
+
+
+
+
+
+
+#endif
+

src/lua/lpeg-1.1.0/lptypes.h

@@ -0,0 +1,150 @@
+/*
+** LPeg - PEG pattern matching for Lua
+** Copyright 2007-2023, Lua.org & PUC-Rio  (see 'lpeg.html' for license)
+** written by Roberto Ierusalimschy
+*/
+
+#if !defined(lptypes_h)
+#define lptypes_h
+
+
+#include <assert.h>
+#include <limits.h>
+#include <string.h>
+
+#include "lua.h"
+
+
+#define VERSION         "1.1.0"
+
+
+#define PATTERN_T	"lpeg-pattern"
+#define MAXSTACKIDX	"lpeg-maxstack"
+
+
+/*
+** compatibility with Lua 5.1
+*/
+#if (LUA_VERSION_NUM == 501)
+
+#define lp_equal	lua_equal
+
+#define lua_getuservalue	lua_getfenv
+#define lua_setuservalue	lua_setfenv
+
+#define lua_rawlen		lua_objlen
+
+#define luaL_setfuncs(L,f,n)	luaL_register(L,NULL,f)
+#define luaL_newlib(L,f)	luaL_register(L,"lpeg",f)
+
+typedef size_t lua_Unsigned;
+
+#endif
+
+
+#if !defined(lp_equal)
+#define lp_equal(L,idx1,idx2)  lua_compare(L,(idx1),(idx2),LUA_OPEQ)
+#endif
+
+
+/* default maximum size for call/backtrack stack */
+#if !defined(MAXBACK)
+#define MAXBACK         400
+#endif
+
+
+/* maximum number of rules in a grammar (limited by 'unsigned short') */
+#if !defined(MAXRULES)
+#define MAXRULES        1000
+#endif
+
+
+
+/* initial size for capture's list */
+#define INITCAPSIZE	32
+
+
+/* index, on Lua stack, for subject */
+#define SUBJIDX		2
+
+/* number of fixed arguments to 'match' (before capture arguments) */
+#define FIXEDARGS	3
+
+/* index, on Lua stack, for capture list */
+#define caplistidx(ptop)	((ptop) + 2)
+
+/* index, on Lua stack, for pattern's ktable */
+#define ktableidx(ptop)		((ptop) + 3)
+
+/* index, on Lua stack, for backtracking stack */
+#define stackidx(ptop)	((ptop) + 4)
+
+
+
+typedef unsigned char byte;
+
+typedef unsigned int uint;
+
+
+#define BITSPERCHAR		8
+
+#define CHARSETSIZE		((UCHAR_MAX/BITSPERCHAR) + 1)
+
+
+
+typedef struct Charset {
+  byte cs[CHARSETSIZE];
+} Charset;
+
+
+
+#define loopset(v,b)    { int v; for (v = 0; v < CHARSETSIZE; v++) {b;} }
+
+#define fillset(s,c)	memset(s,c,CHARSETSIZE)
+#define clearset(s)	fillset(s,0)
+
+/* number of slots needed for 'n' bytes */
+#define bytes2slots(n)  (((n) - 1u) / (uint)sizeof(TTree) + 1u)
+
+/* set 'b' bit in charset 'cs' */
+#define setchar(cs,b)   ((cs)[(b) >> 3] |= (1 << ((b) & 7)))
+
+
+/*
+** in capture instructions, 'kind' of capture and its offset are
+** packed in field 'aux', 4 bits for each
+*/
+#define getkind(op)		((op)->i.aux1 & 0xF)
+#define getoff(op)		(((op)->i.aux1 >> 4) & 0xF)
+#define joinkindoff(k,o)	((k) | ((o) << 4))
+
+#define MAXOFF		0xF
+#define MAXAUX		0xFF
+
+
+/* maximum number of bytes to look behind */
+#define MAXBEHIND	MAXAUX
+
+
+/* maximum size (in elements) for a pattern */
+#define MAXPATTSIZE	(SHRT_MAX - 10)
+
+
+/* size (in instructions) for l bytes (l > 0) */
+#define instsize(l) ((int)(((l) + (uint)sizeof(Instruction) - 1u) \
+			/ (uint)sizeof(Instruction)))
+
+
+/* size (in elements) for a ISet instruction */
+#define CHARSETINSTSIZE		(1 + instsize(CHARSETSIZE))
+
+/* size (in elements) for a IFunc instruction */
+#define funcinstsize(p)		((p)->i.aux + 2)
+
+
+
+#define testchar(st,c)	((((uint)(st)[((c) >> 3)]) >> ((c) & 7)) & 1)
+
+
+#endif
+

src/lua/lpeg-1.1.0/lpvm.c

@@ -0,0 +1,455 @@
+
+#include <limits.h>
+#include <string.h>
+
+
+#include "lua.h"
+#include "lauxlib.h"
+
+#include "lpcap.h"
+#include "lptypes.h"
+#include "lpvm.h"
+#include "lpprint.h"
+
+
+/* initial size for call/backtrack stack */
+#if !defined(INITBACK)
+#define INITBACK	MAXBACK
+#endif
+
+
+#define getoffset(p)	(((p) + 1)->offset)
+
+static const Instruction giveup = {{IGiveup, 0, {0}}};
+
+
+int charinset (const Instruction *i, const byte *buff, uint c) {
+  c -= i->i.aux2.set.offset;
+  if (c >= ((uint)i->i.aux2.set.size  /* size in instructions... */
+           * (uint)sizeof(Instruction)  /* in bytes... */
+           * 8u))  /* in bits */
+    return i->i.aux1;  /* out of range; return default value */
+  return testchar(buff, c);
+}
+
+
+/*
+** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
+*/
+static const char *utf8_decode (const char *o, int *val) {
+  static const uint limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFFu};
+  const unsigned char *s = (const unsigned char *)o;
+  uint c = s[0];  /* first byte */
+  uint res = 0;  /* final result */
+  if (c < 0x80)  /* ascii? */
+    res = c;
+  else {
+    int count = 0;  /* to count number of continuation bytes */
+    while (c & 0x40) {  /* still have continuation bytes? */
+      int cc = s[++count];  /* read next byte */
+      if ((cc & 0xC0) != 0x80)  /* not a continuation byte? */
+        return NULL;  /* invalid byte sequence */
+      res = (res << 6) | (cc & 0x3F);  /* add lower 6 bits from cont. byte */
+      c <<= 1;  /* to test next bit */
+    }
+    res |= (c & 0x7F) << (count * 5);  /* add first byte */
+    if (count > 3 || res > 0x10FFFFu || res <= limits[count])
+      return NULL;  /* invalid byte sequence */
+    s += count;  /* skip continuation bytes read */
+  }
+  *val = res;
+  return (const char *)s + 1;  /* +1 to include first byte */
+}
+
+
+/*
+** {======================================================
+** Virtual Machine
+** =======================================================
+*/
+
+
+typedef struct Stack {
+  const char *s;  /* saved position (or NULL for calls) */
+  const Instruction *p;  /* next instruction */
+  int caplevel;
+} Stack;
+
+
+#define getstackbase(L, ptop)	((Stack *)lua_touserdata(L, stackidx(ptop)))
+
+
+/*
+** Ensures the size of array 'capture' (with size '*capsize' and
+** 'captop' elements being used) is enough to accomodate 'n' extra
+** elements plus one.  (Because several opcodes add stuff to the capture
+** array, it is simpler to ensure the array always has at least one free
+** slot upfront and check its size later.)
