Some new regexp infrastructure.

src/ast.cc

Index: src/ast.cc
diff --git a/src/ast.cc b/src/ast.cc
index 2f21d926c396aec37a05ff4335f82342aa2339fb..5bc68810da773eea493678443d5711b92bbc4208 100644
--- a/src/ast.cc
+++ b/src/ast.cc
@@ -29,6 +29,7 @@
 
 #include "ast.h"
 #include "scopes.h"
+#include "string-stream.h"
 
 namespace v8 { namespace internal {
 
@@ -179,4 +180,156 @@ void Visitor::VisitExpressions(ZoneList<Expression*>* expressions) {
 }
 
 
+// ----------------------------------------------------------------------------
+// Regular expressions
+
+#define MAKE_ACCEPT(Name)                                            \
+  void* RegExp##Name::Accept(RegExpVisitor* visitor, void* data) {   \
+    return visitor->Visit##Name(this, data);                         \
+  }
+FOR_EACH_REG_EXP_NODE_TYPE(MAKE_ACCEPT)
+#undef MAKE_ACCEPT
+
+
+RegExpEmpty RegExpEmpty::kInstance;
+
+
+class RegExpUnparser: public RegExpVisitor {

[Lasse Reichstein, 2008/10/27 13:12:58]

Would be nice with a comment describing the syntax of the unparsed output.

[Christian Plesner Hansen, 2008/10/27 18:57:02]

I can't be bothered to write the full format but I've added a comment that says
that it prints the trees as sexps.  This code should be moved to the test file
later but I think it will be useful for us at this point to have it as part of
the library as well.

+ public:
+  RegExpUnparser();
+  void VisitCharacterRange(CharacterRange that);
+  SmartPointer<char> ToString() { return stream_.ToCString(); }
+#define MAKE_CASE(Name) virtual void* Visit##Name(RegExp##Name*, void* data);
+  FOR_EACH_REG_EXP_NODE_TYPE(MAKE_CASE)
+#undef MAKE_CASE
+ private:
+  StringStream* stream() { return &stream_; }
+  HeapStringAllocator alloc_;
+  StringStream stream_;
+};
+
+
+RegExpUnparser::RegExpUnparser() : stream_(&alloc_) {
+}
+
+
+void* RegExpUnparser::VisitDisjunction(RegExpDisjunction* that, void* data) {
+  stream()->Add("(|");
+  for (int i = 0; i <  that->nodes()->length(); i++) {
+    stream()->Add(" ");
+    that->nodes()->at(i)->Accept(this, data);
+  }
+  stream()->Add(")");
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitAlternative(RegExpAlternative* that, void* data) {
+  stream()->Add("(:");
+  for (int i = 0; i <  that->nodes()->length(); i++) {
+    stream()->Add(" ");
+    that->nodes()->at(i)->Accept(this, data);
+  }
+  stream()->Add(")");
+  return NULL;
+}
+
+
+void RegExpUnparser::VisitCharacterRange(CharacterRange that) {
+  if (that.is_special()) {
+    stream()->Add("&%c", that.from());
+  } else if (that.IsSingleton()) {
+    stream()->Add("%c", that.from());
+  } else {
+    stream()->Add("%c-%c", that.from(), that.to());
+  }
+}
+
+
+
+void* RegExpUnparser::VisitCharacterClass(RegExpCharacterClass* that,
+                                          void* data) {
+  if (that->is_negated())
+    stream()->Add("^");

[Lasse Reichstein, 2008/10/27 13:12:58]

Needs sqiggly quotes around the then-statement or put on same line as the "if".

[Christian Plesner Hansen, 2008/10/27 18:57:02]

Nope, the style guide allows both forms.  "In general, curly braces are not
required for single-line statements, but they are allowed if you like them".  I
don't like them.

+  stream()->Add("[");
+  bool first = true;
+  for (int i = 0; i < that->ranges()->length(); i++) {
+    if (first) first = false;

[Lasse Reichstein, 2008/10/27 13:12:58]

"first" is equivalent to "i == 0", so it can be omitted.

[Christian Plesner Hansen, 2008/10/27 18:57:02]

Good point.

+    else stream()->Add(" ");
+    VisitCharacterRange(that->ranges()->at(i));
+  }
+  stream()->Add("]");
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitAssertion(RegExpAssertion* that, void* data) {
+  switch (that->type()) {
+    case RegExpAssertion::START:
+      stream()->Add("@^");
+      break;
+    case RegExpAssertion::END:
+      stream()->Add("@$");
+      break;
+    case RegExpAssertion::BOUNDARY:
+      stream()->Add("@b");
+      break;
+    case RegExpAssertion::NON_BOUNDARY:
+      stream()->Add("@B");
+      break;
+  }
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitAtom(RegExpAtom* that, void* data) {
+  stream()->Add("'%w'", that->data());
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitQuantifier(RegExpQuantifier* that, void* data) {
+  stream()->Add("(# %i ", that->min());
+  if (that->max() == RegExpQuantifier::kInfinity) {
+    stream()->Add("- ");
+  } else {
+    stream()->Add("%i ", that->max());
+  }
+  stream()->Add(that->is_greedy() ? "g " : "n ");
+  that->body()->Accept(this, data);
+  stream()->Add(")");
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitCapture(RegExpCapture* that, void* data) {
+  stream()->Add("(^ ");
+  that->body()->Accept(this, data);
+  stream()->Add(")");
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitLookahead(RegExpLookahead* that, void* data) {
+  stream()->Add("(-> ");
+  stream()->Add(that->is_positive() ? "+ " : "- ");
+  that->body()->Accept(this, data);
+  stream()->Add(")");
+  return NULL;
+}
+
+
+void* RegExpUnparser::VisitEmpty(RegExpEmpty* that, void* data) {
+  stream()->Put('%');
+  return NULL;
+}
+
+
+SmartPointer<char> RegExpTree::ToString() {
+  RegExpUnparser unparser;
+  Accept(&unparser, NULL);
+  return unparser.ToString();
+}
+
+
 } }  // namespace v8::internal