Parse module name out of .beam files

.gitignore

@@ -1,6 +1,7 @@
 /tmp
 # user-specific project metadata
 /.idea/dictionaries/
+/.idea/markdown-navigator*
 /.idea/workspace.xml
 /.idea/shelf/
 # ignore idea.xml output

src/org/elixir_lang/beam/Beam.java

@@ -0,0 +1,105 @@
+package org.elixir_lang.beam;
+
+import com.ericsson.otp.erlang.OtpErlangDecodeException;
+import gnu.trove.THashMap;
+import org.elixir_lang.beam.chunk.Atoms;
+import org.elixir_lang.beam.chunk.Chunk;
+import org.jetbrains.annotations.NotNull;
+import org.jetbrains.annotations.Nullable;
+
+import java.io.DataInputStream;
+import java.io.IOException;
+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.length;
+import static org.elixir_lang.beam.chunk.Chunk.typeID;
+
+/**
+ * See http://beam-wisdoms.clau.se/en/latest/indepth-beam-file.html
+ */
+public class Beam {
+    /*
+     * CONSTANTS
+     */
+
+    private static final String HEADER = "FOR1";
+
+    /*
+     * Fields
+     */
+
+    private Map<String, Chunk> chunkByTypeID;
+
+    /*
+     * Constructors
+     */
+
+    private Beam(@NotNull Collection<Chunk> chunkCollection) {
+        chunkByTypeID = new THashMap<String, Chunk>(chunkCollection.size());
+
+        for (Chunk chunk : chunkCollection) {
+            chunkByTypeID.put(chunk.typeID, chunk);
+        }
+    }
+
+    /*
+     * Static Methods
+     */
+
+    @Nullable
+    public static Beam from(@NotNull DataInputStream dataInputStream) throws IOException, OtpErlangDecodeException {
+        if (!HEADER.equals(typeID(dataInputStream))) {
+            return null;
+        }
+
+        length(dataInputStream);
+
+        if (!"BEAM".equals(typeID(dataInputStream))) {
+            return null;
+        }
+
+        List<Chunk> chunkList = new ArrayList<Chunk>();
+        while (true) {
+            Chunk chunk = Chunk.from(dataInputStream);
+
+            if (chunk != null) {
+                chunkList.add(chunk);
+            } else {
+                break;
+            }
+        }
+
+        return new Beam(chunkList);
+    }
+
+    /*
+     * Instance Methods
+     */
+
+    @Nullable
+    public Atoms atoms() {
+        Atoms atoms = null;
+
+        Chunk chunk = chunk(ATOM);
+
+        if (chunk != null) {
+            atoms = Atoms.from(chunk);
+        }
+
+        return atoms;
+    }
+
+    @Nullable
+    private Chunk chunk(@NotNull String typeID) {
+        return chunkByTypeID.get(typeID);
+    }
+
+    @Nullable
+    private Chunk chunk(@NotNull Chunk.TypeID typeID) {
+        return chunk(typeID.toString());
+    }
+}

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

@@ -0,0 +1,73 @@
+package org.elixir_lang.beam.chunk;
+
+import org.jetbrains.annotations.NotNull;
+import org.jetbrains.annotations.Nullable;
+
+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;
+
+public class Atoms {
+    /*
+     * Fields
+     */
+
+    @NotNull
+    private List<String> atomList;
+
+    /*
+     * Constructors
+     */
+
+    public Atoms(@NotNull List<String> atomList) {
+        this.atomList = Collections.unmodifiableList(atomList);
+    }
+
+    /*
+     * Static Methods
+     */
+
+    @Nullable
+    public static Atoms from(@NotNull Chunk chunk) {
+        Atoms atoms = null;
+
+        if (chunk.typeID.equals(ATOM.toString()) && chunk.data.length >= 4) {
+            int offset = 0;
+            long atomCount = unsignedInt(chunk.data, offset);
+            List<String> atomList = new ArrayList<String>();
+            offset += 4;
+
+            for (long i = 0; i < atomCount; i++) {
+                int atomLength = unsignedByte(chunk.data[offset]);
+                offset += 1;
+
+                String entry = new String(chunk.data, offset, atomLength);
+                offset += atomLength;
+                atomList.add(entry);
+            }
+
+            atoms = new Atoms(atomList);
+        }
+
+        return atoms;
+    }
+
+    /*
+     * Instance Methods
+     */
+
+    @Nullable
+    public String moduleName() {
+        String moduleName = null;
+
+        if (atomList.size() > 0) {
+            moduleName = atomList.get(0);
+        }
+
+        return moduleName;
+    }
+}

