You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tika.apache.org by ta...@apache.org on 2016/06/27 16:48:23 UTC
[1/2] tika git commit: TIKA-2023 -- clean up RTFParser to use
EndianUtils
Repository: tika
Updated Branches:
refs/heads/master e6c2839c0 -> 7db0ab628
TIKA-2023 -- clean up RTFParser to use EndianUtils
Project: http://git-wip-us.apache.org/repos/asf/tika/repo
Commit: http://git-wip-us.apache.org/repos/asf/tika/commit/c1cea20e
Tree: http://git-wip-us.apache.org/repos/asf/tika/tree/c1cea20e
Diff: http://git-wip-us.apache.org/repos/asf/tika/diff/c1cea20e
Branch: refs/heads/master
Commit: c1cea20efb5ed28cfd3fe889120d7ddbef9966a8
Parents: e6c2839
Author: tballison <ta...@mitre.org>
Authored: Mon Jun 27 12:46:47 2016 -0400
Committer: tballison <ta...@mitre.org>
Committed: Mon Jun 27 12:46:47 2016 -0400
----------------------------------------------------------------------
.../java/org/apache/tika/io/EndianUtils.java | 20 ++++++++++
.../org/apache/tika/io/EndianUtilsTest.java | 19 +++++++++
.../tika/parser/rtf/RTFEmbObjHandler.java | 7 +---
.../tika/parser/rtf/RTFObjDataParser.java | 41 ++++++++------------
.../apache/tika/parser/rtf/TextExtractor.java | 11 +-----
5 files changed, 60 insertions(+), 38 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/tika/blob/c1cea20e/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
----------------------------------------------------------------------
diff --git a/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java b/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
index 1e33986..3416f55 100644
--- a/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
+++ b/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
@@ -70,6 +70,26 @@ public class EndianUtils {
return (ch1 << 8) + (ch2 << 0);
}
+
+ /**
+ * Get a LE unsigned int value from an InputStream
+ *
+ * @param stream the InputStream from which the int is to be read
+ * @return the int (32-bit) value
+ * @exception IOException will be propagated back to the caller
+ * @exception BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static long readUIntLE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return ((ch4 << 24) + (ch3<<16) + (ch2 << 8) + (ch1 << 0)) & 0x00FFFFFFFFl;
+ }
+
/**
* Get a LE int value from an InputStream
*
http://git-wip-us.apache.org/repos/asf/tika/blob/c1cea20e/tika-core/src/test/java/org/apache/tika/io/EndianUtilsTest.java
----------------------------------------------------------------------
diff --git a/tika-core/src/test/java/org/apache/tika/io/EndianUtilsTest.java b/tika-core/src/test/java/org/apache/tika/io/EndianUtilsTest.java
index 7f10cdd..8f481c3 100644
--- a/tika-core/src/test/java/org/apache/tika/io/EndianUtilsTest.java
+++ b/tika-core/src/test/java/org/apache/tika/io/EndianUtilsTest.java
@@ -18,6 +18,7 @@
package org.apache.tika.io;
import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.fail;
import java.io.ByteArrayInputStream;
@@ -37,4 +38,22 @@ public class EndianUtilsTest {
data = new byte[] { (byte)0xac, (byte)0xbe, 0x17 };
assertEquals((long)728855, EndianUtils.readUE7(new ByteArrayInputStream(data)));
}
+
+ @Test
+ public void testReadUIntLE() throws Exception {
+ byte[] data = new byte[] {(byte)0x08, (byte)0x00, (byte)0x00, (byte)0x00 };
+ assertEquals((long) 8, EndianUtils.readUIntLE(new ByteArrayInputStream(data)));
+
+ data = new byte[] {(byte)0xFF, (byte)0xFF, (byte)0xFF, (byte)0xFF };
+ assertEquals(4294967295L, EndianUtils.readUIntLE(new ByteArrayInputStream(data)));
+
+ data = new byte[] {(byte)0xFF, (byte)0xFF, (byte)0xFF };
+ try {
+ EndianUtils.readUIntLE(new ByteArrayInputStream(data));
