Drop unsupported characters from XML

Author: AlexDenisov

include/Properties.h

@@ -7,7 +7,7 @@ namespace llvm2graphml {
 
 class Properties {
 public:
-  void setStringProperty(const std::string &key, std::string value);
+  void setStringProperty(const std::string &key, const std::string &value);
   void setBooleanProperty(const std::string &key, bool value);
   void setLongProperty(const std::string &key, uint64_t value);
 

src/Properties.cpp

@@ -3,8 +3,21 @@
 
 using namespace llvm2graphml;
 
-void Properties::setStringProperty(const std::string& key, std::string value) {
-  strings[key] = std::move(value);
+static std::string dropUnsupportedXmlChars(const std::string &input) {
+  std::string output;
+  output.reserve(input.size());
+  for (char c : input) {
+    if (iscntrl(c) && (c != 0x9) && (c != 0xa) && (c != 0xd)) {
+      continue;
+    }
+    output.push_back(c);
+  }
+
+  return output;
+}
+
+void Properties::setStringProperty(const std::string &key, const std::string &value) {
+  strings[key] = dropUnsupportedXmlChars(value);
 }
 
 void Properties::setBooleanProperty(const std::string &key, bool value) {

tests/CMakeLists.txt

@@ -7,8 +7,8 @@ else()
   include(ExternalProject)
 
   ExternalProject_Add(gremlin
-    URL http://mirror.funkfreundelandshut.de/apache/tinkerpop/3.4.5/apache-tinkerpop-gremlin-console-3.4.5-bin.zip
-    URL_HASH MD5=1ad2e9b2414846aed32213e1b7b331eb
+    URL https://downloads.apache.org/tinkerpop/3.4.6/apache-tinkerpop-gremlin-console-3.4.6-bin.zip
+    URL_HASH MD5=e810f5cc9890a23bdf5f4aca1b4f0471
     CONFIGURE_COMMAND ""
     BUILD_COMMAND ""
     INSTALL_COMMAND ""
@@ -26,7 +26,7 @@ set(LIT_COMMAND
   GREMLIN_CONSOLE_EXEC=${GREMLIN_CONSOLE_EXEC}
   FILECHECK_EXEC=${FILECHECK_EXEC}
   ${LIT_EXEC}
-  -vv -j2
+  -vv
   ${CMAKE_CURRENT_LIST_DIR}/integration-tests
   )
 

tests/integration-tests/regression/main.ll

@@ -0,0 +1,9 @@
+%0 = type opaque
+
+define internal i32 @"\01-[Shape square]"(%0*, i8*) {
+  %3 = alloca %0*, align 8
+  %4 = alloca i8*, align 8
+  store %0* %0, %0** %3, align 8
+  store i8* %1, i8** %4, align 8
+  ret i32 42
+}

tests/integration-tests/regression/test.groovy

@@ -0,0 +1,23 @@
+/**
+; RUN: cd %p
+; RUN: %LLVM2GRAPHML_EXEC --output=%t.xml main.ll
+; RUN: %GREMLIN_CONSOLE_EXEC -e %s %t.xml | %FILECHECK_EXEC %s --match-full-lines
+*/
+
+/// Prepare
+
+graph = TinkerGraph.open()
+g = graph.traversal()
+g.io(args[0]).read().iterate()
+
+def dump(tr) {
+  tr.unfold().sideEffect{println it}.iterate()
+}
+
+/// Assertions
+
+dump(g.V().has('function', 'name', '-[Shape square]').valueMap('name'))
+// CHECK: name=[-[Shape square]]
+
+:exit
+// CHECK-EMPTY: