Update to LPeg 0.10.2

LuaDist · Aug 19, 2011 · 14b436c · 14b436c
1 parent 5019d49
commit 14b436c
Show file tree

Hide file tree

Showing 7 changed files with 105 additions and 19 deletions.
diff --git a/HISTORY b/HISTORY
@@ -10,6 +10,8 @@ HISTORY for LPeg 0.10
   + 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;

diff --git a/dist.info b/dist.info
@@ -1,7 +1,7 @@
 --- This file is part of LuaDist project
 
 name = "lpeg"
-version = "0.10"
+version = "0.10.2"
 
 desc = "Parsing Expression Grammars For Lua"
 author = "Roberto Ierusalimschy"

diff --git a/lpeg.c b/lpeg.c
@@ -1,5 +1,5 @@
 /*
-** $Id: lpeg.c,v 1.112 2010/11/03 17:07:50 roberto Exp $
+** $Id: lpeg.c,v 1.114 2011/02/16 15:02:20 roberto Exp $
 ** LPeg - PEG pattern matching for Lua
 ** Copyright 2007, Lua.org & PUC-Rio  (see 'lpeg.html' for license)
 ** written by Roberto Ierusalimschy
@@ -209,6 +209,8 @@ typedef struct Capture {
 #define setchar(st,c)	((st)[(c) >> 3] |= (1 << ((c) & 7)))
 
 
+static int target (Instruction *p, int i);
+
 
 static int sizei (const Instruction *i) {
   switch((Opcode)i->i.code) {
@@ -245,7 +247,7 @@ static int getposition (lua_State *L, int t, int i) {
 
 /*
 ** {======================================================
-** Printing patterns
+** Printing patterns (for debugging)
 ** =======================================================
 */
 
@@ -350,6 +352,7 @@ static void printpatt (Instruction *p) {
 }
 
 
+#if 0
 static void printcap (Capture *cap) {
   printcapkind(cap->kind);
   printf(" (idx: %d - size: %d) -> %p\n", cap->idx, cap->siz, cap->s);
@@ -359,6 +362,7 @@ static void printcap (Capture *cap) {
 static void printcaplist (Capture *cap) {
   for (; cap->s; cap++) printcap(cap);
 }
+#endif
 
 /* }====================================================== */
 
@@ -785,7 +789,8 @@ static void checkrule (lua_State *L, Instruction *op, int from, int to,
   for (i = from; i < to; i += sizei(op + i)) {
     if (op[i].i.code == IPartialCommit && op[i].i.offset < 0) {  /* loop? */
       int start = dest(op, i);
-      assert(op[start - 1].i.code == IChoice && dest(op, start - 1) == i + 1);
+      assert(op[start - 1].i.code == IChoice &&
+             dest(op, start - 1) == target(op, i + 1));
       if (start <= lastopen) {  /* loop does contain an open call? */
         if (!verify(L, op, op + start, op + i, postable, rule)) /* check body */
           luaL_error(L, "possible infinite loop in %s", val2str(L, rule));
@@ -2345,7 +2350,7 @@ static int matchl (lua_State *L) {
 }
 
 
-static struct luaL_reg pattreg[] = {
+static struct luaL_Reg pattreg[] = {
   {"match", matchl},
   {"print", printpat_l},
   {"locale", locale_l},
@@ -2371,7 +2376,7 @@ static struct luaL_reg pattreg[] = {
 };
 
 
-static struct luaL_reg metapattreg[] = {
+static struct luaL_Reg metapattreg[] = {
   {"__add", union_l},
   {"__pow", star_l},
   {"__sub", diff_l},

diff --git a/lpeg.html b/lpeg.html
@@ -10,7 +10,7 @@
 </head>
 <body>
 
-<!-- $Id: lpeg.html,v 1.60 2010/11/03 17:42:07 roberto Exp $ -->
+<!-- $Id: lpeg.html,v 1.64 2011/02/16 15:01:58 roberto Exp $ -->
 
 <div id="container">
 
@@ -705,7 +705,7 @@ <h3><a name="cap-b"></a><code>lpeg.Cb (name)</code></h3>
 </p>
 
 
-<h3><a name="cap-cc"></a><code>lpeg.Cc ({value})</code></h3>
+<h3><a name="cap-cc"></a><code>lpeg.Cc ([value, ...])</code></h3>
 <p>
 Creates a <em>constant capture</em>.
 This pattern matches the empty string and
@@ -1243,8 +1243,9 @@ <h3>Lua's long strings</h3>
 </p>
 