+ fail("Should have thrown exception");
+ } catch (EndianUtils.BufferUnderrunException e) {
+
+ }
+
+ }
}
http://git-wip-us.apache.org/repos/asf/tika/blob/c1cea20e/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFEmbObjHandler.java
----------------------------------------------------------------------
diff --git a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFEmbObjHandler.java b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFEmbObjHandler.java
index 183df62..1334906 100644
--- a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFEmbObjHandler.java
+++ b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFEmbObjHandler.java
@@ -22,6 +22,7 @@ import java.io.InputStream;
import java.util.concurrent.atomic.AtomicInteger;
import org.apache.commons.io.FilenameUtils;
+import org.apache.commons.io.IOUtils;
import org.apache.tika.config.TikaConfig;
import org.apache.tika.detect.Detector;
import org.apache.tika.exception.TikaException;
@@ -157,11 +158,7 @@ class RTFEmbObjHandler {
}
byte[] bytes = new byte[len];
- int bytesRead = is.read(bytes);
- if (bytesRead < len) {
- throw new TikaException("unexpected end of file: need " + len +
- " bytes of binary data, found " + (len - bytesRead));
- }
+ IOUtils.readFully(is, bytes);
os.write(bytes);
}
http://git-wip-us.apache.org/repos/asf/tika/blob/c1cea20e/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFObjDataParser.java
----------------------------------------------------------------------
diff --git a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFObjDataParser.java b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFObjDataParser.java
index cc9d62f..147d2e8 100644
--- a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFObjDataParser.java
+++ b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/RTFObjDataParser.java
@@ -36,6 +36,7 @@ import org.apache.poi.poifs.filesystem.NPOIFSFileSystem;
import org.apache.poi.poifs.filesystem.Ole10Native;
import org.apache.poi.poifs.filesystem.Ole10NativeException;
import org.apache.poi.util.IOUtils;
+import org.apache.tika.io.EndianUtils;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.metadata.RTFMetadata;
@@ -48,10 +49,6 @@ import org.apache.tika.parser.microsoft.OfficeParser.POIFSDocumentType;
*/
class RTFObjDataParser {
- private final static int[] INT_LE_POWS = new int[]{
- 1, 256, 65536, 16777216
- };
-
private final static String WIN_ASCII = "WINDOWS-1252";
/**
@@ -195,7 +192,12 @@ class RTFObjDataParser {
//should we add this to the metadata?
readAnsiString(is); //iconFilePath
- readUShort(is); //iconIndex
+ try {
+ //iconIndex
+ EndianUtils.readUShortBE(is);
+ } catch (EndianUtils.BufferUnderrunException e) {
+ throw new IOException(e);
+ }
int type = readUShort(is); //type
//1 is link, 3 is embedded object
@@ -209,7 +211,7 @@ class RTFObjDataParser {
String ansiFilePath = readAnsiString(is); //filePath
long bytesLen = readUInt(is);
byte[] objBytes = initByteArray(bytesLen);
- is.read(objBytes);
+ IOUtils.readFully(is, objBytes);
StringBuilder unicodeFilePath = new StringBuilder();
try {
@@ -248,24 +250,19 @@ class RTFObjDataParser {
private int readUShort(InputStream is) throws IOException {
- int lo = is.read();
- int hi = is.read() * 256;
- if (lo == -1 || hi == -1) {
- throw new IOException("Hit end of stream before reading little endian unsigned short.");
+ try {
+ return EndianUtils.readUShortLE(is);
+ } catch (EndianUtils.BufferUnderrunException e) {
+ throw new IOException(e);
}
- return hi + lo;
}
private long readUInt(InputStream is) throws IOException {
- long sum = 0;
- for (int i = 0; i < 4; i++) {
- int v = is.read();
