[Ruby] do-end Blocks (with implicit &Proc and yield handling) #3928

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/AstCreator.scala

@@ -27,7 +27,6 @@ class AstCreator(
     with AstForExpressionsCreator
     with AstForFunctionsCreator
     with AstForTypesCreator
-    with FreshVariableCreator
     with AstSummaryVisitor
     with AstNodeBuilder[RubyNode, AstCreator] {
 

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/AstForExpressionsCreator.scala

@@ -7,9 +7,12 @@ import io.joern.rubysrc2cpg.passes.Defines.{RubyOperators, getBuiltInType}
 import io.joern.x2cpg.{Ast, ValidationMode, Defines as XDefines}
 import io.shiftleft.codepropertygraph.generated.nodes.*
 import io.shiftleft.codepropertygraph.generated.{ControlStructureTypes, DispatchTypes, Operators, PropertyNames}
+import io.joern.rubysrc2cpg.utils.FreshNameGenerator
 
 trait AstForExpressionsCreator(implicit withSchemaValidation: ValidationMode) { this: AstCreator =>
 
+  val tmpGen = FreshNameGenerator(i => s"<tmp-$i>")
+
   protected def astForExpression(node: RubyNode): Ast = node match
     case node: StaticLiteral            => astForStaticLiteral(node)
     case node: HereDocNode              => astForHereDoc(node)
@@ -26,6 +29,7 @@ trait AstForExpressionsCreator(implicit withSchemaValidation: ValidationMode) {
     case node: SimpleCall               => astForSimpleCall(node)
     case node: RequireCall              => astForRequireCall(node)
     case node: IncludeCall              => astForIncludeCall(node)
+    case node: YieldExpr                => astForYield(node)
     case node: RangeExpression          => astForRange(node)
     case node: ArrayLiteral             => astForArrayLiteral(node)
     case node: HashLiteral              => astForHashLiteral(node)
@@ -197,7 +201,7 @@ trait AstForExpressionsCreator(implicit withSchemaValidation: ValidationMode) {
     val block = blockNode(node)
     scope.pushNewScope(BlockScope(block))
 
-    val tmp = SimpleIdentifier(Option(className))(node.span.spanStart(freshVariableName))
+    val tmp = SimpleIdentifier(Option(className))(node.span.spanStart(tmpGen.fresh))
     def tmpIdentifier = {
       val tmpAst = astForSimpleIdentifier(tmp)
       tmpAst.root.collect { case x: NewIdentifier => x.typeFullName(receiverTypeFullName) }
@@ -365,6 +369,19 @@ trait AstForExpressionsCreator(implicit withSchemaValidation: ValidationMode) {
     astForSimpleCall(node.asSimpleCall)
   }
 
+  protected def astForYield(node: YieldExpr): Ast = {
+    scope.useProcParam match {
+      case Some(param) =>
+        astForMemberCall(
+          MemberCall(SimpleIdentifier()(node.span.spanStart(param)), ".", "call", node.arguments)(node.span)
+        )
+      case None =>
+        logger.warn(s"Yield expression outside of method scope: ${code(node)} ($relativeFileName), skipping")
+        astForUnknown(node)
+
+    }
+  }
+
   protected def astForRange(node: RangeExpression): Ast = {
     val lbAst = astForExpression(node.lowerBound)
     val ubAst = astForExpression(node.upperBound)
@@ -398,7 +415,7 @@ trait AstForExpressionsCreator(implicit withSchemaValidation: ValidationMode) {
   }
 
   protected def astForHashLiteral(node: HashLiteral): Ast = {
-    val tmp = freshVariableName
+    val tmp = tmpGen.fresh
 
     def tmpAst(tmpNode: Option[RubyNode] = None) = astForSimpleIdentifier(
       SimpleIdentifier()(tmpNode.map(_.span).getOrElse(node.span).spanStart(tmp))

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/AstForFunctionsCreator.scala

@@ -7,9 +7,12 @@ import io.joern.x2cpg.utils.NodeBuilders.{newClosureBindingNode, newLocalNode, n
 import io.joern.x2cpg.{Ast, AstEdge, ValidationMode, Defines as XDefines}
 import io.shiftleft.codepropertygraph.generated.nodes.*
 import io.shiftleft.codepropertygraph.generated.{EdgeTypes, EvaluationStrategies, ModifierTypes, NodeTypes}
+import io.joern.rubysrc2cpg.utils.FreshNameGenerator
 
 trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { this: AstCreator =>
 
+  val procParamGen = FreshNameGenerator(i => Left(s"<proc-param-$i>"))
+
   /** Creates method declaration related structures.
     * @param node
     *   the node to create the AST structure from.
@@ -40,7 +43,7 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
     )
 
     if (methodName == XDefines.ConstructorMethodName) scope.pushNewScope(ConstructorScope(fullName))
-    else scope.pushNewScope(MethodScope(fullName))
+    else scope.pushNewScope(MethodScope(fullName, procParamGen.fresh))
 
     val parameterAsts = astForParameters(node.parameters)
 
