Repository: spark Updated Branches: refs/heads/branch-2.2 175b221bc -> 367a16118
[SPARK-23649][SQL] Skipping chars disallowed in UTF-8 The mapping of UTF-8 char's first byte to char's size doesn't cover whole range 0-255. It is defined only for 0-253: https://github.com/apache/spark/blob/master/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java#L60-L65 https://github.com/apache/spark/blob/master/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java#L190 If the first byte of a char is 253-255, IndexOutOfBoundsException is thrown. Besides of that values for 244-252 are not correct according to recent unicode standard for UTF-8: http://www.unicode.org/versions/Unicode10.0.0/UnicodeStandard-10.0.pdf As a consequence of the exception above, the length of input string in UTF-8 encoding cannot be calculated if the string contains chars started from 253 code. It is visible on user's side as for example crashing of schema inferring of csv file which contains such chars but the file can be read if the schema is specified explicitly or if the mode set to multiline. The proposed changes build correct mapping of first byte of UTF-8 char to its size (now it covers all cases) and skip disallowed chars (counts it as one octet). Added a test and a file with a char which is disallowed in UTF-8 - 0xFF. Author: Maxim Gekk <maxim.g...@databricks.com> Closes #20796 from MaxGekk/skip-wrong-utf8-chars. (cherry picked from commit 5e7bc2acef4a1e11d0d8056ef5c12cd5c8f220da) Signed-off-by: Wenchen Fan <wenc...@databricks.com> Project: http://git-wip-us.apache.org/repos/asf/spark/repo Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/367a1611 Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/367a1611 Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/367a1611 Branch: refs/heads/branch-2.2 Commit: 367a16118289e1c03507c14f966e8b1ebd688489 Parents: 175b221 Author: Maxim Gekk <maxim.g...@databricks.com> Authored: Tue Mar 20 10:34:56 2018 -0700 Committer: Wenchen Fan <wenc...@databricks.com> Committed: Tue Mar 20 10:37:29 2018 -0700 ---------------------------------------------------------------------- .../apache/spark/unsafe/types/UTF8String.java | 48 ++++++++++++++++---- .../spark/unsafe/types/UTF8StringSuite.java | 23 +++++++++- 2 files changed, 62 insertions(+), 9 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/spark/blob/367a1611/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java ---------------------------------------------------------------------- diff --git a/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java b/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java index 23636ca..a11e63c 100644 --- a/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java +++ b/common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java @@ -57,12 +57,43 @@ public final class UTF8String implements Comparable<UTF8String>, Externalizable, public Object getBaseObject() { return base; } public long getBaseOffset() { return offset; } - private static int[] bytesOfCodePointInUTF8 = {2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, - 4, 4, 4, 4, 4, 4, 4, 4, - 5, 5, 5, 5, - 6, 6}; + /** + * A char in UTF-8 encoding can take 1-4 bytes depending on the first byte which + * indicates the size of the char. See Unicode standard in page 126, Table 3-6: + * http://www.unicode.org/versions/Unicode10.0.0/UnicodeStandard-10.0.pdf + * + * Binary Hex Comments + * 0xxxxxxx 0x00..0x7F Only byte of a 1-byte character encoding + * 10xxxxxx 0x80..0xBF Continuation bytes (1-3 continuation bytes) + * 110xxxxx 0xC0..0xDF First byte of a 2-byte character encoding + * 1110xxxx 0xE0..0xEF First byte of a 3-byte character encoding + * 11110xxx 0xF0..0xF7 First byte of a 4-byte character encoding + * + * As a consequence of the well-formedness conditions specified in + * Table 3-7 (page 126), the following byte values are disallowed in UTF-8: + * C0âC1, F5âFF. + */ + private static byte[] bytesOfCodePointInUTF8 = { + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x00..0x0F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x10..0x1F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x20..0x2F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x30..0x3F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x40..0x4F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x50..0x5F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x60..0x6F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x70..0x7F + // Continuation bytes cannot appear as the first byte + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x80..0x8F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x90..0x9F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xA0..0xAF + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xB0..0xBF + 0, 0, // 0xC0..0xC1 - disallowed in UTF-8 + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xC2..0xCF + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xD0..0xDF + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // 0xE0..0xEF + 4, 4, 4, 4, 4, // 0xF0..0xF4 + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 // 0xF5..0xFF - disallowed in UTF-8 + }; private static boolean isLittleEndian = ByteOrder.nativeOrder() == ByteOrder.LITTLE_ENDIAN; @@ -186,8 +217,9 @@ public final class UTF8String implements Comparable<UTF8String>, Externalizable, * @param b The first byte of a code point */ private static int numBytesForFirstByte(final byte b) { - final int offset = (b & 0xFF) - 192; - return (offset >= 0) ? bytesOfCodePointInUTF8[offset] : 1; + final int offset = b & 0xFF; + byte numBytes = bytesOfCodePointInUTF8[offset]; + return (numBytes == 0) ? 1: numBytes; // Skip the first byte disallowed in UTF-8 } /** http://git-wip-us.apache.org/repos/asf/spark/blob/367a1611/common/unsafe/src/test/java/org/apache/spark/unsafe/types/UTF8StringSuite.java ---------------------------------------------------------------------- diff --git a/common/unsafe/src/test/java/org/apache/spark/unsafe/types/UTF8StringSuite.java b/common/unsafe/src/test/java/org/apache/spark/unsafe/types/UTF8StringSuite.java index c376371..33b9e11 100644 --- a/common/unsafe/src/test/java/org/apache/spark/unsafe/types/UTF8StringSuite.java +++ b/common/unsafe/src/test/java/org/apache/spark/unsafe/types/UTF8StringSuite.java @@ -58,8 +58,12 @@ public class UTF8StringSuite { @Test public void basicTest() { checkBasic("", 0); - checkBasic("hello", 5); + checkBasic("¡", 1); // 2 bytes char + checkBasic("кÑ", 2); // 2 * 2 bytes chars + checkBasic("hello", 5); // 5 * 1 byte chars checkBasic("大 å ä¸ ç", 7); + checkBasic("︽ï¹ï¼ ", 3); // 3 * 3 bytes chars + checkBasic("\uD83E\uDD19", 1); // 4 bytes char } @Test @@ -730,4 +734,21 @@ public class UTF8StringSuite { assertFalse(negativeInput, UTF8String.fromString(negativeInput).toLong(wrapper)); } } + + @Test + public void skipWrongFirstByte() { + int[] wrongFirstBytes = { + 0x80, 0x9F, 0xBF, // Skip Continuation bytes + 0xC0, 0xC2, // 0xC0..0xC1 - disallowed in UTF-8 + // 0xF5..0xFF - disallowed in UTF-8 + 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, + 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF + }; + byte[] c = new byte[1]; + + for (int i = 0; i < wrongFirstBytes.length; ++i) { + c[0] = (byte)wrongFirstBytes[i]; + assertEquals(fromBytes(c).numChars(), 1); + } + } } --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org For additional commands, e-mail: commits-h...@spark.apache.org