- if (v == -1) {
- throw new IOException("Hit end of stream before finishing little endian unsigned int.");
- }
- sum += v * (long) INT_LE_POWS[i];
+ try {
+ return EndianUtils.readUIntLE(is);
+ } catch (EndianUtils.BufferUnderrunException e) {
+ throw new IOException(e);
}
- return sum;
}
private String readAnsiString(InputStream is) throws IOException {
@@ -296,11 +293,7 @@ class RTFObjDataParser {
private byte[] readBytes(InputStream is, long len) throws IOException {
//initByteArray tests for "reading of too many bytes"
byte[] bytes = initByteArray(len);
- int read = is.read(bytes);
- if (read != len) {
- throw new IOException("Hit end of stream before reading all bytes");
- }
-
+ IOUtils.readFully(is, bytes);
return bytes;
}
http://git-wip-us.apache.org/repos/asf/tika/blob/c1cea20e/tika-parsers/src/main/java/org/apache/tika/parser/rtf/TextExtractor.java
----------------------------------------------------------------------
diff --git a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/TextExtractor.java b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/TextExtractor.java
index eeb58ce..6c86765 100644
--- a/tika-parsers/src/main/java/org/apache/tika/parser/rtf/TextExtractor.java
+++ b/tika-parsers/src/main/java/org/apache/tika/parser/rtf/TextExtractor.java
@@ -33,6 +33,7 @@ import java.util.Locale;
import java.util.Map;
import java.util.TimeZone;
+import org.apache.commons.io.IOUtils;
import org.apache.tika.exception.TikaException;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.metadata.Office;
@@ -952,15 +953,7 @@ final class TextExtractor {
embObjHandler.reset();
}
} else {
- int bytesToRead = param;
- byte[] tmpArray = new byte[Math.min(1024, bytesToRead)];
- while (bytesToRead > 0) {
- int r = in.read(tmpArray, 0, Math.min(bytesToRead, tmpArray.length));
- if (r < 0) {
- throw new TikaException("unexpected end of file: need " + param + " bytes of binary data, found " + (param - bytesToRead));
- }
- bytesToRead -= r;
- }
+ IOUtils.skipFully(in, param);
}
} else {
// log some warning?
[2/2] tika git commit: TIKA-2023 -- clean up newlines and indenting
Posted by ta...@apache.org.
TIKA-2023 -- clean up newlines and indenting
Project: http://git-wip-us.apache.org/repos/asf/tika/repo
Commit: http://git-wip-us.apache.org/repos/asf/tika/commit/7db0ab62
Tree: http://git-wip-us.apache.org/repos/asf/tika/tree/7db0ab62
Diff: http://git-wip-us.apache.org/repos/asf/tika/diff/7db0ab62
Branch: refs/heads/master
Commit: 7db0ab6287b2c795ccd7a6383d344434e79611ae
Parents: c1cea20
Author: tballison <ta...@mitre.org>
Authored: Mon Jun 27 12:48:14 2016 -0400
Committer: tballison <ta...@mitre.org>
Committed: Mon Jun 27 12:48:14 2016 -0400
----------------------------------------------------------------------
.../java/org/apache/tika/io/EndianUtils.java | 831 ++++++++++---------
1 file changed, 421 insertions(+), 410 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/tika/blob/7db0ab62/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
----------------------------------------------------------------------
diff --git a/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java b/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
index 3416f55..2ab85b3 100644
--- a/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
+++ b/tika-core/src/main/java/org/apache/tika/io/EndianUtils.java
@@ -25,418 +25,429 @@ import org.apache.tika.exception.TikaException;
* General Endian Related Utilties.
* <p>
* This class provides static utility methods for input/output operations
- * on numbers in Big and Little Endian formats.
+ * on numbers in Big and Little Endian formats.
