Use full name syntax for methods declared in header files (#36)

* Use full name syntax for methods declared in header files Signed-off-by: Prabhu Subramanian <prabhu@appthreat.com> * Added c/c++ support in cdxpass Signed-off-by: Prabhu Subramanian <prabhu@appthreat.com> * Tag library calls Signed-off-by: Prabhu Subramanian <prabhu@appthreat.com> --------- Signed-off-by: Prabhu Subramanian <prabhu@appthreat.com>
AppThreat · Nov 28, 2023 · 037b9da · 037b9da
1 parent dd5612e
commit 037b9da
Show file tree

Hide file tree

Showing 12 changed files with 104 additions and 28 deletions.
diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@ Code Hierarchy Exploration Net (chen) is an advanced exploration toolkit for you
 
 ## Requirements
 
-- Java 17 - 20
+- Java 17 - 21
 - Python > 3.8.1
 - Node.js > 16 (To run [atom](https://github.com/AppThreat/atom))
 - Minimum 16GB RAM

diff --git a/build.sbt b/build.sbt
@@ -1,6 +1,6 @@
 name                     := "chen"
 ThisBuild / organization := "io.appthreat"
-ThisBuild / version      := "1.0.2"
+ThisBuild / version      := "1.0.3"
 ThisBuild / scalaVersion := "3.3.1"
 
 val cpgVersion = "1.4.22"

diff --git a/ci/Dockerfile b/ci/Dockerfile
@@ -4,7 +4,7 @@ LABEL maintainer="appthreat" \
       org.opencontainers.image.authors="Team AppThreat <cloud@appthreat.com>" \
       org.opencontainers.image.source="https://github.com/appthreat/chen" \
       org.opencontainers.image.url="https://github.com/appthreat/chen" \
-      org.opencontainers.image.version="1.0.0" \
+      org.opencontainers.image.version="1.0.x" \
       org.opencontainers.image.vendor="appthreat" \
       org.opencontainers.image.licenses="Apache-2.0" \
       org.opencontainers.image.title="chen" \

diff --git a/codemeta.json b/codemeta.json
@@ -7,7 +7,7 @@
   "downloadUrl": "https://github.com/AppThreat/chen",
   "issueTracker": "https://github.com/AppThreat/chen/issues",
   "name": "chen",
-  "version": "1.0.2",
+  "version": "1.0.3",
   "description": "Code Hierarchy Exploration Net (chen) is an advanced exploration toolkit for your application source code and its dependency hierarchy.",
   "applicationCategory": "code-analysis",
   "keywords": [

diff --git a/console/src/main/scala/io/appthreat/console/Console.scala b/console/src/main/scala/io/appthreat/console/Console.scala
@@ -502,18 +502,22 @@ class Console[T <: Project](
             atom.file.whereNot(_.name("<unknown>")).foreach { f =>
                 val childTree = richTreeLib.Tree(f.name, highlight = true)
                 f.method.foreach(m =>
-                    val mtree = childTree.add(m.fullName)
+                    val addedMethods = mutable.Map.empty[String, Boolean]
+                    val mtree        = childTree.add(m.fullName)
                     if includeCalls then
                         m.call
                             .filterNot(_.name.startsWith("<operator"))
+                            .filterNot(_.methodFullName == "NULL")
                             .toSet
                             .foreach(c =>
-                                mtree
-                                    .add(
-                                      c.methodFullName + (if c.callee(NoResolve).head.isExternal
-                                                          then " :right_arrow_curving_up:"
-                                                          else "")
-                                    )
+                                if !addedMethods.contains(c.methodFullName) then
+                                    mtree
+                                        .add(
+                                          c.methodFullName + (if c.callee(NoResolve).head.isExternal
+                                                              then " :right_arrow_curving_up:"
+                                                              else "")
+                                        )
+                                    addedMethods += c.methodFullName -> true
                             )
                 )
                 rootTree.add(childTree)

diff --git a/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/C2Cpg.scala b/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/C2Cpg.scala
@@ -1,14 +1,18 @@
 package io.appthreat.c2cpg
 
 import io.appthreat.c2cpg.datastructures.CGlobal
-import io.appthreat.c2cpg.passes.{AstCreationPass, PreprocessorPass, TypeDeclNodePass}
+import io.appthreat.c2cpg.passes.{
+    AstCreationPass,
+    ConfigFileCreationPass,
+    PreprocessorPass,
+    TypeDeclNodePass
+}
 import io.appthreat.c2cpg.utils.Report
-import io.appthreat.c2cpg.passes.{AstCreationPass, PreprocessorPass, TypeDeclNodePass}
-import io.shiftleft.codepropertygraph.Cpg
-import io.shiftleft.codepropertygraph.generated.Languages
-import io.appthreat.x2cpg.passes.frontend.{MetaDataPass, TypeNodePass}
 import io.appthreat.x2cpg.X2Cpg.withNewEmptyCpg
 import io.appthreat.x2cpg.X2CpgFrontend
+import io.appthreat.x2cpg.passes.frontend.{MetaDataPass, TypeNodePass}
+import io.shiftleft.codepropertygraph.Cpg
+import io.shiftleft.codepropertygraph.generated.Languages
 
