Open heart surgery on IDs/naming

This commit institutes a more structured system for managing namespaces.
It fixes a handful of namespace collision bugs, and also improves
performance by abandoning ungainly string-based IDs.

1 files changed, 35 insertions, 0 deletions

diff --git a/src/main/scala/Chisel/Namespace.scala b/src/main/scala/Chisel/Namespace.scala
new file mode 100644
index 00000000..9d867bdd
--- /dev/null
+++ b/src/main/scala/Chisel/Namespace.scala
@@ -0,0 +1,35 @@
+package Chisel
+
+abstract class Namespace {
+  final def name(n: String): String = {
+    if (nameExists(n)) name(rename(n))
+    else {
+      names += n
+      n
+    }
+  }
+
+  def nameExists(n: String): Boolean
+
+  protected final def nameExistsHere(n: String): Boolean =
+    names contains n
+
+  private def rename(n: String): String = {
+    i += 1
+    s"${n}_${i}"
+  }
+
+  private var i = 0L
+  protected val names = collection.mutable.HashSet[String]()
+}
+
+class RootNamespace(initialNames: String*) extends Namespace {
+  names ++= initialNames
+  def nameExists(n: String) = nameExistsHere(n)
+}
+
+class ChildNamespace(parent: Namespace) extends Namespace {
+  def nameExists(n: String) = nameExistsHere(n) || parent.nameExists(n)
+}
+
+class FIRRTLNamespace extends RootNamespace("mem", "node", "wire", "reg", "inst")