+*/
+
+/* new size in number of elements cannot overflow integers, and new
+   size in bytes cannot overflow size_t. */
+#define MAXNEWSIZE  \
+    (((size_t)INT_MAX) <= (~(size_t)0 / sizeof(Capture)) ?  \
+     ((size_t)INT_MAX) : (~(size_t)0 / sizeof(Capture)))
+
+static Capture *growcap (lua_State *L, Capture *capture, int *capsize,
+                                       int captop, int n, int ptop) {
+  if (*capsize - captop > n)
+    return capture;  /* no need to grow array */
+  else {  /* must grow */
+    Capture *newc;
+    uint newsize = captop + n + 1;  /* minimum size needed */
+    if (newsize < (MAXNEWSIZE / 3) * 2)
+      newsize += newsize / 2;  /* 1.5 that size, if not too big */
+    else if (newsize < (MAXNEWSIZE / 9) * 8)
+      newsize += newsize / 8;  /* else, try 9/8 that size */
+    else
+      luaL_error(L, "too many captures");
+    newc = (Capture *)lua_newuserdata(L, newsize * sizeof(Capture));
+    memcpy(newc, capture, captop * sizeof(Capture));
+    *capsize = newsize;
+    lua_replace(L, caplistidx(ptop));
+    return newc;
+  }
+}
+
+
+/*
+** Double the size of the stack
+*/
+static Stack *doublestack (lua_State *L, Stack **stacklimit, int ptop) {
+  Stack *stack = getstackbase(L, ptop);
+  Stack *newstack;
+  int n = *stacklimit - stack;  /* current stack size */
+  int max, newn;
+  lua_getfield(L, LUA_REGISTRYINDEX, MAXSTACKIDX);
+  max = lua_tointeger(L, -1);  /* maximum allowed size */
+  lua_pop(L, 1);
+  if (n >= max)  /* already at maximum size? */
+    luaL_error(L, "backtrack stack overflow (current limit is %d)", max);
+  newn = 2 * n;  /* new size */
+  if (newn > max) newn = max;
+  newstack = (Stack *)lua_newuserdata(L, newn * sizeof(Stack));
+  memcpy(newstack, stack, n * sizeof(Stack));
+  lua_replace(L, stackidx(ptop));
+  *stacklimit = newstack + newn;
+  return newstack + n;  /* return next position */
+}
+
+
+/*
+** Interpret the result of a dynamic capture: false -> fail;
+** true -> keep current position; number -> next position.
+** Return new subject position. 'fr' is stack index where
+** is the result; 'curr' is current subject position; 'limit'
+** is subject's size.
+*/
+static int resdyncaptures (lua_State *L, int fr, int curr, int limit) {
+  lua_Integer res;
+  if (!lua_toboolean(L, fr)) {  /* false value? */
+    lua_settop(L, fr - 1);  /* remove results */
+    return -1;  /* and fail */
+  }
+  else if (lua_isboolean(L, fr))  /* true? */
+    res = curr;  /* keep current position */
+  else {
+    res = lua_tointeger(L, fr) - 1;  /* new position */
+    if (res < curr || res > limit)
+      luaL_error(L, "invalid position returned by match-time capture");
+  }
+  lua_remove(L, fr);  /* remove first result (offset) */
+  return res;
+}
+
+
+/*
+** Add capture values returned by a dynamic capture to the list
+** 'capture', nested inside a group. 'fd' indexes the first capture
+** value, 'n' is the number of values (at least 1). The open group
+** capture is already in 'capture', before the place for the new entries.
+*/
+static void adddyncaptures (Index_t index, Capture *capture, int n, int fd) {
+  int i;
+  assert(capture[-1].kind == Cgroup && capture[-1].siz == 0);
+  capture[-1].idx = 0;  /* make group capture an anonymous group */
+  for (i = 0; i < n; i++) {  /* add runtime captures */
+    capture[i].kind = Cruntime;
+    capture[i].siz = 1;  /* mark it as closed */
+    capture[i].idx = fd + i;  /* stack index of capture value */
+    capture[i].index = index;
+  }
+  capture[n].kind = Cclose;  /* close group */
+  capture[n].siz = 1;
+  capture[n].index = index;
+}
+
+
+/*
+** Remove dynamic captures from the Lua stack (called in case of failure)
+*/
+static int removedyncap (lua_State *L, Capture *capture,
+                         int level, int last) {
+  int id = finddyncap(capture + level, capture + last);  /* index of 1st cap. */
+  int top = lua_gettop(L);
+  if (id == 0) return 0;  /* no dynamic captures? */
+  lua_settop(L, id - 1);  /* remove captures */
+  return top - id + 1;  /* number of values removed */
+}
+
+
+/*
+** Find the corresponding 'open' capture before 'cap', when that capture
+** can become a full capture. If a full capture c1 is followed by an
+** empty capture c2, there is no way to know whether c2 is inside
+** c1. So, full captures can enclose only captures that start *before*
+** its end.
