[ruby] Type/Method Identifier Ref Self Access (#4642)

This change moves the type/method identifier references for entities exportable from the script to prefix the respective entity at the definition.

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

@@ -87,7 +87,16 @@ trait AstCreatorHelper(implicit withSchemaValidation: ValidationMode) { this: As
     lineNumber: Option[Integer],
     columnNumber: Option[Integer]
   ): Ast = {
-    val code = Seq(lhs, rhs).collect { case x: AstNodeNew => x.code }.mkString(" = ")
+    astForAssignment(Ast(lhs), Ast(rhs), lineNumber, columnNumber)
+  }
+
+  protected def astForAssignment(
+    lhs: Ast,
+    rhs: Ast,
+    lineNumber: Option[Integer],
+    columnNumber: Option[Integer]
+  ): Ast = {
+    val code = Seq(lhs, rhs).flatMap(_.root).collect { case x: ExpressionNew => x.code }.mkString(" = ")
     val assignment = NewCall()
       .name(Operators.assignment)
       .methodFullName(Operators.assignment)
@@ -96,7 +105,7 @@ trait AstCreatorHelper(implicit withSchemaValidation: ValidationMode) { this: As
       .lineNumber(lineNumber)
       .columnNumber(columnNumber)
 
-    callAst(assignment, Seq(Ast(lhs), Ast(rhs)))
+    callAst(assignment, Seq(lhs, rhs))
   }
 
   protected val UnaryOperatorNames: Map[String, String] = Map(

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

@@ -12,7 +12,14 @@ import io.joern.x2cpg.utils.NodeBuilders.{
 }
 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.shiftleft.codepropertygraph.generated.{
+  DispatchTypes,
+  EdgeTypes,
+  EvaluationStrategies,
+  ModifierTypes,
+  NodeTypes,
+  Operators
+}
 import io.joern.rubysrc2cpg.utils.FreshNameGenerator
 
 import scala.collection.mutable
@@ -378,19 +385,30 @@ trait AstForFunctionsCreator(implicit withSchemaValidation: ValidationMode) { th
 
   private def createMethodRefPointer(method: NewMethod): Ast = {
     if (scope.isSurroundedByProgramScope) {
-      val methodRefNode = NewMethodRef()
-        .code(s"def ${method.name} (...)")
-        .methodFullName(method.fullName)
-        .typeFullName(method.fullName)
-        .lineNumber(method.lineNumber)
-        .columnNumber(method.columnNumber)
-
-      val methodRefIdent = NewIdentifier()
-        .code(method.name)
-        .name(method.name)
-        .typeFullName(method.fullName)
-        .lineNumber(method.lineNumber)
-        .columnNumber(method.columnNumber)
+      val methodRefNode = Ast(
+        NewMethodRef()
+          .code(s"def ${method.name} (...)")
+          .methodFullName(method.fullName)
+          .typeFullName(method.fullName)
+          .lineNumber(method.lineNumber)
+          .columnNumber(method.columnNumber)
+      )
+
+      val methodRefIdent = {
+        val self = NewIdentifier().name(Defines.Self).code(Defines.Self).typeFullName(Defines.Any)
+        val fi = NewFieldIdentifier()
+          .code(method.name)
+          .canonicalName(method.name)
+          .lineNumber(method.lineNumber)
+          .columnNumber(method.columnNumber)
+        val fieldAccess = NewCall()
+          .name(Operators.fieldAccess)
+          .code(s"${Defines.Self}.${method.name}")
+          .methodFullName(Operators.fieldAccess)
+          .dispatchType(DispatchTypes.STATIC_DISPATCH)
+          .typeFullName(Defines.Any)
+        callAst(fieldAccess, Seq(Ast(self), Ast(fi)))
+      }
 
       astForAssignment(methodRefIdent, methodRefNode, method.lineNumber, method.columnNumber)
     } else {

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

@@ -4,7 +4,13 @@ import io.joern.rubysrc2cpg.astcreation.RubyIntermediateAst.*
 import io.joern.rubysrc2cpg.datastructures.{BlockScope, MethodScope, ModuleScope, TypeScope}
 import io.joern.rubysrc2cpg.passes.Defines
 import io.joern.x2cpg.{Ast, ValidationMode, Defines as XDefines}
-import io.shiftleft.codepropertygraph.generated.nodes.{NewIdentifier, NewTypeDecl, NewTypeRef}
+import io.shiftleft.codepropertygraph.generated.nodes.{
+  NewCall,
+  NewFieldIdentifier,
+  NewIdentifier,
+  NewTypeDecl,
+  NewTypeRef
+}
 import io.shiftleft.codepropertygraph.generated.{DispatchTypes, EvaluationStrategies, Operators}
 
