[ruby] Move Post-Processing to x2cpg (#4851)

Following the other frontends, migrates the non-deprecated post-processing passes to `x2cpg`
joernio · Aug 15, 2024 · e967d4c · e967d4c
1 parent ca4c172
commit e967d4c
Show file tree

Hide file tree

Showing 7 changed files with 119 additions and 23 deletions.
diff --git a/joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/RubySrc2Cpg.scala b/joern-cli/frontends/rubysrc2cpg/src/main/scala/io/joern/rubysrc2cpg/RubySrc2Cpg.scala
@@ -11,14 +11,17 @@ import io.joern.rubysrc2cpg.passes.{
   AstCreationPass,
   ConfigFileCreationPass,
   DependencyPass,
-  DependencySummarySolverPass,
+  DependencySummarySolverPass
+}
+import io.joern.rubysrc2cpg.utils.DependencyDownloader
+import io.joern.x2cpg.X2Cpg.withNewEmptyCpg
+import io.joern.x2cpg.frontendspecific.rubysrc2cpg.{
   ImplicitRequirePass,
   ImportsPass,
   RubyImportResolverPass,
-  RubyTypeHintCallLinker
+  RubyTypeHintCallLinker,
+  RubyTypeRecoveryPassGenerator
 }
-import io.joern.rubysrc2cpg.utils.DependencyDownloader
-import io.joern.x2cpg.X2Cpg.withNewEmptyCpg
 import io.joern.x2cpg.passes.base.AstLinkerPass
 import io.joern.x2cpg.passes.callgraph.NaiveCallLinker
 import io.joern.x2cpg.passes.frontend.{MetaDataPass, TypeNodePass, XTypeRecoveryConfig}
@@ -175,7 +178,7 @@ object RubySrc2Cpg {
     } else {
       val implicitRequirePass = if (cpg.dependency.name.contains("zeitwerk")) ImplicitRequirePass(cpg) :: Nil else Nil
       implicitRequirePass ++ List(ImportsPass(cpg), RubyImportResolverPass(cpg)) ++
-        new passes.RubyTypeRecoveryPassGenerator(cpg, config = XTypeRecoveryConfig(iterations = 4))
+        new RubyTypeRecoveryPassGenerator(cpg, config = XTypeRecoveryConfig(iterations = 4))
           .generate() ++ List(new RubyTypeHintCallLinker(cpg), new NaiveCallLinker(cpg), new AstLinkerPass(cpg))
     }
   }

diff --git a/...rontends/x2cpg/src/main/scala/io/joern/x2cpg/frontendspecific/rubysrc2cpg/Constants.scala b/...rontends/x2cpg/src/main/scala/io/joern/x2cpg/frontendspecific/rubysrc2cpg/Constants.scala
@@ -0,0 +1,93 @@
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
+
+object Constants {
+
+  val builtinPrefix = "__core"
+  val kernelPrefix  = s"$builtinPrefix.Kernel"
+  val Initialize    = "initialize"
+  val Main          = "<main>"
+
+  /* Source: https://ruby-doc.org/3.2.2/Kernel.html
+   *
+   * We comment-out methods that require an explicit "receiver" (target of member access.)
+   */
+  val kernelFunctions: Set[String] = Set(
+    "Array",
+    "Complex",
+    "Float",
+    "Hash",
+    "Integer",
+    "Rational",
+    "String",
+    "__callee__",
+    "__dir__",
+    "__method__",
+    "abort",
+    "at_exit",
+    "autoload",
+    "autoload?",
+    "binding",
+    "block_given?",
+    "callcc",
+    "caller",
+    "caller_locations",
+    "catch",
+    "chomp",
+    "chomp!",
+    "chop",
+    "chop!",
+    // "class",
+    // "clone",
+    "eval",
+    "exec",
+    "exit",
+    "exit!",
+    "fail",
+    "fork",
+    "format",
+    // "frozen?",
+    "gets",
+    "global_variables",
+    "gsub",
+    "gsub!",
+    "iterator?",
+    "lambda",
+    "load",
+    "local_variables",
+    "loop",
+    "open",
+    "p",
+    "print",
+    "printf",
+    "proc",
+    "putc",
+    "puts",
+    "raise",
+    "rand",
+    "readline",
+    "readlines",
+    "require",
+    "require_all",
+    "require_relative",
+    "select",
+    "set_trace_func",
+    "sleep",
+    "spawn",
+    "sprintf",
+    "srand",
+    "sub",
+    "sub!",
+    "syscall",
+    "system",
+    "tap",
+    "test",
+    // "then",
+    "throw",
+    "trace_var",
+    // "trap",
+    "untrace_var",
+    "warn"
+    // "yield_self",
+  )
+
+}
diff --git a/...ysrc2cpg/passes/ImplicitRequirePass.scala → ...fic/rubysrc2cpg/ImplicitRequirePass.scala b/...ysrc2cpg/passes/ImplicitRequirePass.scala → ...fic/rubysrc2cpg/ImplicitRequirePass.scala
@@ -1,6 +1,7 @@
-package io.joern.rubysrc2cpg.passes
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
 