 import scala.util.Try
 
@@ -20,6 +24,7 @@ class C2Cpg extends X2CpgFrontend[Config]:
         withNewEmptyCpg(config.outputPath, config) { (cpg, config) =>
             new MetaDataPass(cpg, Languages.NEWC, config.inputPath).createAndApply()
             new AstCreationPass(cpg, config, report).createAndApply()
+            new ConfigFileCreationPass(cpg).createAndApply()
             TypeNodePass.withRegisteredTypes(CGlobal.typesSeen(), cpg).createAndApply()
             new TypeDeclNodePass(cpg)(config.schemaValidation).createAndApply()
             report.print()

diff --git a/...form/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/astcreation/AstCreatorHelper.scala b/...form/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/astcreation/AstCreatorHelper.scala
@@ -1,6 +1,7 @@
 package io.appthreat.c2cpg.astcreation
 
 import io.appthreat.c2cpg.datastructures.CGlobal
+import io.appthreat.c2cpg.parser.FileDefaults
 import io.shiftleft.codepropertygraph.generated.nodes.{ExpressionNew, NewNode}
 import io.shiftleft.codepropertygraph.generated.{DispatchTypes, Operators}
 import io.appthreat.x2cpg.{Ast, SourceFiles, ValidationMode}
@@ -280,6 +281,9 @@ trait AstCreatorHelper(implicit withSchemaValidation: ValidationMode):
         cleanedName.split(Defines.qualifiedNameSeparator).lastOption.getOrElse(cleanedName)
 
     protected def fullName(node: IASTNode): String =
+        val filename           = fileName(node)
+        val lineNo: Integer    = line(node).getOrElse(-1)
+        val lineNoEnd: Integer = lineEnd(node).getOrElse(-1)
         val qualifiedName: String = node match
             case d: CPPASTIdExpression if d.getEvaluation.isInstanceOf[EvalBinding] =>
                 val evaluation = d.getEvaluation.asInstanceOf[EvalBinding]
@@ -324,9 +328,21 @@ trait AstCreatorHelper(implicit withSchemaValidation: ValidationMode):
                 if ASTStringUtil.getSimpleName(
                   f.getName
                 ).isEmpty && f.getNestedDeclarator != null =>
-                s"${fullName(f.getParent)}.${shortName(f.getNestedDeclarator)}"
+                val parentFullName = fullName(f.getParent)
+                val sn             = shortName(f.getNestedDeclarator)
+                val fnWithParent =
+                    if parentFullName.nonEmpty then s"${parentFullName}.${sn}" else sn
+                if FileDefaults.isHeaderFile(filename) then
+                    s"$filename:$lineNo:$lineNoEnd:${fnWithParent}"
+                else fnWithParent
             case f: IASTFunctionDeclarator =>
-                s"${fullName(f.getParent)}.${ASTStringUtil.getSimpleName(f.getName)}"
+                val parentFullName = fullName(f.getParent)
+                val sn             = ASTStringUtil.getSimpleName(f.getName)
+                val fnWithParent =
+                    if parentFullName.nonEmpty then s"${parentFullName}.${sn}" else sn
+                if FileDefaults.isHeaderFile(filename) then
+                    s"$filename:$lineNo:$lineNoEnd:${fnWithParent}"
+                else fnWithParent
             case f: IASTFunctionDefinition if f.getDeclarator != null =>
                 s"${fullName(f.getParent)}.${ASTStringUtil.getQualifiedName(f.getDeclarator.getName)}"
             case f: IASTFunctionDefinition =>

diff --git a/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/astcreation/MacroHandler.scala b/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/astcreation/MacroHandler.scala
@@ -158,7 +158,7 @@ trait MacroHandler(implicit withSchemaValidation: ValidationMode):
         val filename           = fileName(macroDef)
         val lineNo: Integer    = line(macroDef).getOrElse(-1)
         val lineNoEnd: Integer = lineEnd(macroDef).getOrElse(-1)
-        s"$filename:$lineNo:$lineNoEnd:$name:${argAsts.size}"
+        if name != "NULL" then s"$filename:$lineNo:$lineNoEnd:$name:${argAsts.size}" else name
 
     /** The CDT utility method is unfortunately in a class that is marked as deprecated, however,
       * this is because the CDT team would like to discourage its use but at the same time does not

diff --git a/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/parser/FileDefaults.scala b/platform/frontends/c2cpg/src/main/scala/io/appthreat/c2cpg/parser/FileDefaults.scala
@@ -12,7 +12,8 @@ object FileDefaults:
 
     val SOURCE_FILE_EXTENSIONS: Set[String] = Set(C_EXT, CC_EXT, CPP_EXT)
 