* <p>
* Origin of code: Based on the version in POI
*/
public class EndianUtils {
- /**
- * Get a LE short value from an InputStream
- *
- * @param stream the InputStream from which the short is to be read
- * @return the short (16-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static short readShortLE(InputStream stream) throws IOException, BufferUnderrunException {
- return (short) readUShortLE(stream);
- }
- /**
- * Get a BE short value from an InputStream
- *
- * @param stream the InputStream from which the short is to be read
- * @return the short (16-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static short readShortBE(InputStream stream) throws IOException, BufferUnderrunException {
- return (short) readUShortBE(stream);
- }
-
- public static int readUShortLE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- if ((ch1 | ch2) < 0) {
- throw new BufferUnderrunException();
- }
- return (ch2 << 8) + (ch1 << 0);
- }
- public static int readUShortBE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- if ((ch1 | ch2) < 0) {
- throw new BufferUnderrunException();
- }
- return (ch1 << 8) + (ch2 << 0);
- }
-
-
- /**
- * Get a LE unsigned int value from an InputStream
- *
- * @param stream the InputStream from which the int is to be read
- * @return the int (32-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static long readUIntLE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- int ch3 = stream.read();
- int ch4 = stream.read();
- if ((ch1 | ch2 | ch3 | ch4) < 0) {
- throw new BufferUnderrunException();
- }
- return ((ch4 << 24) + (ch3<<16) + (ch2 << 8) + (ch1 << 0)) & 0x00FFFFFFFFl;
- }
-
- /**
- * Get a LE int value from an InputStream
- *
- * @param stream the InputStream from which the int is to be read
- * @return the int (32-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static int readIntLE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- int ch3 = stream.read();
- int ch4 = stream.read();
- if ((ch1 | ch2 | ch3 | ch4) < 0) {
- throw new BufferUnderrunException();
- }
- return (ch4 << 24) + (ch3<<16) + (ch2 << 8) + (ch1 << 0);
- }
- /**
- * Get a BE int value from an InputStream
- *
- * @param stream the InputStream from which the int is to be read
- * @return the int (32-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static int readIntBE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- int ch3 = stream.read();
- int ch4 = stream.read();
- if ((ch1 | ch2 | ch3 | ch4) < 0) {
- throw new BufferUnderrunException();
- }
- return (ch1 << 24) + (ch2<<16) + (ch3 << 8) + (ch4 << 0);
- }
-
- /**
- * Get a LE long value from an InputStream
- *
- * @param stream the InputStream from which the long is to be read
- * @return the long (64-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static long readLongLE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- int ch3 = stream.read();
- int ch4 = stream.read();
- int ch5 = stream.read();
- int ch6 = stream.read();
- int ch7 = stream.read();
- int ch8 = stream.read();
- if ((ch1 | ch2 | ch3 | ch4 | ch5 | ch6 | ch7 | ch8) < 0) {
- throw new BufferUnderrunException();
- }
-
- return
- ((long)ch8 << 56) +
- ((long)ch7 << 48) +
- ((long)ch6 << 40) +
- ((long)ch5 << 32) +
- ((long)ch4 << 24) + // cast to long to preserve bit 31 (sign bit for ints)
- (ch3 << 16) +
- (ch2 << 8) +
- (ch1 << 0);
- }
- /**
- * Get a NE long value from an InputStream
- *
- * @param stream the InputStream from which the long is to be read
- * @return the long (64-bit) value
- * @exception IOException will be propagated back to the caller
- * @exception BufferUnderrunException if the stream cannot provide enough bytes
- */
- public static long readLongBE(InputStream stream) throws IOException, BufferUnderrunException {
- int ch1 = stream.read();
- int ch2 = stream.read();
- int ch3 = stream.read();
- int ch4 = stream.read();
- int ch5 = stream.read();
- int ch6 = stream.read();
- int ch7 = stream.read();
- int ch8 = stream.read();
- if ((ch1 | ch2 | ch3 | ch4 | ch5 | ch6 | ch7 | ch8) < 0) {
- throw new BufferUnderrunException();
- }
-
- return
- ((long)ch1 << 56) +
- ((long)ch2 << 48) +
- ((long)ch3 << 40) +
- ((long)ch4 << 32) +
- ((long)ch5 << 24) + // cast to long to preserve bit 31 (sign bit for ints)
- (ch6 << 16) +
- (ch7 << 8) +
- (ch8 << 0);
- }
-
- /**
- * Gets the integer value that is stored in UTF-8 like fashion, in Big Endian
- * but with the high bit on each number indicating if it continues or not
- */
- public static long readUE7(InputStream stream) throws IOException {
- int i;
- long v = 0;
- while ((i = stream.read()) >= 0) {
- v = v << 7;
- if ((i & 128) == 128) {
- // Continues
- v += (i&127);
- } else {
- // Last value
- v += i;
- break;
- }
- }
- return v;
- }
-
-
- /**
- * Get a LE short value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the short (16-bit) value
- */
- public static short getShortLE(byte[] data) {
- return getShortLE(data, 0);
- }
- /**
- * Get a LE short value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the short (16-bit) value
- */
- public static short getShortLE(byte[] data, int offset) {
- return (short)getUShortLE(data, offset);
- }
-
- /**
- * Get a LE unsigned short value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the unsigned short (16-bit) value in an int
- */
- public static int getUShortLE(byte[] data) {
- return getUShortLE(data, 0);
- }
- /**
- * Get a LE unsigned short value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the unsigned short (16-bit) value in an integer
- */
- public static int getUShortLE(byte[] data, int offset) {
- int b0 = data[offset] & 0xFF;
- int b1 = data[offset+1] & 0xFF;
- return (b1 << 8) + (b0 << 0);
- }
-
- /**
- * Get a BE short value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the short (16-bit) value
- */
- public static short getShortBE(byte[] data) {
- return getShortBE(data, 0);
- }
- /**
- * Get a BE short value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the short (16-bit) value
- */
- public static short getShortBE(byte[] data, int offset) {
- return (short)getUShortBE(data, offset);
- }
-
- /**
- * Get a BE unsigned short value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the unsigned short (16-bit) value in an int
- */
- public static int getUShortBE(byte[] data) {
- return getUShortBE(data, 0);
- }
- /**
- * Get a BE unsigned short value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the unsigned short (16-bit) value in an integer
- */
- public static int getUShortBE(byte[] data, int offset) {
- int b0 = data[offset] & 0xFF;
- int b1 = data[offset+1] & 0xFF;
- return (b0 << 8) + (b1 << 0);
- }
-
- /**
- * Get a LE int value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the int (32-bit) value
- */
- public static int getIntLE(byte[] data) {
- return getIntLE(data, 0);
- }
- /**
- * Get a LE int value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the int (32-bit) value
- */
- public static int getIntLE(byte[] data, int offset) {
- int i=offset;
- int b0 = data[i++] & 0xFF;
- int b1 = data[i++] & 0xFF;
- int b2 = data[i++] & 0xFF;
- int b3 = data[i++] & 0xFF;
- return (b3 << 24) + (b2 << 16) + (b1 << 8) + (b0 << 0);
- }
-
- /**
- * Get a BE int value from the beginning of a byte array
- *
- *@param data the byte array
- *@return the int (32-bit) value
- */
- public static int getIntBE(byte[] data) {
- return getIntBE(data, 0);
- }
- /**
- * Get a BE int value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the int (32-bit) value
- */
- public static int getIntBE(byte[] data, int offset) {
- int i=offset;
- int b0 = data[i++] & 0xFF;
- int b1 = data[i++] & 0xFF;
- int b2 = data[i++] & 0xFF;
- int b3 = data[i++] & 0xFF;
- return (b0 << 24) + (b1 << 16) + (b2 << 8) + (b3 << 0);
- }
-
- /**
- * Get a LE unsigned int value from a byte array
- *
- *@param data the byte array
- *@return the unsigned int (32-bit) value in a long
- */
- public static long getUIntLE(byte[] data) {
- return getUIntLE(data,0);
- }
- /**
- * Get a LE unsigned int value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the unsigned int (32-bit) value in a long
- */
- public static long getUIntLE(byte[] data, int offset) {
- long retNum = getIntLE(data, offset);
- return retNum & 0x00FFFFFFFFl;
- }
-
- /**
- * Get a BE unsigned int value from a byte array
- *
- *@param data the byte array
- *@return the unsigned int (32-bit) value in a long
- */
- public static long getUIntBE(byte[] data) {
- return getUIntBE(data,0);
- }
- /**
- * Get a BE unsigned int value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the unsigned int (32-bit) value in a long
- */
- public static long getUIntBE(byte[] data, int offset) {
- long retNum = getIntBE(data, offset);
- return retNum & 0x00FFFFFFFFl;
- }
-
- /**
- * Get a LE long value from a byte array
- *
- *@param data the byte array
- *@param offset a starting offset into the byte array
- *@return the long (64-bit) value
- */
- public static long getLongLE(byte[] data, int offset) {
- long result = 0;
-
- for (int j = offset + LONG_SIZE - 1; j >= offset; j--) {
- result <<= 8;
- result |= 0xff & data[j];
- }
- return result;
- }
- private static final int LONG_SIZE = 8;
-
-
- /**
- * Convert an 'unsigned' byte to an integer. ie, don't carry across the
- * sign.