-import io.joern.rubysrc2cpg.passes.GlobalTypes.{builtinPrefix, kernelPrefix}
+import io.joern.x2cpg.Defines
+import io.joern.x2cpg.frontendspecific.rubysrc2cpg.Constants.*
 import io.shiftleft.codepropertygraph.generated.nodes.*
 import io.shiftleft.codepropertygraph.generated.{Cpg, DispatchTypes, EdgeTypes, Operators}
 import io.shiftleft.passes.ForkJoinParallelCpgPass

diff --git a/...oern/rubysrc2cpg/passes/ImportsPass.scala → ...endspecific/rubysrc2cpg/ImportsPass.scala b/...oern/rubysrc2cpg/passes/ImportsPass.scala → ...endspecific/rubysrc2cpg/ImportsPass.scala
@@ -1,4 +1,4 @@
-package io.joern.rubysrc2cpg.passes
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
 
 import io.joern.x2cpg.Imports.createImportNodeAndLink
 import io.joern.x2cpg.X2Cpg.stripQuotes

diff --git a/...c2cpg/passes/RubyImportResolverPass.scala → .../rubysrc2cpg/RubyImportResolverPass.scala b/...c2cpg/passes/RubyImportResolverPass.scala → .../rubysrc2cpg/RubyImportResolverPass.scala
@@ -1,14 +1,12 @@
-package io.joern.rubysrc2cpg.passes
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
 
 import better.files.File
-import io.joern.rubysrc2cpg.deprecated.utils.PackageTable
-import io.joern.x2cpg.Defines as XDefines
-import io.shiftleft.semanticcpg.language.importresolver.*
+import io.joern.x2cpg.frontendspecific.rubysrc2cpg.Constants.*
 import io.joern.x2cpg.passes.frontend.XImportResolverPass
 import io.shiftleft.codepropertygraph.generated.Cpg
 import io.shiftleft.codepropertygraph.generated.nodes.Call
 import io.shiftleft.semanticcpg.language.*
-import io.joern.rubysrc2cpg.passes.Defines as RDefines
+import io.shiftleft.semanticcpg.language.importresolver.*
 import io.shiftleft.semanticcpg.language.types.structure.NamespaceTraversal
 
 import java.io.File as JFile
@@ -47,7 +45,7 @@ class RubyImportResolverPass(cpg: Cpg) extends XImportResolverPass(cpg) {
         .flatMap(fullName =>
           Seq(
             ResolvedTypeDecl(fullName),
-            ResolvedMethod(s"$fullName.${Defines.Initialize}", "new", fullName.split("[.]").lastOption)
+            ResolvedMethod(s"$fullName.${Initialize}", "new", fullName.split("[.]").lastOption)
           )
         )
         .toSet
@@ -61,7 +59,7 @@ class RubyImportResolverPass(cpg: Cpg) extends XImportResolverPass(cpg) {
       // Expose methods which are directly present in a file, without any module, TypeDecl
       val resolvedMethods = cpg.method
         .where(_.file.name(filePattern))
-        .where(_.nameExact(RDefines.Main))
+        .where(_.nameExact(Main))
         .astChildren
         .astChildren
         .isMethod