@@ -76,12 +79,18 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
       astForMethodBody(node.body, optionalStatementList)
     }
 
+    val anonProcParam = scope.anonProcParam.map { param =>
+      val paramNode = ProcParameter(param)(node.span.spanStart(s"&$param"))
+      val index = parameterAsts.lastOption.flatMap(_.root).map { case m: NewMethodParameterIn => m.index }.getOrElse(0)
+      astForParameter(paramNode, index)
+    }
+
     scope.popScope()
 
     val modifiers =
       ModifierTypes.VIRTUAL :: (if isClosure then ModifierTypes.LAMBDA :: Nil else Nil) map newModifierNode
 
-    methodAst(method, parameterAsts, stmtBlockAst, methodReturn, modifiers) :: refs
+    methodAst(method, parameterAsts ++ anonProcParam, stmtBlockAst, methodReturn, modifiers) :: refs
   }
 
   private def transformAsClosureBody(refs: List[Ast], baseStmtBlockAst: Ast) = {
@@ -141,6 +150,19 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
         )
         scope.addToScope(node.name, parameterIn)
         Ast(parameterIn)
+      case node: ProcParameter =>
+        val parameterIn = parameterInNode(
+          node = node,
+          name = node.name,
+          code = code(node),
+          index = index,
+          isVariadic = false,
+          evaluationStrategy = EvaluationStrategies.BY_REFERENCE,
+          typeFullName = None
+        )
+        scope.addToScope(node.name, parameterIn)
+        scope.setProcParam(node.name)
+        Ast(parameterIn)
       case node: CollectionParameter =>
         val typeFullName = node match {
           case ArrayParameter(_) => prefixAsBuiltin("Array")
@@ -252,7 +274,7 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
           astParentFullName = scope.surroundingScopeFullName
         )
 
-        scope.pushNewScope(MethodScope(fullName))
+        scope.pushNewScope(MethodScope(fullName, procParamGen.fresh))
 
         val thisParameterAst = Ast(
           newThisParameterNode(
@@ -268,8 +290,20 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
 
         val stmtBlockAst = astForMethodBody(node.body, optionalStatementList)
 
+        val anonProcParam = scope.anonProcParam.map { param =>
+          val paramNode = ProcParameter(param)(node.span.spanStart(s"&$param"))
+          val index =
+            parameterAsts.lastOption.flatMap(_.root).map { case m: NewMethodParameterIn => m.index }.getOrElse(1)
+          astForParameter(paramNode, index)
+        }
+
         scope.popScope()
-        methodAst(method, thisParameterAst +: parameterAsts, stmtBlockAst, methodReturnNode(node, Defines.Any))
+        methodAst(
+          method,
+          (thisParameterAst +: parameterAsts) ++ anonProcParam,
+          stmtBlockAst,
+          methodReturnNode(node, Defines.Any)
+        )
 
       case targetNode =>
         logger.warn(

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/AstForStatementsCreator.scala

@@ -160,7 +160,7 @@ trait AstForStatementsCreator(implicit withSchemaValidation: ValidationMode) { t
     }
     def generatedNode: StatementList = node.expression
       .map { e =>
-        val tmp = SimpleIdentifier(None)(e.span.spanStart(freshVariableName))
+        val tmp = SimpleIdentifier(None)(e.span.spanStart(tmpGen.fresh))
         StatementList(
           List(SingleAssignment(tmp, "=", e)(e.span)) ++
             goCase(Some(tmp))
@@ -252,7 +252,7 @@ trait AstForStatementsCreator(implicit withSchemaValidation: ValidationMode) { t
       case node: MemberCallWithBlock => returnAstForRubyCall(node)
       case node: SimpleCallWithBlock => returnAstForRubyCall(node)
       case _: (LiteralExpr | BinaryExpression | UnaryExpression | SimpleIdentifier | IndexAccess | Association |
-            RubyCall) =>
+            RubyCall | YieldExpr) =>
         astForReturnStatement(ReturnExpression(List(node))(node.span)) :: Nil
       case node: SingleAssignment =>
         astForSingleAssignment(node) :: List(astForReturnStatement(ReturnExpression(List(node.lhs))(node.span)))
@@ -265,6 +265,7 @@ trait AstForStatementsCreator(implicit withSchemaValidation: ValidationMode) { t
       case ret: ReturnExpression => astForReturnStatement(ret) :: Nil
       case node: MethodDeclaration =>
         (astForMethodDeclaration(node) :+ astForReturnMethodDeclarationSymbolName(node)).toList
+
       case node =>
         logger.warn(
           s"Implicit return here not supported yet: ${node.text} (${node.getClass.getSimpleName}), only generating statement"