 <pre class="example">
-open = "[" * lpeg.Cg(lpeg.P"="^0, "init") * "[" * lpeg.P"\n"^-1
+equals = lpeg.P"="^0
-close = "]" * lpeg.C(lpeg.P"="^0) * "]"
+open = "[" * lpeg.Cg(equals, "init") * "[" * lpeg.P"\n"^-1
+close = "]" * lpeg.C(equals) * "]"
 closeeq = lpeg.Cmt(close * lpeg.Cb("init"), function (s, i, a, b) return a == b end)
 string = open * lpeg.C((lpeg.P(1) - closeeq)^0) * close /
   function (s, o) return s end
@@ -1368,7 +1369,7 @@ <h3>Arithmetic expressions</h3>
 <h2><a name="download"></a>Download</h2>
 
 <p>LPeg 
-<a href="http://www.inf.puc-rio.br/~roberto/lpeg/lpeg-0.10.tar.gz">source code</a>.</p>
+<a href="http://www.inf.puc-rio.br/~roberto/lpeg/lpeg-0.10.2.tar.gz">source code</a>.</p>
 
 
 <h2><a name="license">License</a></h2>
@@ -1412,7 +1413,7 @@ <h2><a name="license">License</a></h2>
 
 <div id="about">
 <p><small>
-$Id: lpeg.html,v 1.60 2010/11/03 17:42:07 roberto Exp $
+$Id: lpeg.html,v 1.64 2011/02/16 15:01:58 roberto Exp $
 </small></p>
 </div> <!-- id="about" -->
 

diff --git a/re.html b/re.html
@@ -10,7 +10,7 @@
 </head>
 <body>
 
-<!-- $Id: re.html,v 1.14 2010/11/03 17:07:19 roberto Exp $ -->
+<!-- $Id: re.html,v 1.17 2011/01/10 15:08:06 roberto Exp $ -->
 
 <div id="container">
 
@@ -154,6 +154,13 @@ <h3><code>re.find (subject, pattern [, init])</code></h3>
 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.match (subject, pattern)</code></h3>
 <p>
 Matches the given pattern against the given subject.
@@ -225,13 +232,81 @@ <h3>Lua's long strings</h3>
 </p>
 <pre class="example">
 c = re.compile([[
-  longstring &lt;- ('[' {:eq: '='* :} '[' close) =&gt; void
+  longstring &lt;- ('[' {:eq: '='* :} '[' close) -&gt; void
   close &lt;- ']' =eq ']' / . close
-]], {void = function () return true end})
+]], {void = function () end})
 
 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)* -&gt; {}
+      name &lt;- {[a-z][a-z]*} -&gt; {}
+      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,
@@ -401,7 +476,7 @@ <h2><a name="license">License</a></h2>
 
 <div id="about">
 <p><small>
-$Id: re.html,v 1.14 2010/11/03 17:07:19 roberto Exp $
+$Id: re.html,v 1.17 2011/01/10 15:08:06 roberto Exp $
 </small></p>
 </div> <!-- id="about" -->
 

diff --git a/re.lua b/re.lua
@@ -1,4 +1,4 @@
--- $Id: re.lua,v 1.38 2010/11/03 17:21:07 roberto Exp $
+-- $Id: re.lua,v 1.39 2010/11/04 19:44:18 roberto Exp $
 
 -- imported functions and modules
 local tonumber, type, print, error = tonumber, type, print, error
@@ -102,7 +102,7 @@ local function equalcap (s, i, c)
 end
 
 
-local S = (Predef.space + "--" * (any - Predef.nl)^0)^0
+local S = (m.S(" \f\n\r\t\v") + "--" * (any - Predef.nl)^0)^0
 
 local name = m.R("AZ", "az") * m.R("AZ", "az", "__", "09")^0
 

diff --git a/test.lua b/test.lua
@@ -1,6 +1,6 @@
 #!/usr/bin/env lua5.1
 
--- $Id: test.lua,v 1.81 2010/11/03 17:07:50 roberto Exp $
+-- $Id: test.lua,v 1.82 2010/12/03 14:49:54 roberto Exp $
 
 require"strict"    -- just to be pedantic
 
@@ -348,6 +348,9 @@ checkerr("rule <a table> is not defined", m.match, { m.V({}) }, "")
 print('+')
 
 
+-- bug in 0.10 (rechecking a grammar, after tail-call optimization)
+m.P{ m.P { (m.P(3) + "xuxu")^0 * m.V"xuxu", xuxu = m.P(1) } }
+
 local V = m.V
 
 local Space = m.S(" \n\t")^0