- *
- * @param b Description of the Parameter
- * @return Description of the Return Value
- */
- public static int ubyteToInt(byte b) {
- return b & 0xFF;
- }
-
- /**
- * get the unsigned value of a byte.
- *
- * @param data
- * the byte array.
- * @param offset
- * a starting offset into the byte array.
- * @return the unsigned value of the byte as a 16 bit short
- */
- public static short getUByte( byte[] data, int offset )
- {
- return (short) ( data[offset] & 0xFF );
- }
-
-
- public static class BufferUnderrunException extends TikaException {
- private static final long serialVersionUID = 8358288231138076276L;
- public BufferUnderrunException() {
- super("Insufficient data left in stream for required read");
- }
- }
+ /**
+ * Get a LE short value from an InputStream
+ *
+ * @param stream the InputStream from which the short is to be read
+ * @return the short (16-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static short readShortLE(InputStream stream) throws IOException, BufferUnderrunException {
+ return (short) readUShortLE(stream);
+ }
+
+ /**
+ * Get a BE short value from an InputStream
+ *
+ * @param stream the InputStream from which the short is to be read
+ * @return the short (16-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static short readShortBE(InputStream stream) throws IOException, BufferUnderrunException {
+ return (short) readUShortBE(stream);
+ }
+
+ public static int readUShortLE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ if ((ch1 | ch2) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return (ch2 << 8) + (ch1 << 0);
+ }
+
+ public static int readUShortBE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ if ((ch1 | ch2) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return (ch1 << 8) + (ch2 << 0);
+ }
+
+
+ /**
+ * Get a LE unsigned int value from an InputStream
+ *
+ * @param stream the InputStream from which the int is to be read
+ * @return the int (32-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static long readUIntLE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return ((ch4 << 24) + (ch3 << 16) + (ch2 << 8) + (ch1 << 0)) & 0x00FFFFFFFFl;
+ }
+
+ /**
+ * Get a LE int value from an InputStream
+ *
+ * @param stream the InputStream from which the int is to be read
+ * @return the int (32-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static int readIntLE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return (ch4 << 24) + (ch3 << 16) + (ch2 << 8) + (ch1 << 0);
+ }
+
+ /**
+ * Get a BE int value from an InputStream
+ *
+ * @param stream the InputStream from which the int is to be read
+ * @return the int (32-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static int readIntBE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4) < 0) {
+ throw new BufferUnderrunException();
+ }
+ return (ch1 << 24) + (ch2 << 16) + (ch3 << 8) + (ch4 << 0);
+ }
+
+ /**
+ * Get a LE long value from an InputStream
+ *
+ * @param stream the InputStream from which the long is to be read
+ * @return the long (64-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static long readLongLE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ int ch5 = stream.read();
+ int ch6 = stream.read();
+ int ch7 = stream.read();
+ int ch8 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4 | ch5 | ch6 | ch7 | ch8) < 0) {
+ throw new BufferUnderrunException();
+ }
+
+ return
+ ((long) ch8 << 56) +
+ ((long) ch7 << 48) +