diff --git a/...c2cpg/passes/RubyTypeHintCallLinker.scala → .../rubysrc2cpg/RubyTypeHintCallLinker.scala b/...c2cpg/passes/RubyTypeHintCallLinker.scala → .../rubysrc2cpg/RubyTypeHintCallLinker.scala
@@ -1,6 +1,7 @@
-package io.joern.rubysrc2cpg.passes
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
 
 import io.joern.x2cpg.passes.frontend.XTypeHintCallLinker
+import io.joern.x2cpg.frontendspecific.rubysrc2cpg.Constants.*
 import io.shiftleft.codepropertygraph.generated.Cpg
 import io.shiftleft.codepropertygraph.generated.nodes.{Call, NewMethod}
 import io.shiftleft.semanticcpg.language.*
@@ -26,8 +27,8 @@ class RubyTypeHintCallLinker(cpg: Cpg) extends XTypeHintCallLinker(cpg) {
     }
     val name =
       if (methodName.contains(pathSep) && methodName.length > methodName.lastIndexOf(pathSep) + 1)
-        val strippedMethod = methodName.stripPrefix(s"${GlobalTypes.kernelPrefix}.")
-        if GlobalTypes.kernelFunctions.contains(strippedMethod) then strippedMethod
+        val strippedMethod = methodName.stripPrefix(s"$kernelPrefix.")
+        if kernelFunctions.contains(strippedMethod) then strippedMethod
         else methodName.substring(methodName.lastIndexOf(pathSep) + pathSep.length)
       else methodName
     createMethodStub(name, methodName, call.argumentOut.size, isExternal, builder)

diff --git a/...asses/RubyTypeRecoveryPassGenerator.scala → ...c2cpg/RubyTypeRecoveryPassGenerator.scala b/...asses/RubyTypeRecoveryPassGenerator.scala → ...c2cpg/RubyTypeRecoveryPassGenerator.scala
@@ -1,13 +1,13 @@
-package io.joern.rubysrc2cpg.passes
+package io.joern.x2cpg.frontendspecific.rubysrc2cpg
 
 import io.joern.x2cpg.Defines as XDefines
+import io.joern.x2cpg.frontendspecific.rubysrc2cpg.Constants.*
 import io.joern.x2cpg.passes.frontend.*
 import io.joern.x2cpg.passes.frontend.XTypeRecovery.AllNodeTypesFromNodeExt
-import io.shiftleft.codepropertygraph.generated.{Cpg, Operators, PropertyNames}
 import io.shiftleft.codepropertygraph.generated.nodes.*
-import io.shiftleft.semanticcpg.language.operatorextension.OpNodes.FieldAccess
+import io.shiftleft.codepropertygraph.generated.{Cpg, DiffGraphBuilder, Operators, PropertyNames}
 import io.shiftleft.semanticcpg.language.*
-import io.shiftleft.codepropertygraph.generated.DiffGraphBuilder
+import io.shiftleft.semanticcpg.language.operatorextension.OpNodes.FieldAccess
 
 class RubyTypeRecoveryPassGenerator(cpg: Cpg, config: XTypeRecoveryConfig = XTypeRecoveryConfig())
     extends XTypeRecoveryPassGenerator[File](cpg, config) {
@@ -40,7 +40,7 @@ private class RecoverForRubyFile(cpg: Cpg, cu: File, builder: DiffGraphBuilder,
   /** A heuristic method to determine if a call name is a constructor or not.
     */
   override protected def isConstructor(name: String): Boolean =
-    !name.isBlank && (name == "new" || name == Defines.Initialize)
+    !name.isBlank && (name == "new" || name == Initialize)
 
   override protected def hasTypes(node: AstNode): Boolean = node match {
     case x: Call if !x.methodFullName.startsWith("<operator>") =>