 import scala.collection.immutable.List
@@ -96,18 +102,29 @@ trait AstForTypesCreator(implicit withSchemaValidation: ValidationMode) { this:
 
   private def createTypeRefPointer(typeDecl: NewTypeDecl): Ast = {
     if (scope.isSurroundedByProgramScope) {
-      val typeRefName = typeDecl.name.split("[.]").takeRight(1).head
-      val typeRefNode = NewTypeRef()
-        .code(s"class ${typeDecl.fullName} (...)")
-        .typeFullName(typeDecl.fullName)
-
-      val typeRefIdent = NewIdentifier()
-        .code(typeRefName)
-        .name(typeRefName)
-        .typeFullName(typeDecl.fullName)
-        .lineNumber(typeDecl.lineNumber)
-        .columnNumber(typeDecl.columnNumber)
+      val typeRefNode = Ast(
+        NewTypeRef()
+          .code(s"class ${typeDecl.name} (...)")
+          .typeFullName(typeDecl.fullName)
+          .lineNumber(typeDecl.lineNumber)
+          .columnNumber(typeDecl.columnNumber)
+      )
 
+      val typeRefIdent = {
+        val self = NewIdentifier().name(Defines.Self).code(Defines.Self).typeFullName(Defines.Any)
+        val fi = NewFieldIdentifier()
+          .code(typeDecl.name)
+          .canonicalName(typeDecl.name)
+          .lineNumber(typeDecl.lineNumber)
+          .columnNumber(typeDecl.columnNumber)
+        val fieldAccess = NewCall()
+          .name(Operators.fieldAccess)
+          .code(s"${Defines.Self}.${typeDecl.name}")
+          .methodFullName(Operators.fieldAccess)
+          .dispatchType(DispatchTypes.STATIC_DISPATCH)
+          .typeFullName(Defines.Any)
+        callAst(fieldAccess, Seq(Ast(self), Ast(fi)))
+      }
       astForAssignment(typeRefIdent, typeRefNode, typeDecl.lineNumber, typeDecl.columnNumber)
     } else {
       Ast()

joern-cli/frontends/rubysrc2cpg/src/test/scala/io/joern/rubysrc2cpg/dataflow/MethodReturnTests.scala

@@ -81,8 +81,12 @@ class MethodReturnTests extends RubyCode2CpgFixture(withPostProcessing = true, w
           ("arg > 1", 3),
           ("arg + 1", 4),
           ("RET", 2),
+          ("self.foo = def foo (...)", 2),
+          ("self.foo = def foo (...)", -1),
+          ("foo x", 11),
+          ("foo(self, arg)", 2),
+          ("RET", 2),
           ("foo x", 11),
-          ("y = foo x", 11),
           ("puts y", 12)
         )
       )
@@ -122,8 +126,12 @@ class MethodReturnTests extends RubyCode2CpgFixture(withPostProcessing = true, w
           ("RET", 2),
           ("add(arg)", 8),
           ("RET", 6),
+          ("self.foo = def foo (...)", 6),
+          ("self.foo = def foo (...)", -1),
+          ("foo x", 15),
+          ("foo(self, arg)", 6),
+          ("RET", 6),
           ("foo x", 15),
-          ("y = foo x", 15),
           ("puts y", 16)
         )
       )
@@ -151,8 +159,12 @@ class MethodReturnTests extends RubyCode2CpgFixture(withPostProcessing = true, w
           ("q = p", 3),
           ("return q", 4),
           ("RET", 2),
+          ("self.add = def add (...)", 2),
+          ("self.add = def add (...)", -1),
+          ("add(n)", 8),
+          ("add(self, p)", 2),
+          ("RET", 2),
           ("add(n)", 8),
-          ("ret = add(n)", 8),
           ("puts ret", 9)
         )
       )

joern-cli/frontends/rubysrc2cpg/src/test/scala/io/joern/rubysrc2cpg/querying/MethodTests.scala

@@ -592,27 +592,30 @@ class MethodTests extends RubyCode2CpgFixture {
     "be directly under :program" in {
       inside(cpg.method.name(RDefines.Program).filename("t1.rb").assignment.l) {
         case moduleAssignment :: classAssignment :: methodAssignment :: Nil =>
-          moduleAssignment.code shouldBe "A = class t1.rb:<global>::program.A (...)"
-          classAssignment.code shouldBe "B = class t1.rb:<global>::program.B (...)"
-          methodAssignment.code shouldBe "c = def c (...)"
+          moduleAssignment.code shouldBe "self.A = class A (...)"
+          classAssignment.code shouldBe "self.B = class B (...)"
+          methodAssignment.code shouldBe "self.c = def c (...)"
 