+ ((long) ch6 << 40) +
+ ((long) ch5 << 32) +
+ ((long) ch4 << 24) + // cast to long to preserve bit 31 (sign bit for ints)
+ (ch3 << 16) +
+ (ch2 << 8) +
+ (ch1 << 0);
+ }
+
+ /**
+ * Get a NE long value from an InputStream
+ *
+ * @param stream the InputStream from which the long is to be read
+ * @return the long (64-bit) value
+ * @throws IOException will be propagated back to the caller
+ * @throws BufferUnderrunException if the stream cannot provide enough bytes
+ */
+ public static long readLongBE(InputStream stream) throws IOException, BufferUnderrunException {
+ int ch1 = stream.read();
+ int ch2 = stream.read();
+ int ch3 = stream.read();
+ int ch4 = stream.read();
+ int ch5 = stream.read();
+ int ch6 = stream.read();
+ int ch7 = stream.read();
+ int ch8 = stream.read();
+ if ((ch1 | ch2 | ch3 | ch4 | ch5 | ch6 | ch7 | ch8) < 0) {
+ throw new BufferUnderrunException();
+ }
+
+ return
+ ((long) ch1 << 56) +
+ ((long) ch2 << 48) +
+ ((long) ch3 << 40) +
+ ((long) ch4 << 32) +
+ ((long) ch5 << 24) + // cast to long to preserve bit 31 (sign bit for ints)
+ (ch6 << 16) +
+ (ch7 << 8) +
+ (ch8 << 0);
+ }
+
+ /**
+ * Gets the integer value that is stored in UTF-8 like fashion, in Big Endian
+ * but with the high bit on each number indicating if it continues or not
+ */
+ public static long readUE7(InputStream stream) throws IOException {
+ int i;
+ long v = 0;
+ while ((i = stream.read()) >= 0) {
+ v = v << 7;
+ if ((i & 128) == 128) {
+ // Continues
+ v += (i & 127);
+ } else {
+ // Last value
+ v += i;
+ break;
+ }
+ }
+ return v;
+ }
+
+
+ /**
+ * Get a LE short value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the short (16-bit) value
+ */
+ public static short getShortLE(byte[] data) {
+ return getShortLE(data, 0);
+ }
+
+ /**
+ * Get a LE short value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the short (16-bit) value
+ */
+ public static short getShortLE(byte[] data, int offset) {
+ return (short) getUShortLE(data, offset);
+ }
+
+ /**
+ * Get a LE unsigned short value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the unsigned short (16-bit) value in an int
+ */
+ public static int getUShortLE(byte[] data) {
+ return getUShortLE(data, 0);
+ }
+
+ /**
+ * Get a LE unsigned short value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the unsigned short (16-bit) value in an integer
+ */
+ public static int getUShortLE(byte[] data, int offset) {
+ int b0 = data[offset] & 0xFF;
+ int b1 = data[offset + 1] & 0xFF;
+ return (b1 << 8) + (b0 << 0);
+ }
+
+ /**
+ * Get a BE short value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the short (16-bit) value
+ */
+ public static short getShortBE(byte[] data) {
+ return getShortBE(data, 0);
+ }
+
+ /**
+ * Get a BE short value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the short (16-bit) value
+ */
+ public static short getShortBE(byte[] data, int offset) {
+ return (short) getUShortBE(data, offset);
+ }
+
+ /**
+ * Get a BE unsigned short value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the unsigned short (16-bit) value in an int
+ */
+ public static int getUShortBE(byte[] data) {
+ return getUShortBE(data, 0);
+ }
+
+ /**
+ * Get a BE unsigned short value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the unsigned short (16-bit) value in an integer
+ */
+ public static int getUShortBE(byte[] data, int offset) {
+ int b0 = data[offset] & 0xFF;
+ int b1 = data[offset + 1] & 0xFF;