+*/
+static Capture *findopen (Capture *cap, Index_t currindex) {
+  int i;
+  cap--;  /* check last capture */
+  /* Must it be inside current one, but starts where current one ends? */
+  if (!isopencap(cap) && cap->index == currindex)
+    return NULL;  /* current one cannot be a full capture */
+  /* else, look for an 'open' capture */
+  for (i = 0; i < MAXLOP; i++, cap--) {
+    if (currindex - cap->index >= UCHAR_MAX)
+      return NULL;  /* capture too long for a full capture */
+    else if (isopencap(cap))  /* open capture? */
+      return cap;  /* that's the one to be closed */
+    else if (cap->kind == Cclose)
+      return NULL;  /* a full capture should not nest a non-full one */
+  }
+  return NULL;  /* not found within allowed search limit */
+}
+
+
+/*
+** Opcode interpreter
+*/
+const char *match (lua_State *L, const char *o, const char *s, const char *e,
+                   Instruction *op, Capture *capture, int ptop) {
+  Stack stackbase[INITBACK];
+  Stack *stacklimit = stackbase + INITBACK;
+  Stack *stack = stackbase;  /* point to first empty slot in stack */
+  int capsize = INITCAPSIZE;
+  int captop = 0;  /* point to first empty slot in captures */
+  int ndyncap = 0;  /* number of dynamic captures (in Lua stack) */
+  const Instruction *p = op;  /* current instruction */
+  stack->p = &giveup; stack->s = s; stack->caplevel = 0; stack++;
+  lua_pushlightuserdata(L, stackbase);
+  for (;;) {
+#if defined(DEBUG)
+      printf("-------------------------------------\n");
+      printcaplist(capture, capture + captop);
+      printf("s: |%s| stck:%d, dyncaps:%d, caps:%d  ",
+             s, (int)(stack - getstackbase(L, ptop)), ndyncap, captop);
+      printinst(op, p);
+#endif
+    assert(stackidx(ptop) + ndyncap == lua_gettop(L) && ndyncap <= captop);
+    switch ((Opcode)p->i.code) {
+      case IEnd: {
+        assert(stack == getstackbase(L, ptop) + 1);
+        capture[captop].kind = Cclose;
+        capture[captop].index = MAXINDT;
+        return s;
+      }
+      case IGiveup: {
+        assert(stack == getstackbase(L, ptop));
+        return NULL;
+      }
+      case IRet: {
+        assert(stack > getstackbase(L, ptop) && (stack - 1)->s == NULL);
+        p = (--stack)->p;
+        continue;
+      }
+      case IAny: {
+        if (s < e) { p++; s++; }
+        else goto fail;
+        continue;
+      }
+      case IUTFR: {
+        int codepoint;
+        if (s >= e)
+          goto fail;
+        s = utf8_decode (s, &codepoint);
+        if (s && p[1].offset <= codepoint && codepoint <= utf_to(p))
+          p += 2;
+        else
+          goto fail;
+        continue;
+      }
+      case ITestAny: {
+        if (s < e) p += 2;
+        else p += getoffset(p);
+        continue;
+      }
+      case IChar: {
+        if ((byte)*s == p->i.aux1 && s < e) { p++; s++; }
+        else goto fail;
+        continue;
+      }
+      case ITestChar: {
+        if ((byte)*s == p->i.aux1 && s < e) p += 2;
+        else p += getoffset(p);
+        continue;
+      }
+      case ISet: {
+        uint c = (byte)*s;
+        if (charinset(p, (p+1)->buff, c) && s < e)
+          { p += 1 + p->i.aux2.set.size; s++; }
+        else goto fail;
+        continue;
+      }
+      case ITestSet: {
+        uint c = (byte)*s;
+        if (charinset(p, (p + 2)->buff, c) && s < e)
+          p += 2 + p->i.aux2.set.size;
+        else p += getoffset(p);
+        continue;
+      }
+      case IBehind: {
+        int n = p->i.aux1;
+        if (n > s - o) goto fail;
+        s -= n; p++;
+        continue;
+      }
+      case ISpan: {
+        for (; s < e; s++) {
+          uint c = (byte)*s;
+          if (!charinset(p, (p+1)->buff, c)) break;
+        }
+        p += 1 + p->i.aux2.set.size;
+        continue;
+      }
+      case IJmp: {
+        p += getoffset(p);
+        continue;
+      }
+      case IChoice: {
+        if (stack == stacklimit)
+          stack = doublestack(L, &stacklimit, ptop);
+        stack->p = p + getoffset(p);
+        stack->s = s;
+        stack->caplevel = captop;
+        stack++;
+        p += 2;
+        continue;
+      }
+      case ICall: {
+        if (stack == stacklimit)
+          stack = doublestack(L, &stacklimit, ptop);
+        stack->s = NULL;
+        stack->p = p + 2;  /* save return address */
+        stack++;
+        p += getoffset(p);
+        continue;
+      }
+      case ICommit: {
+        assert(stack > getstackbase(L, ptop) && (stack - 1)->s != NULL);
+        stack--;
+        p += getoffset(p);
+        continue;
+      }
+      case IPartialCommit: {
+        assert(stack > getstackbase(L, ptop) && (stack - 1)->s != NULL);
+        (stack - 1)->s = s;
+        (stack - 1)->caplevel = captop;
+        p += getoffset(p);
+        continue;
+      }
+      case IBackCommit: {
+        assert(stack > getstackbase(L, ptop) && (stack - 1)->s != NULL);
+        s = (--stack)->s;
+        if (ndyncap > 0)  /* are there matchtime captures? */
+          ndyncap -= removedyncap(L, capture, stack->caplevel, captop);
+        captop = stack->caplevel;
+        p += getoffset(p);
+        continue;
+      }
+      case IFailTwice:
+        assert(stack > getstackbase(L, ptop));
+        stack--;
+        /* FALLTHROUGH */
+      case IFail:
+      fail: { /* pattern failed: try to backtrack */
+        do {  /* remove pending calls */
+          assert(stack > getstackbase(L, ptop));
+          s = (--stack)->s;
+        } while (s == NULL);
+        if (ndyncap > 0)  /* is there matchtime captures? */
+          ndyncap -= removedyncap(L, capture, stack->caplevel, captop);
+        captop = stack->caplevel;
+        p = stack->p;
+#if defined(DEBUG)
+        printf("**FAIL**\n");
+#endif
+        continue;
+      }
+      case ICloseRunTime: {