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

@@ -0,0 +1,105 @@
+package org.elixir_lang.beam.chunk;
+
+import org.jetbrains.annotations.Contract;
+import org.jetbrains.annotations.NotNull;
+import org.jetbrains.annotations.Nullable;
+
+import java.io.DataInputStream;
+import java.io.IOException;
+
+/**
+ * Chunk of a `.beam` file.  Same chunk format as base IFF
+ */
+public class Chunk {
+    private static final int ALIGNMENT = 4;
+    @NotNull
+    public final String typeID;
+    @NotNull
+    public final byte[] data;
+    private Chunk(@NotNull String typeID, @NotNull byte[] data) {
+        this.typeID = typeID;
+        this.data = data;
+    }
+
+    @Nullable
+    public static Chunk from(@NotNull DataInputStream dataInputStream) throws IOException {
+        String typeID = typeID(dataInputStream);
+        Chunk chunk = null;
+
+        if (typeID != null) {
+            long length = length(dataInputStream);
+
+            byte[] data = new byte[(int) length];
+            dataInputStream.readFully(data);
+
+            int padding = (int) ((ALIGNMENT - (length % ALIGNMENT)) % ALIGNMENT);
+            dataInputStream.skipBytes(padding);
+
+            chunk = new Chunk(typeID, data);
+        }
+
+        return chunk;
+    }
+
+    public static long length(@NotNull DataInputStream dataInputStream) throws IOException {
+        return readUnsignedInt(dataInputStream);
+    }
+
+    @Nullable
+    public static String typeID(@NotNull DataInputStream dataInputStream) throws IOException {
+        byte[] bytes = new byte[4];
+        int bytesRead = dataInputStream.read(bytes, 0, bytes.length);
+        String typeID = null;
+
+        if (bytesRead == bytes.length) {
+            typeID = new String(bytes);
+        }
+
+        return typeID;
+    }
+
+    private static long readUnsignedInt(@NotNull DataInputStream dataInputStream) throws IOException {
+        byte[] bytes = new byte[32 / 8];
+
+        dataInputStream.readFully(bytes);
+
+        return unsignedInt(bytes);
+    }
+
+    static int unsignedByte(byte signedByte) {
+        return signedByte & 0xFF;
+    }
+
+    static long unsignedInt(@NotNull byte[] bytes) {
+        return unsignedInt(bytes, 0);
+    }
+
+    @Contract(pure = true)
+    static long unsignedInt(@NotNull byte[] bytes, int offset) {
+        assert bytes.length >= offset + 4;
+
+        long unsignedInt = 0;
+
+        for (int i = 0; i < 4; i++) {
+            int unsignedByte = unsignedByte(bytes[offset + i]);
+            unsignedInt += unsignedByte << 8 * (4 - 1 - i);
+        }
+
+        return unsignedInt;
+    }
+
+    public enum TypeID {
+        ATOM("Atom");
+
+        private final String typeID;
+
+        TypeID(@NotNull String typeID) {
+            this.typeID = typeID;
+        }
+
+        @Override
+        public String toString() {
+            return typeID;
+        }
+    }
+}

tests/org/elixir_lang/beam/BeamTest.java

@@ -0,0 +1,52 @@
+package org.elixir_lang.beam;
+
+import com.ericsson.otp.erlang.OtpErlangDecodeException;
+import org.elixir_lang.beam.chunk.Atoms;
+import org.jetbrains.annotations.NotNull;
+import org.junit.Before;
+import org.junit.Test;
+
+import java.io.*;
+
+import static org.junit.Assert.*;
+
+public class BeamTest {
+    @NotNull
+    private String ebinDirectory;
+
+    @Test
+    public void elixirModule() throws IOException, OtpErlangDecodeException {
+        assertModuleName("Elixir.Kernel", "Elixir.Kernel");
+    }
+
+    @Test
+    public void erlangModule() throws IOException, OtpErlangDecodeException {
+        assertModuleName("elixir_interpolation", "elixir_interpolation");
+    }
+
+    private void assertModuleName(@NotNull String baseName, @NotNull String moduleName) throws IOException, OtpErlangDecodeException {
+                DataInputStream dataInputStream = new DataInputStream(
+                new BufferedInputStream(
+                        new FileInputStream(ebinDirectory + baseName + ".beam")
+                )
+        );
+        Beam beam = Beam.from(dataInputStream);
+
+        assertNotNull(beam);
+
+        Atoms atoms = beam.atoms();
+
+        assertNotNull(atoms);
+
+        assertEquals(atoms.moduleName(), moduleName);
+    }
+
+    @Before
+    public void setEbinDirectory() {
+        String ebinDirectory = System.getenv("ELIXIR_EBIN_DIRECTORY");
+
+        assertNotNull("ELIXIR_EBIN_DIRECTORY is not set", ebinDirectory);
+
+        this.ebinDirectory = ebinDirectory;
+    }
+}