-    val HEADER_FILE_EXTENSIONS: Set[String] = Set(C_HEADER_EXT, CPP_HEADER_EXT, OTHER_HEADER_EXT)
+    val HEADER_FILE_EXTENSIONS: Set[String] =
+        Set(C_HEADER_EXT, CPP_HEADER_EXT, OTHER_HEADER_EXT, ".h.in")
 
     private val CPP_FILE_EXTENSIONS = Set(CC_EXT, CPP_EXT, CPP_HEADER_EXT, ".ccm", ".cxxm", ".c++m")
 

diff --git a/...tends/c2cpg/src/test/scala/io/appthreat/c2cpg/passes/ast/HeaderAstCreationPassTests.scala b/...tends/c2cpg/src/test/scala/io/appthreat/c2cpg/passes/ast/HeaderAstCreationPassTests.scala
@@ -42,15 +42,16 @@ class HeaderAstCreationPassTests extends CCodeToCpgSuite {
           // in main.h and included in main.c and we do scan both
           bar.fullName shouldBe "bar"
           bar.filename shouldBe "main.h"
-          foo.fullName shouldBe "foo"
-          foo.filename shouldBe "other.h"
+
+          foo.fullName shouldBe "main"
+          foo.filename shouldBe "main.c"
           // main is include twice. First time for the header file,
           // second time for the actual implementation in the source file
           // We do not de-duplicate this as line/column numbers differ
-          m1.fullName shouldBe "main"
+          m1.fullName shouldBe "main.h:2:2:main"
           m1.filename shouldBe "main.h"
-          m2.fullName shouldBe "main"
-          m2.filename shouldBe "main.c"
+          m2.fullName shouldBe "other.h:2:2:foo"
+          m2.filename shouldBe "other.h"
           printf.fullName shouldBe "printf"
       }
     }

diff --git a/platform/frontends/x2cpg/src/main/scala/io/appthreat/x2cpg/passes/taggers/CdxPass.scala b/platform/frontends/x2cpg/src/main/scala/io/appthreat/x2cpg/passes/taggers/CdxPass.scala
@@ -44,7 +44,8 @@ class CdxPass(atom: Cpg) extends CpgPass(atom):
 
     private def PY_REQUEST_PATTERNS = Array(".*views.py:<module>.*")
 
-    private def containsRegex(str: String) = Pattern.quote(str) == str || str.contains("*")
+    private def containsRegex(str: String) =
+        Pattern.quote(str) == str || str.contains("*") || str.contains("(") || str.contains(")")
 
     private val BOM_JSON_FILE = ".*(bom|cdx).json"
 
@@ -121,7 +122,54 @@ class CdxPass(atom: Cpg) extends CpgPass(atom):
                                 then bpkg = toPyModuleForm(bpkg)
                                 if bpkg.nonEmpty && !donePkgs.contains(bpkg) then
                                     donePkgs.put(bpkg, true)
-                                    if !containsRegex(bpkg) then
+                                    // C/C++
+                                    if language == Languages.NEWC || language == Languages.C
+                                    then
+                                        atom.method.fullNameExact(bpkg).callIn(
+                                          NoResolve
+                                        ).newTagNode(
+                                          compPurl
+                                        ).store()(dstGraph)
+                                        atom.method.fullNameExact(bpkg).callIn(
+                                          NoResolve
+                                        ).newTagNode(
+                                          "library-call"
+                                        ).store()(dstGraph)
+                                        atom.method.fullNameExact(bpkg).newTagNode(
+                                          compPurl
+                                        ).store()(dstGraph)
+                                        if !containsRegex(bpkg) then
+                                            atom.parameter.typeFullName(s"$bpkg.*").newTagNode(
+                                              compPurl
+                                            ).store()(dstGraph)
+                                            atom.parameter.typeFullName(s"$bpkg.*").newTagNode(
+                                              "framework-input"
+                                            ).store()(dstGraph)
+                                            atom.parameter.typeFullName(s"$bpkg.*").method.callIn(
+                                              NoResolve
+                                            ).newTagNode(
+                                              compPurl
+                                            ).store()(dstGraph)
+                                        else
+                                            atom.parameter.typeFullName(
+                                              s"${Pattern.quote(bpkg)}.*"
+                                            ).newTagNode(
+                                              compPurl
+                                            ).store()(dstGraph)
+                                            atom.parameter.typeFullName(
+                                              s"${Pattern.quote(bpkg)}.*"
+                                            ).newTagNode(
+                                              "framework-input"
+                                            ).store()(dstGraph)
+                                            atom.parameter.typeFullName(
+                                              s"${Pattern.quote(bpkg)}.*"
+                                            ).method.callIn(
+                                              NoResolve
+                                            ).newTagNode(
+                                              compPurl
+                                            ).store()(dstGraph)
+                                        end if
+                                    else if !containsRegex(bpkg) then
                                         atom.call.typeFullNameExact(bpkg).newTagNode(
                                           compPurl
                                         ).store()(dstGraph)
@@ -246,4 +294,5 @@ class CdxPass(atom: Cpg) extends CpgPass(atom):
                 }
             }
         }
+    end run
 end CdxPass
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "appthreat-chen"
-version = "1.0.2"
+version = "1.0.3"
 description = "Code Hierarchy Exploration Net (chen)"
 authors = ["Team AppThreat <cloud@appthreat.com>"]
 license = "Apache-2.0"