+        CapState cs;
+        int rem, res, n;
+        int fr = lua_gettop(L) + 1;  /* stack index of first result */
+        cs.reclevel = 0; cs.L = L;
+        cs.s = o; cs.ocap = capture; cs.ptop = ptop;
+        n = runtimecap(&cs, capture + captop, s, &rem);  /* call function */
+        captop -= n;  /* remove nested captures */
+        ndyncap -= rem;  /* update number of dynamic captures */
+        fr -= rem;  /* 'rem' items were popped from Lua stack */
+        res = resdyncaptures(L, fr, s - o, e - o);  /* get result */
+        if (res == -1)  /* fail? */
+          goto fail;
+        s = o + res;  /* else update current position */
+        n = lua_gettop(L) - fr + 1;  /* number of new captures */
+        ndyncap += n;  /* update number of dynamic captures */
+        if (n == 0)  /* no new captures? */
+          captop--;  /* remove open group */
+        else {  /* new captures; keep original open group */
+          if (fr + n >= SHRT_MAX)
+            luaL_error(L, "too many results in match-time capture");
+          /* add new captures + close group to 'capture' list */
+          capture = growcap(L, capture, &capsize, captop, n + 1, ptop);
+          adddyncaptures(s - o, capture + captop, n, fr);
+          captop += n + 1;  /* new captures + close group */
+        }
+        p++;
+        continue;
+      }
+      case ICloseCapture: {
+        Capture *open = findopen(capture + captop, s - o);
+        assert(captop > 0);
+        if (open) {  /* if possible, turn capture into a full capture */
+          open->siz = (s - o) - open->index + 1;
+          p++;
+          continue;
+        }
+        else {  /* must create a close capture */
+          capture[captop].siz = 1;  /* mark entry as closed */
+          capture[captop].index = s - o;
+          goto pushcapture;
+        }
+      }
+      case IOpenCapture:
+        capture[captop].siz = 0;  /* mark entry as open */
+        capture[captop].index = s - o;
+        goto pushcapture;
+      case IFullCapture:
+        capture[captop].siz = getoff(p) + 1;  /* save capture size */
+        capture[captop].index = s - o - getoff(p);
+        /* goto pushcapture; */
+      pushcapture: {
+        capture[captop].idx = p->i.aux2.key;
+        capture[captop].kind = getkind(p);
+        captop++;
+        capture = growcap(L, capture, &capsize, captop, 0, ptop);
+        p++;
+        continue;
+      }
+      default: assert(0); return NULL;
+    }
+  }
+}
+
+/* }====================================================== */
+
+

src/lua/lpeg-1.1.0/lpvm.h

@@ -0,0 +1,79 @@
+
+#if !defined(lpvm_h)
+#define lpvm_h
+
+#include "lpcap.h"
+
+
+/*
+** About Character sets in instructions: a set is a bit map with an
+** initial offset, in bits, and a size, in number of instructions.
+** aux1 has the default value for the bits outsize that range.
+*/
+
+
+/* Virtual Machine's instructions */
+typedef enum Opcode {
+  IAny, /* if no char, fail */
+  IChar,  /* if char != aux1, fail */
+  ISet,  /* if char not in set, fail */
+  ITestAny,  /* in no char, jump to 'offset' */
+  ITestChar,  /* if char != aux1, jump to 'offset' */
+  ITestSet,  /* if char not in set, jump to 'offset' */
+  ISpan,  /* read a span of chars in set */
+  IUTFR,  /* if codepoint not in range [offset, utf_to], fail */
+  IBehind,  /* walk back 'aux1' characters (fail if not possible) */
+  IRet,  /* return from a rule */
+  IEnd,  /* end of pattern */
+  IChoice,  /* stack a choice; next fail will jump to 'offset' */
+  IJmp,  /* jump to 'offset' */
+  ICall,  /* call rule at 'offset' */
+  IOpenCall,  /* call rule number 'key' (must be closed to a ICall) */
+  ICommit,  /* pop choice and jump to 'offset' */
+  IPartialCommit,  /* update top choice to current position and jump */
+  IBackCommit,  /* backtrack like "fail" but jump to its own 'offset' */
+  IFailTwice,  /* pop one choice and then fail */
+  IFail,  /* go back to saved state on choice and jump to saved offset */
+  IGiveup,  /* internal use */
+  IFullCapture,  /* complete capture of last 'off' chars */
+  IOpenCapture,  /* start a capture */
+  ICloseCapture,
+  ICloseRunTime,
+  IEmpty  /* to fill empty slots left by optimizations */
+} Opcode;
+
+
+/*
+** All array of instructions has a 'codesize' as its first element
+** and is referred by a pointer to its second element, which is the
+** first actual opcode.
+*/
+typedef union Instruction {
+  struct Inst {
+    byte code;
+    byte aux1;
+    union {
+      short key;
+      struct {
+        byte offset;
+        byte size;
+      } set;
+    } aux2;
+  } i;
+  int offset;
+  uint codesize;
+  byte buff[1];
+} Instruction;
+
+
+/* extract 24-bit value from an instruction */
+#define utf_to(inst)	(((inst)->i.aux2.key << 8) | (inst)->i.aux1)
+
+
+int charinset (const Instruction *i, const byte *buff, uint c);
+const char *match (lua_State *L, const char *o, const char *s, const char *e,
+                   Instruction *op, Capture *capture, int ptop);
+
+
+#endif
+

src/lua/lpeg-1.1.0/makefile

@@ -0,0 +1,59 @@
+LIBNAME = lpeg
+LUADIR = ./lua/
+
+COPT = -O2 -DNDEBUG
+# COPT = -O0 -DLPEG_DEBUG -g
+
+CWARNS = -Wall -Wextra -pedantic \
+	-Waggregate-return \
+	-Wcast-align \
+	-Wcast-qual \
+	-Wdisabled-optimization \
+	-Wpointer-arith \
+	-Wshadow \
+	-Wredundant-decls \
+	-Wsign-compare \
+	-Wundef \
+	-Wwrite-strings \
+	-Wbad-function-cast \
+	-Wdeclaration-after-statement \
+	-Wmissing-prototypes \
+	-Wmissing-declarations \
+	-Wnested-externs \
+	-Wstrict-prototypes \
+	-Wc++-compat \
+# -Wunreachable-code \
+
+
+CFLAGS = $(CWARNS) $(COPT) -std=c99 -I$(LUADIR) -fPIC
+CC = gcc
+
+FILES = lpvm.o lpcap.o lptree.o lpcode.o lpprint.o lpcset.o
+
+# For Linux
+linux:
+	$(MAKE) lpeg.so "DLLFLAGS = -shared -fPIC"
+
+# For Mac OS
+macosx:
