Change .location, NewLocation, and Location to lazily compute location information

console/src/main/scala/io/joern/console/scan/package.scala

@@ -100,7 +100,7 @@ package object scan {
 
   /** Print human readable list of findings to standard out.
     */
-  def outputFindings(cpg: Cpg)(implicit finder: NodeExtensionFinder): Unit = {
+  def outputFindings(cpg: Cpg)(implicit locationCreator: LocationCreator): Unit = {
     cpg.finding.sortBy(_.score.toInt).foreach { finding =>
       val evidence = finding.evidence.headOption
         .map { e =>

joern-cli/frontends/javasrc2cpg/src/test/scala/io/joern/javasrc2cpg/querying/CodeDumperTests.scala

@@ -7,9 +7,6 @@ import io.shiftleft.semanticcpg.language.*
 class CodeDumperTests extends JavaSrcCode2CpgFixture {
 
   "a Java source code CPG" should {
-
-    implicit val finder: NodeExtensionFinder = DefaultNodeExtensionFinder
-
     val cpg = code("""
         |public class Foo {
         |

joern-cli/src/main/scala/io/joern/joerncli/JoernScan.scala

@@ -9,7 +9,7 @@ import io.joern.joerncli.Scan.{allTag, defaultTag}
 import io.joern.joerncli.console.ReplBridge
 import io.shiftleft.semanticcpg.utils.FileUtil.*
 import io.shiftleft.codepropertygraph.generated.Languages
-import io.shiftleft.semanticcpg.language.{DefaultNodeExtensionFinder, NodeExtensionFinder}
+import io.shiftleft.semanticcpg.language.{DefaultNodeExtensionFinder, NodeExtensionFinder, locationCreator}
 import io.shiftleft.semanticcpg.layers.{LayerCreator, LayerCreatorContext, LayerCreatorOptions}
 import io.shiftleft.semanticcpg.utils.FileUtil
 import org.json4s.native.Serialization
@@ -259,8 +259,6 @@ class ScanOptions(var maxCallDepth: Int, var names: Array[String], var tags: Arr
     extends LayerCreatorOptions {}
 
 class Scan(options: ScanOptions)(implicit engineContext: EngineContext) extends LayerCreator {
-  implicit val finder: NodeExtensionFinder = DefaultNodeExtensionFinder
-
   override val overlayName: String = Scan.overlayName
   override val description: String = Scan.description
 

semanticcpg/src/main/scala/io/shiftleft/semanticcpg/codedumper/CodeDumper.scala

@@ -1,7 +1,7 @@
 package io.shiftleft.semanticcpg.codedumper
 
 import io.shiftleft.codepropertygraph.generated.Languages
-import io.shiftleft.codepropertygraph.generated.nodes.{Expression, Local, Method, NewLocation}
+import io.shiftleft.codepropertygraph.generated.nodes.{Expression, Local, Method}
 import io.shiftleft.semanticcpg.language.*
 import io.shiftleft.utils.IOUtils
 import org.slf4j.{Logger, LoggerFactory}
@@ -39,7 +39,7 @@ object CodeDumper {
   /** Dump string representation of code at given `location`.
     */
   def dump(
-    location: NewLocation,
+    location: LocationInfo,
     language: Option[String],
     rootPath: Option[String],
     highlight: Boolean,

semanticcpg/src/main/scala/io/shiftleft/semanticcpg/language/Location.scala

@@ -0,0 +1,124 @@
+package io.shiftleft.semanticcpg.language
+
+import io.shiftleft.codepropertygraph.generated.nodes.*
+import io.shiftleft.codepropertygraph.generated.Properties
+import io.shiftleft.semanticcpg.language.*
+
+import scala.annotation.tailrec
+
+trait LocationInfo {
+  def node: Option[AbstractNode]
+  def symbol: String
+  def nodeLabel: String
+  def lineNumber: Option[Int]
+  def methodFullName: String
+  def methodShortName: String
+  def packageName: String
+  def className: String
+  def classShortName: String
+  def filename: String
+}
+
+// Previously, the non-stored Location node generated
+// from the CPG schema was named NewLocation. The
+// addition of lazily computed location info in this
+// file brought the deletion of these CPG nodes. This
+// is not directly used in joern, but intended to not
+// break usage in users of the API. Users directly
+// importing NewLocation from the CPG package rather
+// will still have a break.
+@deprecated("NewLocation is deprecated, prefer LocationInfo")
+type NewLocation = LocationInfo
+
+trait HasLocation extends Any {
+  def location: LocationInfo
+}
+
+// This design: 1) lets implementors put all logic for
+// determining how to fill out `LocationInfo` for each
+// type of node in one central place, and 2) allow consumers
+// of joern, which may also provide different node types
+// and/or location information, to provide alternate or
+// extended implementations of LocationCreator.
+trait LocationCreator {
+  implicit def apply(node: StoredNode): LocationInfo
+}
+
+object EmptyLocation extends LocationInfo {
+  override def node: Option[AbstractNode] = None
+  override def symbol: String             = "<empty>"
+  override def nodeLabel: String          = "<empty>"
+  override def lineNumber: Option[Int]    = None
+  override def methodFullName: String     = "<empty>"
+  override def methodShortName: String    = "<empty>"
+  override def packageName: String        = "<empty>"
+  override def className: String          = "<empty>"
+  override def classShortName: String     = "<empty>"
+  override def filename: String           = "<empty>"
+}
+
+object LazyLocation extends LocationCreator {
+  implicit def apply(node: StoredNode): LocationInfo = {
+    new LazyLocation(node)
+  }
+}
+
+implicit val locationCreator: LocationCreator = LazyLocation
+
+class LazyLocation(storedNode: StoredNode) extends LocationInfo {
+  def node: Option[AbstractNode] = Some(storedNode)
+
+  def symbol: String = {
+    storedNode match {
+      case _: Call | _: Literal | _: MethodRef =>
+        storedNode.property(Properties.Code)
+      case _: Identifier | _: Local | _: MethodParameterIn | _: MethodParameterOut | _: Method =>
+        storedNode.property(Properties.Name)
+      case _: MethodReturn => "$ret"
+      case _               => defaultString
+    }
+  }
+
+  def nodeLabel: String = storedNode.label
+
+  def lineNumber: Option[Int] = storedNode match {
+    case astNode: AstNode => astNode.lineNumber
+    case _                => None
+  }
+
+  def methodFullName: String = method.fullName
+
+  def methodShortName: String = method.name
+
+  def packageName: String = namespaceOption.getOrElse(defaultString)
+
+  def className: String = typeOption.map(_.fullName).getOrElse(defaultString)
+
+  def classShortName: String = typeOption.map(_.name).getOrElse(defaultString)
+
+  def filename: String = if (method.filename.isEmpty) "N/A" else method.filename
+
+  final protected val defaultString = "<empty>";
+
+  private lazy val typeOption = findParentTypeDecl(method)
+
+  private lazy val namespaceOption = for {
+    tpe            <- typeOption
+    namespaceBlock <- tpe.namespaceBlock
+    namespace      <- namespaceBlock._namespaceViaRefOut.nextOption()
+  } yield namespace.name
+
+  private lazy val method: Method = storedNode match {
+    case cfgNode: CfgNode => cfgNode.method
+    case local: Local     => local.method.head
+  }
+
+  @tailrec
+  private def findParentTypeDecl(node: StoredNode): Option[TypeDecl] = {
+    node._astIn.iterator.nextOption() match {
+      case Some(head) if head.isInstanceOf[TypeDecl] => Some(head.asInstanceOf[TypeDecl])
+      case Some(head)                                => findParentTypeDecl(head)
+      case None                                      => None
+    }
+  }
+}

