Diff
Modified: branches/safari-605-branch/LayoutTests/imported/w3c/ChangeLog (228835 => 228836)
--- branches/safari-605-branch/LayoutTests/imported/w3c/ChangeLog 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/LayoutTests/imported/w3c/ChangeLog 2018-02-20 22:30:01 UTC (rev 228836)
@@ -1,3 +1,20 @@
+2018-02-20 Jason Marcell <jmarc...@apple.com>
+
+ Cherry-pick r228594. rdar://problem/37697671
+
+ 2018-02-16 Daniel Bates <daba...@apple.com>
+
+ Remove dead code for UTF-32
+ https://bugs.webkit.org/show_bug.cgi?id=182900
+ <rdar://problem/37238717>
+
+ Reviewed by Alexey Proskuryakov.
+
+ Update expected results now that all sub-tests pass.
+
+ * web-platform-tests/encoding/unsupported-encodings-expected.txt:
+ * web-platform-tests/encoding/utf-32-expected.txt:
+
2018-02-12 Jason Marcell <jmarc...@apple.com>
Cherry-pick r228326. rdar://problem/37460483
Modified: branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/unsupported-encodings-expected.txt (228835 => 228836)
--- branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/unsupported-encodings-expected.txt 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/unsupported-encodings-expected.txt 2018-02-20 22:30:01 UTC (rev 228836)
@@ -1,16 +1,16 @@
PASS UTF-7 should not be supported
PASS utf-7 should not be supported
-FAIL UTF-32 with BOM should decode as UTF-16LE assert_equals: Decoding with UTF-32 expected "U+0000/U+0041/U+0000/U+0042/U+0000" but got "U+0041/U+0000/U+0000/U+0000/U+0042/U+0000/U+0000/U+0000"
+PASS UTF-32 with BOM should decode as UTF-16LE
PASS UTF-32 with no BOM should decode as windows-1252
-FAIL utf-32 with BOM should decode as UTF-16LE assert_equals: Decoding with utf-32 expected "U+0000/U+0041/U+0000/U+0042/U+0000" but got "U+0041/U+0000/U+0000/U+0000/U+0042/U+0000/U+0000/U+0000"
+PASS utf-32 with BOM should decode as UTF-16LE
PASS utf-32 with no BOM should decode as windows-1252
-FAIL UTF-32LE with BOM should decode as UTF-16LE assert_equals: Decoding with UTF-32LE expected "U+0000/U+0041/U+0000/U+0042/U+0000" but got "U+0041/U+0000/U+0000/U+0000/U+0042/U+0000/U+0000/U+0000"
+PASS UTF-32LE with BOM should decode as UTF-16LE
PASS UTF-32LE with no BOM should decode as windows-1252
-FAIL utf-32le with BOM should decode as UTF-16LE assert_equals: Decoding with utf-32le expected "U+0000/U+0041/U+0000/U+0042/U+0000" but got "U+0041/U+0000/U+0000/U+0000/U+0042/U+0000/U+0000/U+0000"
+PASS utf-32le with BOM should decode as UTF-16LE
PASS utf-32le with no BOM should decode as windows-1252
PASS UTF-32be with no BOM should decode as windows-1252
-FAIL UTF-32be with BOM should decode as windows-1252 assert_equals: Decoding with UTF-32be expected "U+0000/U+0000/U+00FE/U+00FF/U+0000/U+0000/U+0000/U+0041/U+0000/U+0000/U+0000/U+0042" but got "U+0000/U+0000/U+0000/U+0041/U+0000/U+0000/U+0000/U+0042"
+PASS UTF-32be with BOM should decode as windows-1252
PASS utf-32be with no BOM should decode as windows-1252
-FAIL utf-32be with BOM should decode as windows-1252 assert_equals: Decoding with utf-32be expected "U+0000/U+0000/U+00FE/U+00FF/U+0000/U+0000/U+0000/U+0041/U+0000/U+0000/U+0000/U+0042" but got "U+0000/U+0000/U+0000/U+0041/U+0000/U+0000/U+0000/U+0042"
+PASS utf-32be with BOM should decode as windows-1252
Modified: branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/utf-32-expected.txt (228835 => 228836)
--- branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/utf-32-expected.txt 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/LayoutTests/imported/w3c/web-platform-tests/encoding/utf-32-expected.txt 2018-02-20 22:30:01 UTC (rev 228836)
@@ -3,8 +3,8 @@
PASS Expect resources/utf-32-big-endian-bom.xml to parse as windows-1252
PASS Expect resources/utf-32-big-endian-nobom.html to parse as windows-1252
PASS Expect resources/utf-32-big-endian-nobom.xml to parse as windows-1252
-FAIL Expect resources/utf-32-little-endian-bom.html to parse as UTF-16LE assert_equals: expected "UTF-16LE" but got "windows-1252"
-FAIL Expect resources/utf-32-little-endian-bom.xml to parse as UTF-16LE assert_equals: expected "UTF-16LE" but got "windows-1252"
+PASS Expect resources/utf-32-little-endian-bom.html to parse as UTF-16LE
+PASS Expect resources/utf-32-little-endian-bom.xml to parse as UTF-16LE
PASS Expect resources/utf-32-little-endian-nobom.html to parse as windows-1252
PASS Expect resources/utf-32-little-endian-nobom.xml to parse as windows-1252
Modified: branches/safari-605-branch/Source/WebCore/ChangeLog (228835 => 228836)
--- branches/safari-605-branch/Source/WebCore/ChangeLog 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/Source/WebCore/ChangeLog 2018-02-20 22:30:01 UTC (rev 228836)
@@ -1,5 +1,31 @@
2018-02-20 Jason Marcell <jmarc...@apple.com>
+ Cherry-pick r228594. rdar://problem/37697671
+
+ 2018-02-16 Daniel Bates <daba...@apple.com>
+
+ Remove dead code for UTF-32
+ https://bugs.webkit.org/show_bug.cgi?id=182900
+ <rdar://problem/37238717>
+
+ Reviewed by Alexey Proskuryakov.