+	$(MAKE) lpeg.so "DLLFLAGS = -bundle -undefined dynamic_lookup"
+
+lpeg.so: $(FILES)
+	env $(CC) $(DLLFLAGS) $(FILES) -o lpeg.so
+
+$(FILES): makefile
+
+test: test.lua re.lua lpeg.so
+	./test.lua
+
+clean:
+	rm -f $(FILES) lpeg.so
+
+
+lpcap.o: lpcap.c lpcap.h lptypes.h
+lpcode.o: lpcode.c lptypes.h lpcode.h lptree.h lpvm.h lpcap.h lpcset.h
+lpcset.o: lpcset.c lptypes.h lpcset.h lpcode.h lptree.h lpvm.h lpcap.h
+lpprint.o: lpprint.c lptypes.h lpprint.h lptree.h lpvm.h lpcap.h lpcode.h
+lptree.o: lptree.c lptypes.h lpcap.h lpcode.h lptree.h lpvm.h lpprint.h \
+ lpcset.h
+lpvm.o: lpvm.c lpcap.h lptypes.h lpvm.h lpprint.h lptree.h

src/lua/lpeg-1.1.0/re.html

@@ -0,0 +1,472 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
+   "//www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
+<html>
+<head>
+    <title>LPeg.re - Regex syntax for LPEG</title>
+    <link rel="stylesheet"
+          href="//www.inf.puc-rio.br/~roberto/lpeg/doc.css"
+          type="text/css"/>
+	<meta http-equiv="Content-Type" content="text/html; charset=UTF-8"/>
+</head>
+<body>
+
+
+<div id="container">
+	
+<div id="product">
+  <div id="product_logo">
+    <a href="//www.inf.puc-rio.br/~roberto/lpeg/">
+    <img alt="LPeg logo" src="lpeg-128.gif"/>
+    </a>
+  </div>
+  <div id="product_name"><big><strong>LPeg.re</strong></big></div>
+  <div id="product_description">
+     Regex syntax for LPEG
+  </div>
+</div> <!-- id="product" -->
+
+<div id="main">
+	
+<div id="navigation">
+<h1>re</h1>
+
+<ul>
+  <li><a href="#basic">Basic Constructions</a></li>
+  <li><a href="#func">Functions</a></li>
+  <li><a href="#ex">Some Examples</a></li>
+  <li><a href="#license">License</a></li>
+  </ul>
+  </li>
+</ul>
+</div> <!-- id="navigation" -->
+
+<div id="content">
+
+<h2><a name="basic"></a>The <code>re</code> Module</h2>
+
+<p>
+The <code>re</code> module
+(provided by file <code>re.lua</code> in the distribution)
+supports a somewhat conventional regex syntax
+for pattern usage within <a href="lpeg.html">LPeg</a>.
+</p>
+
+<p>
+The next table summarizes <code>re</code>'s syntax.
+A <code>p</code> represents an arbitrary pattern;
+<code>num</code> represents a number (<code>[0-9]+</code>);
+<code>name</code> represents an identifier
+(<code>[a-zA-Z][a-zA-Z0-9_]*</code>).
+Constructions are listed in order of decreasing precedence.
+<table border="1">
+<tbody><tr><td><b>Syntax</b></td><td><b>Description</b></td></tr>
+<tr><td><code>( p )</code></td> <td>grouping</td></tr>
+<tr><td><code>&amp; p</code></td> <td>and predicate</td></tr>
+<tr><td><code>! p</code></td> <td>not predicate</td></tr>
+<tr><td><code>p1 p2</code></td> <td>concatenation</td></tr>
+<tr><td><code>p1 / p2</code></td> <td>ordered choice</td></tr>
+<tr><td><code>p ?</code></td> <td>optional match</td></tr>
+<tr><td><code>p *</code></td> <td>zero or more repetitions</td></tr>
+<tr><td><code>p +</code></td> <td>one or more repetitions</td></tr>
+<tr><td><code>p^num</code></td>
+      <td>exactly <code>num</code> repetitions</td></tr>
+<tr><td><code>p^+num</code></td>
+      <td>at least <code>num</code> repetitions</td></tr>
+<tr><td><code>p^-num</code></td>
+      <td>at most <code>num</code> repetitions</td></tr>
+<tr><td>(<code>name &lt;- p</code>)<sup>+</sup></td> <td>grammar</td></tr>
+<tr><td><code>'string'</code></td> <td>literal string</td></tr>
+<tr><td><code>"string"</code></td> <td>literal string</td></tr>
+<tr><td><code>[class]</code></td> <td>character class</td></tr>
+<tr><td><code>.</code></td> <td>any character</td></tr>
+<tr><td><code>%name</code></td>
+  <td>pattern <code>defs[name]</code> or a pre-defined pattern</td></tr>
+<tr><td><code>name</code></td><td>non terminal</td></tr>
+<tr><td><code>&lt;name&gt;</code></td><td>non terminal</td></tr>
+
+<tr><td><code>{}</code></td> <td>position capture</td></tr>
+<tr><td><code>{ p }</code></td> <td>simple capture</td></tr>
+<tr><td><code>{: p :}</code></td> <td>anonymous group capture</td></tr>
+<tr><td><code>{:name: p :}</code></td> <td>named group capture</td></tr>
+<tr><td><code>{~ p ~}</code></td> <td>substitution capture</td></tr>
+<tr><td><code>{| p |}</code></td> <td>table capture</td></tr>
+<tr><td><code>=name</code></td> <td>back reference</td></tr>
+
+<tr><td><code>p -&gt; 'string'</code></td> <td>string capture</td></tr>
+<tr><td><code>p -&gt; "string"</code></td> <td>string capture</td></tr>
+<tr><td><code>p -&gt; num</code></td> <td>numbered capture</td></tr>
+<tr><td><code>p -&gt; name</code></td> <td>function/query/string capture
+equivalent to <code>p / defs[name]</code></td></tr>
+<tr><td><code>p =&gt; name</code></td> <td>match-time capture
+equivalent to <code>lpeg.Cmt(p, defs[name])</code></td></tr>
+<tr><td><code>p ~&gt; name</code></td> <td>fold capture
+(deprecated)</td></tr>
+<tr><td><code>p &gt;&gt; name</code></td> <td>accumulator capture
+equivalent to <code>(p % defs[name])</code></td></tr>
+</tbody></table>
+<p>
+Any space appearing in a syntax description can be
+replaced by zero or more space characters and Lua-style short comments
+(<code>--</code> until end of line).
+</p>
+
+<p>
+Character classes define sets of characters.
+An initial <code>^</code> complements the resulting set.
+A range <em>x</em><code>-</code><em>y</em> includes in the set
+all characters with codes between the codes of <em>x</em> and <em>y</em>.
+A pre-defined class <code>%</code><em>name</em> includes all
+characters of that class.
+A simple character includes itself in the set.
+The only special characters inside a class are <code>^</code>
+(special only if it is the first character);
+<code>]</code>
+(can be included in the set as the first character,
+after the optional <code>^</code>);
+<code>%</code> (special only if followed by a letter);
+and <code>-</code>
+(can be included in the set as the first or the last character).