semanticcpg/src/main/scala/io/shiftleft/semanticcpg/language/NodeSteps.scala

@@ -44,7 +44,7 @@ class NodeSteps[NodeType <: StoredNode](val traversal: Iterator[NodeType]) exten
       |on the user's side.
       |"""
   )
-  def location(implicit finder: NodeExtensionFinder): Iterator[NewLocation] =
+  def location(implicit locationCreator: LocationCreator): Iterator[LocationInfo] =
     traversal.map(_.location)
 
   @Doc(
@@ -56,7 +56,7 @@ class NodeSteps[NodeType <: StoredNode](val traversal: Iterator[NodeType]) exten
       |This only works for source frontends.
       |"""
   )
-  def dump(implicit finder: NodeExtensionFinder): List[String] =
+  def dump(implicit locationCreator: LocationCreator): List[String] =
     _dump(highlight = true)
 
   @Doc(
@@ -67,10 +67,10 @@ class NodeSteps[NodeType <: StoredNode](val traversal: Iterator[NodeType]) exten
       |to the expression. No color highlighting.
       |"""
   )
-  def dumpRaw(implicit finder: NodeExtensionFinder): List[String] =
+  def dumpRaw(implicit locationCreator: LocationCreator): List[String] =
     _dump(highlight = false)
 
-  private def _dump(highlight: Boolean)(implicit finder: NodeExtensionFinder): List[String] = {
+  private def _dump(highlight: Boolean)(implicit locationCreator: LocationCreator): List[String] = {
     // initialized on first element as we need the graph for retrieving the metaData node.
     // TODO: there should be a step to retrieve the metaData node for any node
     //  so we could avoid instantiating a new Cpg everytime using dump

semanticcpg/src/main/scala/io/shiftleft/semanticcpg/language/nodemethods/CallMethods.scala

@@ -1,6 +1,6 @@
 package io.shiftleft.semanticcpg.language.nodemethods
 
-import io.shiftleft.codepropertygraph.generated.nodes.{Call, Expression, NewLocation}
+import io.shiftleft.codepropertygraph.generated.nodes.{Call, Expression}
 import io.shiftleft.codepropertygraph.generated.DispatchTypes
 import io.shiftleft.semanticcpg.NodeExtension
 import io.shiftleft.semanticcpg.language.*
@@ -45,6 +45,7 @@ class CallMethods(val node: Call) extends AnyVal with NodeExtension with HasLoca
     node.astChildren.isBlock.maxByOption(_.order).iterator.expressionDown
   }
 
-  override def location: NewLocation =
-    LocationCreator.defaultCreateLocation(node)
+  override def location: LocationInfo = {
+    node.location
+  }
 }

semanticcpg/src/main/scala/io/shiftleft/semanticcpg/language/nodemethods/IdentifierMethods.scala

@@ -1,12 +1,12 @@
 package io.shiftleft.semanticcpg.language.nodemethods
 
-import io.shiftleft.codepropertygraph.generated.nodes.{Declaration, Identifier, NewLocation}
+import io.shiftleft.codepropertygraph.generated.nodes.{Declaration, Identifier}
 import io.shiftleft.semanticcpg.NodeExtension
 import io.shiftleft.semanticcpg.language.*
 
 class IdentifierMethods(val identifier: Identifier) extends AnyVal with NodeExtension with HasLocation {
-  override def location: NewLocation = {
-    LocationCreator.defaultCreateLocation(identifier)
+  override def location: LocationInfo = {
+    identifier.location
   }
 
   def isModuleVariable: Boolean = identifier.refOut.collectAll[Declaration].method.isModule.nonEmpty