Land #777: AtU8 UTF8 atoms chunk decompilation

src/org/elixir_lang/beam/Beam.java

@@ -14,12 +14,14 @@
 import java.io.ByteArrayInputStream;
 import java.io.DataInputStream;
 import java.io.IOException;
+import java.nio.charset.Charset;
 import java.util.ArrayList;
 import java.util.Collection;
 import java.util.List;
 import java.util.Map;
 
 import static org.elixir_lang.beam.chunk.Chunk.TypeID.ATOM;
+import static org.elixir_lang.beam.chunk.Chunk.TypeID.ATU8;
 import static org.elixir_lang.beam.chunk.Chunk.TypeID.EXPT;
 import static org.elixir_lang.beam.chunk.Chunk.length;
 import static org.elixir_lang.beam.chunk.Chunk.typeID;
@@ -166,7 +168,13 @@ public Atoms atoms() {
         Chunk chunk = chunk(ATOM);
 
         if (chunk != null) {
-            atoms = Atoms.from(chunk);
+            atoms = Atoms.from(chunk, ATOM, Charset.forName("LATIN1"));
+        } else {
+            chunk = chunk(ATU8);
+
+            if (chunk != null) {
+                atoms = Atoms.from(chunk, ATU8, Charset.forName("UTF-8"));
+            }
         }
 
         return atoms;

src/org/elixir_lang/beam/Decompiler.java

@@ -19,6 +19,7 @@
 import java.util.List;
 import java.util.SortedSet;
 
+import static org.elixir_lang.beam.chunk.Chunk.TypeID.ATOM;
 import static org.elixir_lang.psi.call.name.Function.DEF;
 import static org.elixir_lang.psi.call.name.Function.DEFMACRO;
 import static org.elixir_lang.psi.call.name.Module.ELIXIR_PREFIX;
@@ -36,7 +37,7 @@ public class Decompiler implements BinaryFileDecompiler {
 
     @NotNull
     private static CharSequence decompiled(@Nullable Beam beam) {
-        StringBuilder decompiled = new StringBuilder("Decompilated Error");
+        StringBuilder decompiled = new StringBuilder("# Decompilation Error: ");
 
         if (beam != null) {
             Atoms atoms = beam.atoms();
@@ -57,8 +58,14 @@ private static CharSequence decompiled(@Nullable Beam beam) {
                     appendExports(decompiled, beam, atoms);
 
                     decompiled.append("end\n");
+                } else {
+                    decompiled.append("No module name found in ").append(ATOM).append(" chunk in BEAM");
                 }
+            } else {
+                decompiled.append("No ").append(ATOM).append(" chunk found in BEAM");
             }
+        } else {
+            decompiled.append("BEAM format could not be read");
         }
 
         return decompiled;

src/org/elixir_lang/beam/chunk/Atoms.java

@@ -4,11 +4,11 @@
 import org.jetbrains.annotations.NotNull;
 import org.jetbrains.annotations.Nullable;
 
+import java.nio.charset.Charset;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.List;
 
-import static org.elixir_lang.beam.chunk.Chunk.TypeID.ATOM;
 import static org.elixir_lang.beam.chunk.Chunk.unsignedByte;
 import static org.elixir_lang.beam.chunk.Chunk.unsignedInt;
 
@@ -18,13 +18,9 @@ public class Atoms {
      */
 
     @NotNull
-    private List<String> atomList;
+    private final List<String> atomList;
 
-    /*
-     * Constructors
-     */
-
-    public Atoms(@NotNull List<String> atomList) {
+    private Atoms(@NotNull List<String> atomList) {
         this.atomList = Collections.unmodifiableList(atomList);
     }
 
@@ -33,23 +29,23 @@ public Atoms(@NotNull List<String> atomList) {
      */
 
     @Nullable
-    public static Atoms from(@NotNull Chunk chunk) {
+    public static Atoms from(@NotNull Chunk chunk, @NotNull Chunk.TypeID typeID, @NotNull Charset charset) {
         Atoms atoms = null;
 
-        if (chunk.typeID.equals(ATOM.toString()) && chunk.data.length >= 4) {
+        if (chunk.typeID.equals(typeID.toString()) && chunk.data.length >= 4) {
             int offset = 0;
             Pair<Long, Integer> atomCountByteCount = unsignedInt(chunk.data, offset);
             long atomCount = atomCountByteCount.first;
             offset += atomCountByteCount.second;
 
-            List<String> atomList = new ArrayList<String>();
+            List<String> atomList = new ArrayList<>();
 
             for (long i = 0; i < atomCount; i++) {
                 Pair<Integer, Integer> atomLengthByteCount = unsignedByte(chunk.data[offset]);
                 int atomLength = atomLengthByteCount.first;
                 offset += atomLengthByteCount.second;
 
-                String entry = new String(chunk.data, offset, atomLength);
+                String entry = new String(chunk.data, offset, atomLength, charset);
                 offset += atomLength;
                 atomList.add(entry);
             }
@@ -65,9 +61,8 @@ public static Atoms from(@NotNull Chunk chunk) {
      */
 
     /**
-     *
      * @param index 1-based index.  1 is reserved for {#link moduleName}
-     * @return atom if
+     * @return atom if it exists
      */
     @Nullable
     public String get(int index) {

src/org/elixir_lang/beam/chunk/Chunk.java

@@ -101,6 +101,7 @@ static Pair<Long, Integer> unsignedInt(@NotNull byte[] bytes, int offset) {
 
     public enum TypeID {
         ATOM("Atom"),
+        ATU8("AtU8"),
         EXPT("ExpT");
 
         private final String typeID;

testData/org/elixir_lang/beam/decompiler/OTP20/Elixir.AtU8Test.ex

@@ -0,0 +1,21 @@
+# Source code recreated from a .beam file by IntelliJ Elixir
+defmodule AtU8Test do
+
+  # Macros
+
+  def __info__(p0) do
+    # body not decompiled
+  end
+
+  def module_info() do
+    # body not decompiled
+  end
+
+  def module_info(p0) do
+    # body not decompiled
+  end
+
+  def こんにちは世界() do
+    # body not decompiled
+  end
+end