+</p>
+
+<p>
+Currently the pre-defined classes are similar to those from the
+Lua's string library
+(<code>%a</code> for letters,
+<code>%A</code> for non letters, etc.).
+There is also a class <code>%nl</code>
+containing only the newline character,
+which is particularly handy for grammars written inside long strings,
+as long strings do not interpret escape sequences like <code>\n</code>.
+</p>
+
+
+<h2><a name="func">Functions</a></h2>
+
+<h3><code>re.compile (string, [, defs])</code></h3>
+<p>
+Compiles the given string and
+returns an equivalent LPeg pattern.
+The given string may define either an expression or a grammar.
+The optional <code>defs</code> table provides extra Lua values
+to be used by the pattern.
+</p>
+
+<h3><code>re.find (subject, pattern [, init])</code></h3>
+<p>
+Searches the given pattern in the given subject.
+If it finds a match,
+returns the index where this occurrence starts and
+the index where it ends.
+Otherwise, returns nil.
+</p>
+
+<p>
+An optional numeric argument <code>init</code> makes the search
+starts at that position in the subject string.
+As usual in Lua libraries,
+a negative value counts from the end.
+</p>
+
+<h3><code>re.gsub (subject, pattern, replacement)</code></h3>
+<p>
+Does a <em>global substitution</em>,
+replacing all occurrences of <code>pattern</code>
+in the given <code>subject</code> by <code>replacement</code>.
+
+<h3><code>re.match (subject, pattern)</code></h3>
+<p>
+Matches the given pattern against the given subject,
+returning all captures.
+</p>
+
+<h3><code>re.updatelocale ()</code></h3>
+<p>
+Updates the pre-defined character classes to the current locale.
+</p>
+
+
+<h2><a name="ex">Some Examples</a></h2>
+
+<h3>A complete simple program</h3>
+<p>
+The next code shows a simple complete Lua program using
+the <code>re</code> module:
+</p>
+<pre class="example">
+local re = require"re"
+
+-- find the position of the first numeral in a string
+print(re.find("the number 423 is odd", "[0-9]+"))  --&gt; 12    14
+
+-- returns all words in a string
+print(re.match("the number 423 is odd", "({%a+} / .)*"))
+--&gt; the    number    is    odd
+
+-- returns the first numeral in a string
+print(re.match("the number 423 is odd", "s &lt;- {%d+} / . s"))
+--&gt; 423
+
+-- substitutes a dot for each vowel in a string
+print(re.gsub("hello World", "[aeiou]", "."))
+--&gt; h.ll. W.rld
+</pre>
+
+
+<h3>Balanced parentheses</h3>
+<p>
+The following call will produce the same pattern produced by the
+Lua expression in the
+<a href="lpeg.html#balanced">balanced parentheses</a> example:
+</p>
+<pre class="example">
+b = re.compile[[  balanced &lt;- "(" ([^()] / balanced)* ")"  ]]
+</pre>
+
+<h3>String reversal</h3>
+<p>
+The next example reverses a string:
+</p>
+<pre class="example">
+rev = re.compile[[ R &lt;- (!.) -&gt; '' / ({.} R) -&gt; '%2%1']]
+print(rev:match"0123456789")   --&gt; 9876543210
+</pre>
+
+<h3>CSV decoder</h3>
+<p>
+The next example replicates the <a href="lpeg.html#CSV">CSV decoder</a>:
+</p>
+<pre class="example">
+record = re.compile[[
+  record &lt;- {| field (',' field)* |} (%nl / !.)
+  field &lt;- escaped / nonescaped
+  nonescaped &lt;- { [^,"%nl]* }
+  escaped &lt;- '"' {~ ([^"] / '""' -&gt; '"')* ~} '"'
+]]
+</pre>
+
+<h3>Lua's long strings</h3>
+<p>
+The next example matches Lua long strings:
+</p>
+<pre class="example">
+c = re.compile([[
+  longstring &lt;- ('[' {:eq: '='* :} '[' close)
+  close &lt;- ']' =eq ']' / . close
+]])
+
+print(c:match'[==[]]===]]]]==]===[]')   --&gt; 17
+</pre>
+
+<h3>Abstract Syntax Trees</h3>
+<p>
+This example shows a simple way to build an
+abstract syntax tree (AST) for a given grammar.
+To keep our example simple,
+let us consider the following grammar
+for lists of names:
+</p>
+<pre class="example">
+p = re.compile[[
+      listname &lt;- (name s)*
+      name &lt;- [a-z][a-z]*
+      s &lt;- %s*
+]]
+</pre>
+<p>
+Now, we will add captures to build a corresponding AST.
+As a first step, the pattern will build a table to
+represent each non terminal;
+terminals will be represented by their corresponding strings:
+</p>
+<pre class="example">
+c = re.compile[[
+      listname &lt;- {| (name s)* |}
+      name &lt;- {| {[a-z][a-z]*} |}
+      s &lt;- %s*
+]]
+</pre>
+<p>
+Now, a match against <code>"hi hello bye"</code>
+results in the table
+<code>{{"hi"}, {"hello"}, {"bye"}}</code>.
+</p>
+<p>
+For such a simple grammar,
+this AST is more than enough;
+actually, the tables around each single name
+are already overkilling.
+More complex grammars,
+however, may need some more structure.
+Specifically,
+it would be useful if each table had
+a <code>tag</code> field telling what non terminal
+that table represents.
+We can add such a tag using
+<a href="lpeg.html#cap-g">named group captures</a>:
+</p>
+<pre class="example">
+x = re.compile[[
+      listname <- {| {:tag: '' -> 'list':} (name s)* |}
+      name <- {| {:tag: '' -> 'id':} {[a-z][a-z]*} |}
+      s <- ' '*
+]]
+</pre>
+<p>
+With these group captures,
+a match against <code>"hi hello bye"</code>
+results in the following table:
+</p>
+<pre class="example">
+{tag="list",
+  {tag="id", "hi"},
+  {tag="id", "hello"},
+  {tag="id", "bye"}
+}
+</pre>
+
+
+<h3>Indented blocks</h3>
+<p>
+This example breaks indented blocks into tables,
+respecting the indentation:
+</p>
+<pre class="example">
+p = re.compile[[
+  block &lt;- {| {:ident:' '*:} line
+           ((=ident !' ' line) / &amp;(=ident ' ') block)* |}
+  line &lt;- {[^%nl]*} %nl
+]]
+</pre>
+<p>
+As an example,
+consider the following text:
+</p>
+<pre class="example">
+t = p:match[[
+first line
+  subline 1
+  subline 2
+second line
+third line
+  subline 3.1
+    subline 3.1.1
+  subline 3.2
+]]
+</pre>
+<p>
+The resulting table <code>t</code> will be like this:
+</p>
+<pre class="example">
+   {'first line'; {'subline 1'; 'subline 2'; ident = '  '};
+    'second line';
+    'third line'; { 'subline 3.1'; {'subline 3.1.1'; ident = '    '};
+                    'subline 3.2'; ident = '  '};
+    ident = ''}
+</pre>
+
+<h3>Macro expander</h3>
+<p>
+This example implements a simple macro expander.