+
+ The UTF-32 encodings were removed from the text codec registry in r224747. So,
+ we no longer need code to encode or decode these variants.
+
+ * dom/TextDecoder.cpp:
+ (WebCore::codeUnitByteSize):
+ * loader/TextResourceDecoder.cpp:
+ (WebCore::TextResourceDecoder::checkForBOM):
+ (WebCore::TextResourceDecoder::checkForHeadCharset):
+ * platform/text/TextEncoding.cpp:
+ (WebCore::TextEncoding::isNonByteBasedEncoding const):
+ (WebCore::UTF32BigEndianEncoding): Deleted.
+ (WebCore::UTF32LittleEndianEncoding): Deleted.
+ * platform/text/TextEncoding.h:
+
+2018-02-20 Jason Marcell <jmarc...@apple.com>
+
Cherry-pick r228586. rdar://problem/37697673
2018-02-16 Chris Dumez <cdu...@apple.com>
Modified: branches/safari-605-branch/Source/WebCore/dom/TextDecoder.cpp (228835 => 228836)
--- branches/safari-605-branch/Source/WebCore/dom/TextDecoder.cpp 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/Source/WebCore/dom/TextDecoder.cpp 2018-02-20 22:30:01 UTC (rev 228836)
@@ -87,11 +87,7 @@
static size_t codeUnitByteSize(const TextEncoding& encoding)
{
- if (encoding.isByteBasedEncoding())
- return 1;
- if (encoding == UTF32BigEndianEncoding() || encoding == UTF32LittleEndianEncoding())
- return 4;
- return 2;
+ return encoding.isByteBasedEncoding() ? 1 : 2;
}
ExceptionOr<String> TextDecoder::decode(std::optional<BufferSource::VariantType> input, DecodeOptions options)
Modified: branches/safari-605-branch/Source/WebCore/loader/TextResourceDecoder.cpp (228835 => 228836)
--- branches/safari-605-branch/Source/WebCore/loader/TextResourceDecoder.cpp 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/Source/WebCore/loader/TextResourceDecoder.cpp 2018-02-20 22:30:01 UTC (rev 228836)
@@ -401,8 +401,10 @@
size_t TextResourceDecoder::checkForBOM(const char* data, size_t len)
{
- // Check for UTF-16/32 or UTF-8 BOM mark at the beginning, which is a sure sign of a Unicode encoding.
+ // Check for UTF-16 or UTF-8 BOM mark at the beginning, which is a sure sign of a Unicode encoding.
// We let it override even a user-chosen encoding.
+ const size_t maximumBOMLength = 3;
+
ASSERT(!m_checkedForBOM);
size_t lengthOfBOM = 0;
@@ -416,31 +418,26 @@
unsigned char c1 = buf1Len ? (--buf1Len, *buf1++) : buf2Len ? (--buf2Len, *buf2++) : 0;
unsigned char c2 = buf1Len ? (--buf1Len, *buf1++) : buf2Len ? (--buf2Len, *buf2++) : 0;
unsigned char c3 = buf1Len ? (--buf1Len, *buf1++) : buf2Len ? (--buf2Len, *buf2++) : 0;
- unsigned char c4 = buf2Len ? (--buf2Len, *buf2++) : 0;
// Check for the BOM.