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/AstForTypesCreator.scala

@@ -125,7 +125,7 @@ trait AstForTypesCreator(implicit withSchemaValidation: ValidationMode) { this:
       astParentType = scope.surroundingAstLabel,
       astParentFullName = scope.surroundingScopeFullName
     )
-    scope.pushNewScope(MethodScope(fullName))
+    scope.pushNewScope(MethodScope(fullName, procParamGen.fresh))
     val block_ = blockNode(node)
     scope.pushNewScope(BlockScope(block_))
     // TODO: Should it be `return this.@abc`?
@@ -155,7 +155,7 @@ trait AstForTypesCreator(implicit withSchemaValidation: ValidationMode) { this:
       astParentType = scope.surroundingAstLabel,
       astParentFullName = scope.surroundingScopeFullName
     )
-    scope.pushNewScope(MethodScope(fullName))
+    scope.pushNewScope(MethodScope(fullName, procParamGen.fresh))
     val parameter = parameterInNode(node, "x", "x", 1, false, EvaluationStrategies.BY_REFERENCE)
     val methodBody = {
       val block_ = blockNode(node)

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/astcreation/RubyIntermediateAst.scala

@@ -110,9 +110,7 @@ object RubyIntermediateAst {
 
   final case class HashParameter(name: String)(span: TextSpan) extends RubyNode(span) with CollectionParameter
 
-  final case class ProcParameter(target: RubyNode)(span: TextSpan) extends RubyNode(span) with MethodParameter {
-    def name: String = target.text
-  }
+  final case class ProcParameter(name: String)(span: TextSpan) extends RubyNode(span) with MethodParameter
 
   final case class SingleAssignment(lhs: RubyNode, op: String, rhs: RubyNode)(span: TextSpan) extends RubyNode(span)
 
@@ -303,6 +301,8 @@ object RubyIntermediateAst {
     */
   final case class ProcOrLambdaExpr(block: Block)(span: TextSpan) extends RubyNode(span)
 
+  final case class YieldExpr(arguments: List[RubyNode])(span: TextSpan) extends RubyNode(span)
+
   /** Represents a call with a block argument.
     */
   sealed trait RubyCallWithBlock[C <: RubyCall] extends RubyCall {

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/datastructures/RubyScope.scala

@@ -4,6 +4,7 @@ import better.files.File
 import io.joern.rubysrc2cpg.astcreation.GlobalTypes
 import io.joern.rubysrc2cpg.astcreation.GlobalTypes.builtinPrefix
 import io.joern.x2cpg.Defines
+import io.joern.rubysrc2cpg.passes.Defines as RDefines
 import io.joern.x2cpg.datastructures.*
 import io.shiftleft.codepropertygraph.generated.NodeTypes
 import io.shiftleft.codepropertygraph.generated.nodes.{DeclarationNew, NewLocal, NewMethodParameterIn}
@@ -117,6 +118,42 @@ class RubyScope(summary: RubyProgramSummary, projectRoot: Option[String])
     case ScopeElement(x: MethodLikeScope, _)    => x.fullName
   }
 
+  /** Locates a position in the stack matching a partial function, modifies it and emits a result
+    * @param pf
+    *   Tests ScopeElements of the stack. If they match, return the new value and the result to emi
+    * @return
+    *   the emitted result if the position was found and modifies
+    */
+  def updateSurrounding[T](
+    pf: PartialFunction[
+      ScopeElement[String, DeclarationNew, TypedScopeElement],
+      (ScopeElement[String, DeclarationNew, TypedScopeElement], T)
+    ]
+  ): Option[T] = {
+    stack.zipWithIndex
+      .collectFirst { case (pf(elem, res), i) =>
+        (elem, res, i)
+      }
+      .map { case (elem, res, i) =>
+        stack = stack.updated(i, elem)
+        res
+      }
+  }
+
+  def useProcParam: Option[String] = updateSurrounding {
+    case scope @ ScopeElement(MethodScope(fullName, param, _), variables) =>
+      (ScopeElement(MethodScope(fullName, param, true), variables), param.fold(x => x, x => x))
+  }
+
+  def anonProcParam: Option[String] = stack.collectFirst { case ScopeElement(MethodScope(_, Left(param), true), _) =>
+    param
+  }
+
+  def setProcParam(param: String): Unit = updateSurrounding {
+    case scope @ ScopeElement(MethodScope(fullName, _, _), variables) =>
+      (ScopeElement(MethodScope(fullName, Right(param)), variables), ())
+  }
+
   def surroundingTypeFullName: Option[String] = stack.collectFirst { case ScopeElement(x: TypeLikeScope, _) =>
     x.fullName
   }

