From 65f438e56b88843fd91287ca60b2e5bdd3af9cb7 Mon Sep 17 00:00:00 2001
From: Guillaume Martres <smarter@ubuntu.com>
Date: Tue, 11 Apr 2017 19:16:44 +0200
Subject: SI-9915 Utf8_info are modified UTF8

Adapted from scalac commit 3c5990ce5839f4bdfca8fed7f2c415a72f6a8bd8 by
Som Snytt:

Use DataInputStream.readUTF to read CONSTANT_Utf8_info.

This fixes reading embedded null char and supplementary chars.
---
 .../tools/dotc/core/classfile/ClassfileParser.scala  |  8 ++++++--
 tests/run/t9915/C_1.java                             | 20 ++++++++++++++++++++
 tests/run/t9915/Test_2.scala                         | 12 ++++++++++++
 3 files changed, 38 insertions(+), 2 deletions(-)
 create mode 100644 tests/run/t9915/C_1.java
 create mode 100644 tests/run/t9915/Test_2.scala

diff --git a/compiler/src/dotty/tools/dotc/core/classfile/ClassfileParser.scala b/compiler/src/dotty/tools/dotc/core/classfile/ClassfileParser.scala
index bebc4ab2c..9415c047f 100644
--- a/compiler/src/dotty/tools/dotc/core/classfile/ClassfileParser.scala
+++ b/compiler/src/dotty/tools/dotc/core/classfile/ClassfileParser.scala
@@ -7,7 +7,7 @@ import Contexts._, Symbols._, Types._, Names._, StdNames._, NameOps._, Scopes._,
 import SymDenotations._, unpickleScala2.Scala2Unpickler._, Constants._, Annotations._, util.Positions._
 import NameKinds.{ModuleClassName, DefaultGetterName}
 import ast.tpd._
-import java.io.{ File, IOException }
+import java.io.{ ByteArrayInputStream, DataInputStream, File, IOException }
 import java.lang.Integer.toHexString
 import scala.collection.{ mutable, immutable }
 import scala.collection.mutable.{ ListBuffer, ArrayBuffer }
@@ -935,12 +935,16 @@ class ClassfileParser(
         case null   =>
           val start = starts(index)
           if (in.buf(start).toInt != CONSTANT_UTF8) errorBadTag(start)
-          val name = termName(in.buf, start + 3, in.getChar(start + 1))
+          val len   = in.getChar(start + 1).toInt
+          val name = termName(fromMUTF8(in.buf, start + 1, len + 2))
           values(index) = name
           name
       }
     }
 
+    private def fromMUTF8(bytes: Array[Byte], offset: Int, len: Int): String =
+      new DataInputStream(new ByteArrayInputStream(bytes, offset, len)).readUTF
+
     /** Return the name found at given index in the constant pool, with '/' replaced by '.'. */
     def getExternalName(index: Int): SimpleTermName = {
       if (index <= 0 || len <= index)
diff --git a/tests/run/t9915/C_1.java b/tests/run/t9915/C_1.java
new file mode 100644
index 000000000..4269cf74e
--- /dev/null
+++ b/tests/run/t9915/C_1.java
@@ -0,0 +1,20 @@
+/*
+ * javac: -encoding UTF-8
+ */
+public class C_1 {
+    public static final String NULLED = "X\000ABC";
+    public static final String SUPPED = "𐒈𐒝𐒑𐒛𐒐𐒘𐒕𐒖";
+
+    public String nulled() {
+        return C_1.NULLED;
+    }
+    public String supped() {
+        return C_1.SUPPED;
+    }
+    public int nulledSize() {
+        return C_1.NULLED.length();
+    }
+    public int suppedSize() {
+        return C_1.SUPPED.length();
+    }
+}
diff --git a/tests/run/t9915/Test_2.scala b/tests/run/t9915/Test_2.scala
new file mode 100644
index 000000000..afed667cc
--- /dev/null
+++ b/tests/run/t9915/Test_2.scala
@@ -0,0 +1,12 @@
+
+object Test extends App {
+  val c = new C_1
+  assert(c.nulled == "X\u0000ABC")    // "X\000ABC"
+  assert(c.supped == "𐒈𐒝𐒑𐒛𐒐𐒘𐒕𐒖")
+
+  assert(C_1.NULLED == "X\u0000ABC")  // "X\000ABC"
+  assert(C_1.SUPPED == "𐒈𐒝𐒑𐒛𐒐𐒘𐒕𐒖")
+
+  assert(C_1.NULLED.size == "XYABC".size)
+  assert(C_1.SUPPED.codePointCount(0, C_1.SUPPED.length) == 8)
+}
-- 
cgit v1.2.3