+ return (b0 << 8) + (b1 << 0);
+ }
+
+ /**
+ * Get a LE int value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the int (32-bit) value
+ */
+ public static int getIntLE(byte[] data) {
+ return getIntLE(data, 0);
+ }
+
+ /**
+ * Get a LE int value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the int (32-bit) value
+ */
+ public static int getIntLE(byte[] data, int offset) {
+ int i = offset;
+ int b0 = data[i++] & 0xFF;
+ int b1 = data[i++] & 0xFF;
+ int b2 = data[i++] & 0xFF;
+ int b3 = data[i++] & 0xFF;
+ return (b3 << 24) + (b2 << 16) + (b1 << 8) + (b0 << 0);
+ }
+
+ /**
+ * Get a BE int value from the beginning of a byte array
+ *
+ * @param data the byte array
+ * @return the int (32-bit) value
+ */
+ public static int getIntBE(byte[] data) {
+ return getIntBE(data, 0);
+ }
+
+ /**
+ * Get a BE int value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the int (32-bit) value
+ */
+ public static int getIntBE(byte[] data, int offset) {
+ int i = offset;
+ int b0 = data[i++] & 0xFF;
+ int b1 = data[i++] & 0xFF;
+ int b2 = data[i++] & 0xFF;
+ int b3 = data[i++] & 0xFF;
+ return (b0 << 24) + (b1 << 16) + (b2 << 8) + (b3 << 0);
+ }
+
+ /**
+ * Get a LE unsigned int value from a byte array
+ *
+ * @param data the byte array
+ * @return the unsigned int (32-bit) value in a long
+ */
+ public static long getUIntLE(byte[] data) {
+ return getUIntLE(data, 0);
+ }
+
+ /**
+ * Get a LE unsigned int value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the unsigned int (32-bit) value in a long
+ */
+ public static long getUIntLE(byte[] data, int offset) {
+ long retNum = getIntLE(data, offset);
+ return retNum & 0x00FFFFFFFFl;
+ }
+
+ /**
+ * Get a BE unsigned int value from a byte array
+ *
+ * @param data the byte array
+ * @return the unsigned int (32-bit) value in a long
+ */
+ public static long getUIntBE(byte[] data) {
+ return getUIntBE(data, 0);
+ }
+
+ /**
+ * Get a BE unsigned int value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the unsigned int (32-bit) value in a long
+ */
+ public static long getUIntBE(byte[] data, int offset) {
+ long retNum = getIntBE(data, offset);
+ return retNum & 0x00FFFFFFFFl;
+ }
+
+ /**
+ * Get a LE long value from a byte array
+ *
+ * @param data the byte array
+ * @param offset a starting offset into the byte array
+ * @return the long (64-bit) value
+ */
+ public static long getLongLE(byte[] data, int offset) {
+ long result = 0;
+
+ for (int j = offset + LONG_SIZE - 1; j >= offset; j--) {
+ result <<= 8;
+ result |= 0xff & data[j];
+ }
+ return result;
+ }
+
+ private static final int LONG_SIZE = 8;
+
+
+ /**
+ * Convert an 'unsigned' byte to an integer. ie, don't carry across the
+ * sign.
+ *
+ * @param b Description of the Parameter
+ * @return Description of the Return Value
+ */
+ public static int ubyteToInt(byte b) {
+ return b & 0xFF;
+ }
+
+ /**
+ * get the unsigned value of a byte.
+ *
+ * @param data the byte array.
+ * @param offset a starting offset into the byte array.
+ * @return the unsigned value of the byte as a 16 bit short
+ */
+ public static short getUByte(byte[] data, int offset) {
+ return (short) (data[offset] & 0xFF);
+ }
+
+
+ public static class BufferUnderrunException extends TikaException {
+ private static final long serialVersionUID = 8358288231138076276L;
+
+ public BufferUnderrunException() {
+ super("Insufficient data left in stream for required read");
+ }
+ }
}