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/datastructures/ScopeElement.scala

@@ -57,7 +57,8 @@ trait MethodLikeScope extends TypedScopeElement {
   def fullName: String
 }
 
-case class MethodScope(fullName: String) extends MethodLikeScope
+case class MethodScope(fullName: String, procParam: Either[String, String], hasYield: Boolean = false)
+    extends MethodLikeScope
 
 case class ConstructorScope(fullName: String) extends MethodLikeScope
 

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/parser/RubyNodeCreator.scala

@@ -2,26 +2,21 @@ package io.joern.rubysrc2cpg.parser
 
 import io.joern.rubysrc2cpg.astcreation.RubyIntermediateAst.*
 import io.joern.rubysrc2cpg.parser.AntlrContextHelpers.*
-import io.joern.rubysrc2cpg.parser.RubyParser.RangeOperatorContext
 import io.joern.rubysrc2cpg.passes.Defines
 import io.joern.rubysrc2cpg.passes.Defines.getBuiltInType
 import org.antlr.v4.runtime.tree.{ParseTree, RuleNode}
 import io.joern.x2cpg.Defines as XDefines;
 
 import scala.jdk.CollectionConverters.*
+import io.joern.rubysrc2cpg.utils.FreshNameGenerator
 
 /** Converts an ANTLR Ruby Parse Tree into the intermediate Ruby AST.
   */
 class RubyNodeCreator extends RubyParserBaseVisitor[RubyNode] {
 
-  private var classCounter: Int = 0
-
-  private def tmpClassTemplate(id: Int): String = s"<anon-class-$id>"
-
+  private val classNameGen = FreshNameGenerator(id => s"<anon-class-$id>")
   protected def freshClassName(span: TextSpan): SimpleIdentifier = {
-    val name = tmpClassTemplate(classCounter)
-    classCounter += 1
-    SimpleIdentifier(None)(span.spanStart(name))
+    SimpleIdentifier(None)(span.spanStart(classNameGen.fresh))
   }
 
   private def defaultTextSpan(code: String = ""): TextSpan = TextSpan(None, None, None, None, code)
@@ -523,6 +518,18 @@ class RubyNodeCreator extends RubyParserBaseVisitor[RubyNode] {
     }
   }
 
+  override def visitYieldExpression(ctx: RubyParser.YieldExpressionContext): RubyNode = {
+    val arguments = Option(ctx.argumentWithParentheses()).iterator.flatMap(_.arguments).map(visit).toList
+    YieldExpr(arguments)(ctx.toTextSpan)
+  }
+
+  override def visitYieldMethodInvocationWithoutParentheses(
+    ctx: RubyParser.YieldMethodInvocationWithoutParenthesesContext
+  ): RubyNode = {
+    val arguments = ctx.primaryValueList().primaryValue().asScala.map(visit).toList
+    YieldExpr(arguments)(ctx.toTextSpan)
+  }
+
   override def visitConstantIdentifierVariable(ctx: RubyParser.ConstantIdentifierVariableContext): RubyNode = {
     SimpleIdentifier()(ctx.toTextSpan)
   }
@@ -870,7 +877,9 @@ class RubyNodeCreator extends RubyParserBaseVisitor[RubyNode] {
   }
 
   override def visitProcParameter(ctx: RubyParser.ProcParameterContext): RubyNode = {
-    ProcParameter(visit(ctx.procParameterName()))(ctx.toTextSpan)
+    ProcParameter(
+      Option(ctx.procParameterName).map(_.LOCAL_VARIABLE_IDENTIFIER()).map(_.getText()).getOrElse(ctx.getText())
+    )(ctx.toTextSpan)
   }
 
   override def visitHashParameter(ctx: RubyParser.HashParameterContext): RubyNode = {

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/passes/Defines.scala

@@ -27,6 +27,8 @@ object Defines {
 
   val Resolver: String = "<dependency-resolver>"
 
+  val AnonymousProcParameter = "<anonymous-proc-param>"
+
   def getBuiltInType(typeInString: String) = s"${GlobalTypes.builtinPrefix}.$typeInString"
 
   object RubyOperators {

joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/utils/FreshNameGenerator.scala

@@ -0,0 +1,10 @@
+package io.joern.rubysrc2cpg.utils
+
+class FreshNameGenerator[T](template: Int => T) {
+  private var counter: Int = 0
+  def fresh: T = {
+    val name = template(counter)
+    counter += 1
+    name
+  }
+}