if (c1 == 0xFF && c2 == 0xFE) {
- if (c3 != 0 || c4 != 0) {
- setEncoding(UTF16LittleEndianEncoding(), AutoDetectedEncoding);
- lengthOfBOM = 2;
- } else {
- setEncoding(UTF32LittleEndianEncoding(), AutoDetectedEncoding);
- lengthOfBOM = 4;
- }
+ ASSERT(UTF16LittleEndianEncoding().isValid());
+ setEncoding(UTF16LittleEndianEncoding(), AutoDetectedEncoding);
+ lengthOfBOM = 2;
+ } else if (c1 == 0xFE && c2 == 0xFF) {
+ ASSERT(UTF16BigEndianEncoding().isValid());
+ setEncoding(UTF16BigEndianEncoding(), AutoDetectedEncoding);
+ lengthOfBOM = 2;
} else if (c1 == 0xEF && c2 == 0xBB && c3 == 0xBF) {
+ ASSERT(UTF8Encoding().isValid());
setEncoding(UTF8Encoding(), AutoDetectedEncoding);
lengthOfBOM = 3;
- } else if (c1 == 0xFE && c2 == 0xFF) {
- setEncoding(UTF16BigEndianEncoding(), AutoDetectedEncoding);
- lengthOfBOM = 2;
- } else if (c1 == 0 && c2 == 0 && c3 == 0xFE && c4 == 0xFF) {
- setEncoding(UTF32BigEndianEncoding(), AutoDetectedEncoding);
- lengthOfBOM = 4;
}
- if (lengthOfBOM || bufferLength + len >= 4)
+ if (lengthOfBOM || bufferLength + len >= maximumBOMLength)
m_checkedForBOM = true;
+ ASSERT(lengthOfBOM <= maximumBOMLength);
return lengthOfBOM;
}
@@ -533,12 +530,6 @@
} else if (bytesEqual(ptr, 0, '<', 0, '?', 0, 'x')) {
setEncoding(UTF16BigEndianEncoding(), AutoDetectedEncoding);
return true;
- } else if (bytesEqual(ptr, '<', 0, 0, 0, '?', 0, 0, 0)) {
- setEncoding(UTF32LittleEndianEncoding(), AutoDetectedEncoding);
- return true;
- } else if (bytesEqual(ptr, 0, 0, 0, '<', 0, 0, 0, '?')) {
- setEncoding(UTF32BigEndianEncoding(), AutoDetectedEncoding);
- return true;
}
// The HTTP-EQUIV meta has no effect on XHTML.
Modified: branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.cpp (228835 => 228836)
--- branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.cpp 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.cpp 2018-02-20 22:30:01 UTC (rev 228836)
@@ -145,15 +145,7 @@
bool TextEncoding::isNonByteBasedEncoding() const
{
- if (noExtendedTextEncodingNameUsed()) {
- return *this == UTF16LittleEndianEncoding()
- || *this == UTF16BigEndianEncoding();
- }
-
- return *this == UTF16LittleEndianEncoding()
- || *this == UTF16BigEndianEncoding()
- || *this == UTF32BigEndianEncoding()
- || *this == UTF32LittleEndianEncoding();
+ return *this == UTF16LittleEndianEncoding() || *this == UTF16BigEndianEncoding();
}
bool TextEncoding::isUTF7Encoding() const
@@ -207,18 +199,6 @@
return globalUTF16LittleEndianEncoding;
}
-const TextEncoding& UTF32BigEndianEncoding()
-{
- static TextEncoding globalUTF32BigEndianEncoding("UTF-32BE");
- return globalUTF32BigEndianEncoding;
-}
-
-const TextEncoding& UTF32LittleEndianEncoding()
-{
- static TextEncoding globalUTF32LittleEndianEncoding("UTF-32LE");
- return globalUTF32LittleEndianEncoding;
-}
-
const TextEncoding& UTF8Encoding()
{
static TextEncoding globalUTF8Encoding("UTF-8");
Modified: branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.h (228835 => 228836)
--- branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.h 2018-02-20 22:29:56 UTC (rev 228835)
+++ branches/safari-605-branch/Source/WebCore/platform/text/TextEncoding.h 2018-02-20 22:30:01 UTC (rev 228836)
@@ -67,8 +67,6 @@
const TextEncoding& Latin1Encoding();
const TextEncoding& UTF16BigEndianEncoding();
const TextEncoding& UTF16LittleEndianEncoding();
-const TextEncoding& UTF32BigEndianEncoding();
-const TextEncoding& UTF32LittleEndianEncoding();
WEBCORE_EXPORT const TextEncoding& UTF8Encoding();
WEBCORE_EXPORT const TextEncoding& WindowsLatin1Encoding();