+Macros must be defined as part of the pattern,
+following some simple rules:
+</p>
+<pre class="example">
+p = re.compile[[
+      text &lt;- {~ item* ~}
+      item &lt;- macro / [^()] / '(' item* ')'
+      arg &lt;- ' '* {~ (!',' item)* ~}
+      args &lt;- '(' arg (',' arg)* ')'
+      -- now we define some macros
+      macro &lt;- ('apply' args) -&gt; '%1(%2)'
+             / ('add' args) -&gt; '%1 + %2'
+             / ('mul' args) -&gt; '%1 * %2'
+]]
+
+print(p:match"add(mul(a,b), apply(f,x))")   --&gt; a * b + f(x)
+</pre>
+<p>
+A <code>text</code> is a sequence of items,
+wherein we apply a substitution capture to expand any macros.
+An <code>item</code> is either a macro,
+any character different from parentheses,
+or a parenthesized expression.
+A macro argument (<code>arg</code>) is a sequence
+of items different from a comma.
+(Note that a comma may appear inside an item,
+e.g., inside a parenthesized expression.)
+Again we do a substitution capture to expand any macro
+in the argument before expanding the outer macro.
+<code>args</code> is a list of arguments separated by commas.
+Finally we define the macros.
+Each macro is a string substitution;
+it replaces the macro name and its arguments by its corresponding string,
+with each <code>%</code><em>n</em> replaced by the <em>n</em>-th argument.
+</p>
+
+<h3>Patterns</h3>
+<p>
+This example shows the complete syntax
+of patterns accepted by <code>re</code>.
+</p>
+<pre class="example">
+p = [=[
+
+pattern         &lt;- exp !.
+exp             &lt;- S (grammar / alternative)
+
+alternative     &lt;- seq ('/' S seq)*
+seq             &lt;- prefix*
+prefix          &lt;- '&amp;' S prefix / '!' S prefix / suffix
+suffix          &lt;- primary S (([+*?]
+                            / '^' [+-]? num
+                            / '-&gt;' S (string / '{}' / name)
+                            / '&gt&gt;' S name
+                            / '=&gt;' S name) S)*
+
+primary         &lt;- '(' exp ')' / string / class / defined
+                 / '{:' (name ':')? exp ':}'
+                 / '=' name
+                 / '{}'
+                 / '{~' exp '~}'
+                 / '{|' exp '|}'
+                 / '{' exp '}'
+                 / '.'
+                 / name S !arrow
+                 / '&lt;' name '&gt;'          -- old-style non terminals
+
+grammar         &lt;- definition+
+definition      &lt;- name S arrow exp
+
+class           &lt;- '[' '^'? item (!']' item)* ']'
+item            &lt;- defined / range / .
+range           &lt;- . '-' [^]]
+
+S               &lt;- (%s / '--' [^%nl]*)*   -- spaces and comments
+name            &lt;- [A-Za-z_][A-Za-z0-9_]*
+arrow           &lt;- '&lt;-'
+num             &lt;- [0-9]+
+string          &lt;- '"' [^"]* '"' / "'" [^']* "'"
+defined         &lt;- '%' name
+
+]=]
+
+print(re.match(p, p))   -- a self description must match itself
+</pre>
+
+
+
+<h2><a name="license">License</a></h2>
+
+<p>
+This module is part of the <a href="lpeg.html">LPeg</a> package and shares
+its <a href="lpeg.html#license">license</a>.
+
+
+</div> <!-- id="content" -->
+
+</div> <!-- id="main" -->
+
+</div> <!-- id="container" -->
+
+</body>
+</html> 

src/lua/lpeg-1.1.0/re.lua

@@ -0,0 +1,270 @@
+--
+-- Copyright 2007-2023, Lua.org & PUC-Rio  (see 'lpeg.html' for license)
+-- written by Roberto Ierusalimschy
+--
+
+-- imported functions and modules
+local tonumber, type, print, error = tonumber, type, print, error
+local setmetatable = setmetatable
+local m = require"lpeg"
+
+-- 'm' will be used to parse expressions, and 'mm' will be used to
+-- create expressions; that is, 're' runs on 'm', creating patterns
+-- on 'mm'
+local mm = m
+
+-- patterns' metatable
+local mt = getmetatable(mm.P(0))
+
+
+local version = _VERSION
+
+-- No more global accesses after this point
+_ENV = nil     -- does no harm in Lua 5.1
+
+
+local any = m.P(1)
+
+
+-- Pre-defined names
+local Predef = { nl = m.P"\n" }
+
+
+local mem
+local fmem
+local gmem
+
+
+local function updatelocale ()
+  mm.locale(Predef)
+  Predef.a = Predef.alpha
+  Predef.c = Predef.cntrl
+  Predef.d = Predef.digit
+  Predef.g = Predef.graph
+  Predef.l = Predef.lower
+  Predef.p = Predef.punct
+  Predef.s = Predef.space
+  Predef.u = Predef.upper
+  Predef.w = Predef.alnum
+  Predef.x = Predef.xdigit
+  Predef.A = any - Predef.a
+  Predef.C = any - Predef.c
+  Predef.D = any - Predef.d
+  Predef.G = any - Predef.g
+  Predef.L = any - Predef.l
+  Predef.P = any - Predef.p
+  Predef.S = any - Predef.s
+  Predef.U = any - Predef.u
+  Predef.W = any - Predef.w
+  Predef.X = any - Predef.x
+  mem = {}    -- restart memoization
+  fmem = {}
+  gmem = {}
+  local mt = {__mode = "v"}
+  setmetatable(mem, mt)
+  setmetatable(fmem, mt)
+  setmetatable(gmem, mt)
+end
+
+
+updatelocale()
+
+
+
+local I = m.P(function (s,i) print(i, s:sub(1, i-1)); return i end)
+
+
+local function patt_error (s, i)
+  local msg = (#s < i + 20) and s:sub(i)
+                             or s:sub(i,i+20) .. "..."
