diff --git a/org.eclipse.jgit.test/tst/org/eclipse/jgit/util/RawParseUtils_LineMapTest.java b/org.eclipse.jgit.test/tst/org/eclipse/jgit/util/RawParseUtils_LineMapTest.java index e8e191d78..2e9cbb503 100644 --- a/org.eclipse.jgit.test/tst/org/eclipse/jgit/util/RawParseUtils_LineMapTest.java +++ b/org.eclipse.jgit.test/tst/org/eclipse/jgit/util/RawParseUtils_LineMapTest.java @@ -79,6 +79,13 @@ public class RawParseUtils_LineMapTest { assertArrayEquals(new int[]{Integer.MIN_VALUE, 0, 4, buf.length}, asInts(map)); } + @Test + public void testBinary() throws UnsupportedEncodingException { + final byte[] buf = "xxxfoo\nb\0ar".getBytes("ISO-8859-1"); + final IntList map = RawParseUtils.lineMap(buf, 3, buf.length); + assertArrayEquals(new int[]{Integer.MIN_VALUE, 3, buf.length}, asInts(map)); + } + @Test public void testFourLineBlanks() throws UnsupportedEncodingException { final byte[] buf = "foo\n\n\nbar\n".getBytes("ISO-8859-1"); diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/util/RawParseUtils.java b/org.eclipse.jgit/src/org/eclipse/jgit/util/RawParseUtils.java index 86777b9cd..ad138bbf1 100644 --- a/org.eclipse.jgit/src/org/eclipse/jgit/util/RawParseUtils.java +++ b/org.eclipse.jgit/src/org/eclipse/jgit/util/RawParseUtils.java @@ -618,6 +618,10 @@ public final class RawParseUtils { *

* The last element (index map.size()-1) always contains * end. + *

+ * If the data contains a '\0' anywhere, the whole region is considered binary + * and a LineMap corresponding to a single line is returned. + *

* * @param buf * buffer to scan. @@ -629,14 +633,29 @@ public final class RawParseUtils { * @return a line map indexing the start position of each line. */ public static final IntList lineMap(final byte[] buf, int ptr, int end) { + 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. - // - final IntList map = new IntList((end - ptr) / 36); - map.fillTo(1, Integer.MIN_VALUE); - for (; ptr < end; ptr = nextLF(buf, ptr)) - map.add(ptr); + IntList map = new IntList((end - ptr) / 36); + map.add(Integer.MIN_VALUE); + boolean foundLF = true; + for (; ptr < end; ptr++) { + if (foundLF) { + map.add(ptr); + } + + if (buf[ptr] == '\0') { + // binary data. + map = new IntList(3); + map.add(Integer.MIN_VALUE); + map.add(start); + break; + } + + foundLF = (buf[ptr] == '\n'); + } map.add(end); return map; }