From e0abee7f9f7c23fbeeeceb3345ed90ebcd78bbf7 Mon Sep 17 00:00:00 2001
From: Paul Beckingham <paul@beckingham.net>
Date: Mon, 13 Jun 2011 23:06:54 -0400
Subject: [PATCH] Expressions

- Command line arguments are now handled a little differently.  Each
  argument is subjected to further splitting, to break up expressions
  early in the process, prior to categorization.
- Patterns are now treated as quoted string, where the quote character
  is /.
---
 src/Arguments.cpp  | 59 ++++++++++++++++++++++++++++++++++++++--------
 src/Arguments.h    |  1 +
 src/DOM.cpp        |  4 +++-
 src/Expression.cpp | 23 ++++++++++--------
 4 files changed, 66 insertions(+), 21 deletions(-)

diff --git a/src/Arguments.cpp b/src/Arguments.cpp
index 803d4f257..f765eacd7 100644
--- a/src/Arguments.cpp
+++ b/src/Arguments.cpp
@@ -148,7 +148,17 @@ Arguments::~Arguments ()
 void Arguments::capture (int argc, const char** argv)
 {
   for (int i = 0; i < argc; ++i)
-    this->push_back (std::make_pair (argv[i], ""));
+  {
+    std::vector <std::string> parts;
+    if (is_multipart (argv[i], parts))
+    {
+      std::vector <std::string>::iterator part;
+      for (part = parts.begin (); part != parts.end (); ++part)
+        this->push_back (std::make_pair (*part, ""));
+    }
+    else
+      this->push_back (std::make_pair (argv[i], ""));
+  }
 
   categorize ();
 }
@@ -157,7 +167,16 @@ void Arguments::capture (int argc, const char** argv)
 // Add a pair with a category of "".
 void Arguments::capture (const std::string& arg)
 {
-  this->push_back (std::make_pair (arg, ""));
+  std::vector <std::string> parts;
+  if (is_multipart (arg, parts))
+  {
+    std::vector <std::string>::iterator part;
+    for (part = parts.begin (); part != parts.end (); ++part)
+      this->push_back (std::make_pair (*part, ""));
+  }
+  else
+    this->push_back (std::make_pair (arg, ""));
+
   categorize ();
 }
 
@@ -617,6 +636,26 @@ bool Arguments::find_command (std::string& command)
   return false;
 }
 
+////////////////////////////////////////////////////////////////////////////////
+bool Arguments::is_multipart (
+  const std::string& input,
+  std::vector <std::string>& parts)
+{
+  parts.clear ();
+  Nibbler n (input);
+  std::string part;
+  while (n.getQuoted ('"', part)  ||
+         n.getQuoted ('\'', part) ||
+         n.getQuoted ('/', part) ||
+         n.getUntilWS (part))
+  {
+    n.skipWS ();
+    parts.push_back (part);
+  }
+
+  return parts.size () > 1 ? true : false;
+}
+
 ////////////////////////////////////////////////////////////////////////////////
 bool Arguments::is_command (
   const std::vector <std::string>& keywords,
@@ -758,12 +797,10 @@ bool Arguments::is_subst (const std::string& input)
 // /<pattern>/
 bool Arguments::is_pattern (const std::string& input)
 {
-  unsigned int length = input.length ();
-
-  if (input[0] == '/'          &&
-      length  > 2              &&
-      input[length - 1] == '/' &&
-      input.find ('/', 1) == length - 1)
+  Nibbler n (input);
+  std::string pattern;
+  if (input.length () > 2 &&
+      n.getQuoted ('/', pattern, true))
     return true;
 
   return false;
@@ -974,7 +1011,8 @@ bool Arguments::extract_attr (
     {
       // Both quoted and unquoted Att's are accepted.
       // Consider removing this for a stricter parse.
-      if (n.getQuoted   ('"', value) ||
+      if (n.getQuoted   ('"', value)  ||
+          n.getQuoted   ('\'', value) ||
           n.getUntilEOS (value))
       {
         return true;
@@ -1031,7 +1069,8 @@ bool Arguments::extract_attmod (
     {
       // Both quoted and unquoted Att's are accepted.
       // Consider removing this for a stricter parse.
-      if (n.getQuoted   ('"', value) ||
+      if (n.getQuoted   ('"', value)  ||
+          n.getQuoted   ('\'', value) ||
           n.getUntilEOS (value))
       {
         return true;
diff --git a/src/Arguments.h b/src/Arguments.h
index e0e042328..e7d1999c4 100644
--- a/src/Arguments.h
+++ b/src/Arguments.h
@@ -56,6 +56,7 @@ public:
 
   bool find_command (std::string&);
 
+  static bool is_multipart (const std::string&, std::vector <std::string>&);
   static bool is_command (const std::vector <std::string>&, std::string&);
   static bool is_attr (const std::string&);
   static bool is_attmod (const std::string&);
diff --git a/src/DOM.cpp b/src/DOM.cpp
index 0a6e3957f..b1e7b38e1 100644
--- a/src/DOM.cpp
+++ b/src/DOM.cpp
@@ -230,7 +230,9 @@ bool DOM::is_primitive (const std::string& input)
 
   // String?
   Nibbler n (input);
-  if (n.getQuoted ('"', s) && n.depleted ())
+  if ((n.getQuoted ('"', s) ||
+       n.getQuoted ('\'', s)) &&
+      n.depleted ())
     return true;
 
   // Number?
diff --git a/src/Expression.cpp b/src/Expression.cpp
index 4cadc10a8..e71468b70 100644
--- a/src/Expression.cpp
+++ b/src/Expression.cpp
@@ -45,26 +45,26 @@ Expression::Expression (Arguments& arguments)
 {
   _args.dump ("Expression::Expression");
 
-  if (is_new_style () && context.config.getBoolean ("expressions"))
-  {
+  bool new_style = is_new_style () && context.config.getBoolean ("expressions");
+  if (new_style)
     context.debug ("Filter --> new");
-    expand_sequence ();
-    expand_tokens ();
-    postfix ();
-  }
   else
-  {
     context.debug ("Filter --> old");
-    expand_sequence ();
+
+  expand_sequence ();
+
+  if (new_style)
+  {
     implicit_and ();
     expand_tag ();
     expand_pattern ();
     expand_attr ();
     expand_attmod ();
     expand_word ();
-    expand_tokens ();
-    postfix ();
   }
+
+  expand_tokens ();
+  postfix ();
 }
 
 ////////////////////////////////////////////////////////////////////////////////
@@ -226,6 +226,9 @@ void Expression::expand_tokens ()
             n.getQuoted ('\'', s, true))
           temp.push_back (std::make_pair (s, "string"));
 
+        else if (n.getQuoted ('/', s, true))
+          temp.push_back (std::make_pair (s, "pattern"));
+
         else if (n.getOneOf (operators, s))
           temp.push_back (std::make_pair (s, "op"));