Throw BinaryBlobException from RawParseUtils#lineMap.

This makes detection of binaries exact for ResolveMerger and
DiffFormatter: they will classify files as binary regardless of where
the '\0' occurs in the text.

Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Change-Id: Id4342a199628d9406bfa04af1b023c27a47d4014

org.eclipse.jgit.test/tst/org/eclipse/jgit/diff/RawTextTest.java

@@ -64,6 +64,16 @@ public class RawTextTest {
 		assertEquals(0, r.size());
 	}
 
+	@Test
+	public void testBinary() {
+		String input = "foo-a\nf\0o-b\n";
+		byte[] data = Constants.encodeASCII(input);
+		final RawText a = new RawText(data);
+		assertEquals(a.content, data);
+		assertEquals(a.size(), 1);
+		assertEquals(a.getString(0, 1, false), input);
+	}
+
 	@Test
 	public void testEquals() {
 		final RawText a = new RawText(Constants.encodeASCII("foo-a\nfoo-b\n"));

org.eclipse.jgit.test/tst/org/eclipse/jgit/util/RawParseUtils_LineMapTest.java

@@ -48,45 +48,45 @@ import static org.junit.Assert.assertNotNull;
 
 import java.io.UnsupportedEncodingException;
 
+import org.eclipse.jgit.errors.BinaryBlobException;
 import org.junit.Test;
 
 public class RawParseUtils_LineMapTest {
 	@Test
-	public void testEmpty() {
+	public void testEmpty() throws Exception {
 		final IntList map = RawParseUtils.lineMap(new byte[] {}, 0, 0);
 		assertNotNull(map);
 		assertArrayEquals(new int[]{Integer.MIN_VALUE, 0}, asInts(map));
 	}
 
 	@Test
-	public void testOneBlankLine() {
+	public void testOneBlankLine() throws Exception  {
 		final IntList map = RawParseUtils.lineMap(new byte[] { '\n' }, 0, 1);
 		assertArrayEquals(new int[]{Integer.MIN_VALUE, 0, 1}, asInts(map));
 	}
 
 	@Test
-	public void testTwoLineFooBar() throws UnsupportedEncodingException {
+	public void testTwoLineFooBar() throws Exception {
 		final byte[] buf = "foo\nbar\n".getBytes("ISO-8859-1");
 		final IntList map = RawParseUtils.lineMap(buf, 0, buf.length);
 		assertArrayEquals(new int[]{Integer.MIN_VALUE, 0, 4, buf.length}, asInts(map));
 	}
 
 	@Test
-	public void testTwoLineNoLF() throws UnsupportedEncodingException {
+	public void testTwoLineNoLF() throws Exception {
 		final byte[] buf = "foo\nbar".getBytes("ISO-8859-1");
 		final IntList map = RawParseUtils.lineMap(buf, 0, buf.length);
 		assertArrayEquals(new int[]{Integer.MIN_VALUE, 0, 4, buf.length}, asInts(map));
 	}
 
-	@Test
+	@Test(expected = BinaryBlobException.class)
-	public void testBinary() throws UnsupportedEncodingException {
+	public void testBinary() throws Exception {
 		final byte[] buf = "xxxfoo\nb\0ar".getBytes("ISO-8859-1");
-		final IntList map = RawParseUtils.lineMap(buf, 3, buf.length);
+		RawParseUtils.lineMap(buf, 3, buf.length);
-		assertArrayEquals(new int[]{Integer.MIN_VALUE, 3, buf.length}, asInts(map));
 	}
 
 	@Test
-	public void testFourLineBlanks() throws UnsupportedEncodingException {
+	public void testFourLineBlanks() throws Exception {
 		final byte[] buf = "foo\n\n\nbar\n".getBytes("ISO-8859-1");
 		final IntList map = RawParseUtils.lineMap(buf, 0, buf.length);
 

org.eclipse.jgit/src/org/eclipse/jgit/diff/RawText.java

@@ -93,7 +93,29 @@ public class RawText extends Sequence {
 	 */
 	public RawText(final byte[] input) {
 		content = input;
-		lines = RawParseUtils.lineMap(content, 0, content.length);
+		IntList map;
+		try {
+			map = RawParseUtils.lineMap(content, 0, content.length);
+		} catch (BinaryBlobException e) {
+			map = new IntList(3);
+			map.add(Integer.MIN_VALUE);
+			map.add(0);
+			map.add(content.length);
+		}
+		lines = map;
+	}
+
+	/**
+	 * Construct a new RawText if the line map is already known.
+	 *
+	 * @param data
+	 *   the blob data.
+	 * @param lineMap
+	 *   Indices of line starts, with indexed by base-1 linenumber.
+	 */
+	private RawText(final byte[] data, final IntList lineMap) {
+		content = data;
+		lines = lineMap;
 	}
 
 	/**
@@ -357,7 +379,8 @@ public class RawText extends Sequence {
 
 			System.arraycopy(head, 0, data, 0, head.length);
 			IO.readFully(stream, data, off, (int) (sz-off));
-			return new RawText(data);
+			IntList lineMap = RawParseUtils.lineMap(data, 0, data.length);
+			return new RawText(data, lineMap);
 		}
 	}
 }

org.eclipse.jgit/src/org/eclipse/jgit/util/RawParseUtils.java

@@ -63,6 +63,7 @@ import java.util.HashMap;
 import java.util.Map;
 
 import org.eclipse.jgit.annotations.Nullable;
+import org.eclipse.jgit.errors.BinaryBlobException;
 import org.eclipse.jgit.lib.Constants;
 import org.eclipse.jgit.lib.PersonIdent;
 
@@ -618,9 +619,6 @@ public final class RawParseUtils {
 	 * <p>
 	 * The last element (index <code>map.size()-1</code>) always contains
 	 * <code>end</code>.
-	 * <p>
-	 * If the data contains a '\0' anywhere, the whole region is considered binary
-	 * and a LineMap corresponding to a single line is returned.
 	 * </p>
 	 *
 	 * @param buf
@@ -631,10 +629,9 @@ public final class RawParseUtils {
 	 * @param end
 	 *            1 past the end of the content within <code>buf</code>.
 	 * @return a line map indexing the start position of each line.
+	 * @throws BinaryBlobException if any '\0' is found.
 	 */
-	public static final IntList lineMap(final byte[] buf, int ptr, int end) {
+	public static final IntList lineMap(final byte[] buf, int ptr, int end) throws BinaryBlobException {
-		int start = ptr;
-
 		// Experimentally derived from multiple source repositories
 		// the average number of bytes/line is 36. Its a rough guess
 		// to initially size our map close to the target.
@@ -647,11 +644,7 @@ public final class RawParseUtils {
 			}
 
 			if (buf[ptr] == '\0') {
-				// binary data.
+				throw new BinaryBlobException();
-				map = new IntList(3);
-				map.add(Integer.MIN_VALUE);
-				map.add(start);
-				break;
 			}
 
 			foundLF = (buf[ptr] == '\n');