           inside(moduleAssignment.argument.l) {
-            case (lhs: Identifier) :: (rhs: TypeRef) :: Nil =>
-              lhs.name shouldBe "A"
+            case (lhs: Call) :: (rhs: TypeRef) :: Nil =>
+              lhs.code shouldBe "self.A"
+              lhs.name shouldBe Operators.fieldAccess
               rhs.typeFullName shouldBe "t1.rb:<global>::program.A"
             case xs => fail(s"Expected lhs and rhs, instead got ${xs.code.mkString(",")}")
           }
 
           inside(classAssignment.argument.l) {
-            case (lhs: Identifier) :: (rhs: TypeRef) :: Nil =>
-              lhs.name shouldBe "B"
+            case (lhs: Call) :: (rhs: TypeRef) :: Nil =>
+              lhs.code shouldBe "self.B"
+              lhs.name shouldBe Operators.fieldAccess
               rhs.typeFullName shouldBe "t1.rb:<global>::program.B"
             case xs => fail(s"Expected lhs and rhs, instead got ${xs.code.mkString(",")}")
           }
 
           inside(methodAssignment.argument.l) {
-            case (lhs: Identifier) :: (rhs: MethodRef) :: Nil =>
-              lhs.name shouldBe "c"
+            case (lhs: Call) :: (rhs: MethodRef) :: Nil =>
+              lhs.code shouldBe "self.c"
+              lhs.name shouldBe Operators.fieldAccess
               rhs.methodFullName shouldBe "t1.rb:<global>::program:c"
               rhs.typeFullName shouldBe "t1.rb:<global>::program:c"
             case xs => fail(s"Expected lhs and rhs, instead got ${xs.code.mkString(",")}")
@@ -625,19 +628,21 @@ class MethodTests extends RubyCode2CpgFixture {
     "not be present in other files" in {
       inside(cpg.method.name(RDefines.Program).filename("t2.rb").assignment.l) {
         case classAssignment :: methodAssignment :: Nil =>
-          classAssignment.code shouldBe "D = class t2.rb:<global>::program.D (...)"
-          methodAssignment.code shouldBe "e = def e (...)"
+          classAssignment.code shouldBe "self.D = class D (...)"
+          methodAssignment.code shouldBe "self.e = def e (...)"
 
           inside(classAssignment.argument.l) {
-            case (lhs: Identifier) :: (rhs: TypeRef) :: Nil =>
-              lhs.name shouldBe "D"
+            case (lhs: Call) :: (rhs: TypeRef) :: Nil =>
+              lhs.code shouldBe "self.D"
+              lhs.name shouldBe Operators.fieldAccess
               rhs.typeFullName shouldBe "t2.rb:<global>::program.D"
             case xs => fail(s"Expected lhs and rhs, instead got ${xs.code.mkString(",")}")
           }
 
           inside(methodAssignment.argument.l) {
-            case (lhs: Identifier) :: (rhs: MethodRef) :: Nil =>
-              lhs.name shouldBe "e"
+            case (lhs: Call) :: (rhs: MethodRef) :: Nil =>
+              lhs.code shouldBe "self.e"
+              lhs.name shouldBe Operators.fieldAccess
               rhs.methodFullName shouldBe "t2.rb:<global>::program:e"
               rhs.typeFullName shouldBe "t2.rb:<global>::program:e"
             case xs => fail(s"Expected lhs and rhs, instead got ${xs.code.mkString(",")}")
@@ -650,9 +655,9 @@ class MethodTests extends RubyCode2CpgFixture {
     "be placed directly before each entity's definition" in {
       inside(cpg.method.name(RDefines.Program).filename("t1.rb").block.astChildren.l) {
         case (a1: Call) :: (_: TypeDecl) :: (a2: Call) :: (_: TypeDecl) :: (a3: Call) :: (_: Method) :: (_: TypeDecl) :: Nil =>
-          a1.code shouldBe "A = class t1.rb:<global>::program.A (...)"
-          a2.code shouldBe "B = class t1.rb:<global>::program.B (...)"
-          a3.code shouldBe "c = def c (...)"
+          a1.code shouldBe "self.A = class A (...)"
+          a2.code shouldBe "self.B = class B (...)"
+          a3.code shouldBe "self.c = def c (...)"
         case xs => fail(s"Expected assignments to appear before definitions, instead got [$xs]")
       }
     }