+  msg = ("pattern error near '%s'"):format(msg)
+  error(msg, 2)
+end
+
+local function mult (p, n)
+  local np = mm.P(true)
+  while n >= 1 do
+    if n%2 >= 1 then np = np * p end
+    p = p * p
+    n = n/2
+  end
+  return np
+end
+
+local function equalcap (s, i, c)
+  if type(c) ~= "string" then return nil end
+  local e = #c + i
+  if s:sub(i, e - 1) == c then return e else return nil end
+end
+
+
+local S = (Predef.space + "--" * (any - Predef.nl)^0)^0
+
+local name = m.R("AZ", "az", "__") * m.R("AZ", "az", "__", "09")^0
+
+local arrow = S * "<-"
+
+local seq_follow = m.P"/" + ")" + "}" + ":}" + "~}" + "|}" + (name * arrow) + -1
+
+name = m.C(name)
+
+
+-- a defined name only have meaning in a given environment
+local Def = name * m.Carg(1)
+
+
+local function getdef (id, defs)
+  local c = defs and defs[id]
+  if not c then error("undefined name: " .. id) end
+  return c
+end
+
+-- match a name and return a group of its corresponding definition
+-- and 'f' (to be folded in 'Suffix')
+local function defwithfunc (f)
+  return m.Cg(Def / getdef * m.Cc(f))
+end
+
+
+local num = m.C(m.R"09"^1) * S / tonumber
+
+local String = "'" * m.C((any - "'")^0) * "'" +
+               '"' * m.C((any - '"')^0) * '"'
+
+
+local defined = "%" * Def / function (c,Defs)
+  local cat =  Defs and Defs[c] or Predef[c]
+  if not cat then error ("name '" .. c .. "' undefined") end
+  return cat
+end
+
+local Range = m.Cs(any * (m.P"-"/"") * (any - "]")) / mm.R
+
+local item = (defined + Range + m.C(any)) / m.P
+
+local Class =
+    "["
+  * (m.C(m.P"^"^-1))    -- optional complement symbol
+  * (item * ((item % mt.__add) - "]")^0) /
+                          function (c, p) return c == "^" and any - p or p end
+  * "]"
+
+local function adddef (t, k, exp)
+  if t[k] then
+    error("'"..k.."' already defined as a rule")
+  else
+    t[k] = exp
+  end
+  return t
+end
+
+local function firstdef (n, r) return adddef({n}, n, r) end
+
+
+local function NT (n, b)
+  if not b then
+    error("rule '"..n.."' used outside a grammar")
+  else return mm.V(n)
+  end
+end
+
+
+local exp = m.P{ "Exp",
+  Exp = S * ( m.V"Grammar"
+            + m.V"Seq" * ("/" * S * m.V"Seq" % mt.__add)^0 );
+  Seq = (m.Cc(m.P"") * (m.V"Prefix" % mt.__mul)^0)
+        * (#seq_follow + patt_error);
+  Prefix = "&" * S * m.V"Prefix" / mt.__len
+         + "!" * S * m.V"Prefix" / mt.__unm
+         + m.V"Suffix";
+  Suffix = m.V"Primary" * S *
+          ( ( m.P"+" * m.Cc(1, mt.__pow)
+            + m.P"*" * m.Cc(0, mt.__pow)
+            + m.P"?" * m.Cc(-1, mt.__pow)
+            + "^" * ( m.Cg(num * m.Cc(mult))
+                    + m.Cg(m.C(m.S"+-" * m.R"09"^1) * m.Cc(mt.__pow))
+                    )
+            + "->" * S * ( m.Cg((String + num) * m.Cc(mt.__div))
+                         + m.P"{}" * m.Cc(nil, m.Ct)
+                         + defwithfunc(mt.__div)
+                         )
+            + "=>" * S * defwithfunc(mm.Cmt)
+            + ">>" * S * defwithfunc(mt.__mod)
+            + "~>" * S * defwithfunc(mm.Cf)
+            ) % function (a,b,f) return f(a,b) end * S
+          )^0;
+  Primary = "(" * m.V"Exp" * ")"
+            + String / mm.P
+            + Class
+            + defined
+            + "{:" * (name * ":" + m.Cc(nil)) * m.V"Exp" * ":}" /
+                     function (n, p) return mm.Cg(p, n) end
+            + "=" * name / function (n) return mm.Cmt(mm.Cb(n), equalcap) end
+            + m.P"{}" / mm.Cp
+            + "{~" * m.V"Exp" * "~}" / mm.Cs
+            + "{|" * m.V"Exp" * "|}" / mm.Ct
+            + "{" * m.V"Exp" * "}" / mm.C
+            + m.P"." * m.Cc(any)
+            + (name * -arrow + "<" * name * ">") * m.Cb("G") / NT;
+  Definition = name * arrow * m.V"Exp";
+  Grammar = m.Cg(m.Cc(true), "G") *
+            ((m.V"Definition" / firstdef) * (m.V"Definition" % adddef)^0) / mm.P
+}
+
+local pattern = S * m.Cg(m.Cc(false), "G") * exp / mm.P * (-any + patt_error)
+
+
+local function compile (p, defs)
+  if mm.type(p) == "pattern" then return p end   -- already compiled
+  local cp = pattern:match(p, 1, defs)
+  if not cp then error("incorrect pattern", 3) end
+  return cp
+end
+
+local function match (s, p, i)
+  local cp = mem[p]
+  if not cp then
+    cp = compile(p)
+    mem[p] = cp
+  end
+  return cp:match(s, i or 1)
+end
+
+local function find (s, p, i)
+  local cp = fmem[p]
+  if not cp then
+    cp = compile(p) / 0
+    cp = mm.P{ mm.Cp() * cp * mm.Cp() + 1 * mm.V(1) }
+    fmem[p] = cp
+  end
+  local i, e = cp:match(s, i or 1)
+  if i then return i, e - 1
+  else return i
+  end
+end
+
+local function gsub (s, p, rep)
+  local g = gmem[p] or {}   -- ensure gmem[p] is not collected while here
+  gmem[p] = g
+  local cp = g[rep]
+  if not cp then
+    cp = compile(p)
+    cp = mm.Cs((cp / rep + 1)^0)
+    g[rep] = cp
+  end
+  return cp:match(s)
+end
+
+
+-- exported names
+local re = {
+  compile = compile,
+  match = match,
+  find = find,
+  gsub = gsub,
+  updatelocale = updatelocale,
+}
+
+if version == "Lua 5.1" then _G.re = re end
+
+return re