- Revision
- 267964
- Author
- achristen...@apple.com
- Date
- 2020-10-05 07:43:04 -0700 (Mon, 05 Oct 2020)
Log Message
When appending a Windows drive letter to a file URL, remove any path we may have copied from the base URL
https://bugs.webkit.org/show_bug.cgi?id=217284
Reviewed by Darin Adler.
LayoutTests/imported/w3c:
* web-platform-tests/url/a-element-expected.txt:
* web-platform-tests/url/a-element-xhtml-expected.txt:
* web-platform-tests/url/resources/urltestdata.json:
* web-platform-tests/url/url-constructor-expected.txt:
Source/WTF:
This fixes a web platform test and adds another to verify that we remove the entire path, not just the last segment like we do with ..
Edge has implemented this, and in order to have a platform-independent URL we should do it too.
* wtf/URLParser.cpp:
(WTF::URLParser::isWindowsDriveLetter):
(WTF::URLParser::appendWindowsDriveLetter):
(WTF::URLParser::parse):
Modified Paths
Diff
Modified: trunk/LayoutTests/imported/w3c/ChangeLog (267963 => 267964)
--- trunk/LayoutTests/imported/w3c/ChangeLog 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/LayoutTests/imported/w3c/ChangeLog 2020-10-05 14:43:04 UTC (rev 267964)
@@ -1,5 +1,17 @@
2020-10-05 Alex Christensen <achristen...@webkit.org>
+ When appending a Windows drive letter to a file URL, remove any path we may have copied from the base URL
+ https://bugs.webkit.org/show_bug.cgi?id=217284
+
+ Reviewed by Darin Adler.
+
+ * web-platform-tests/url/a-element-expected.txt:
+ * web-platform-tests/url/a-element-xhtml-expected.txt:
+ * web-platform-tests/url/resources/urltestdata.json:
+ * web-platform-tests/url/url-constructor-expected.txt:
+
+2020-10-05 Alex Christensen <achristen...@webkit.org>
+
Fix UTF-8 encoding in URL parsing
https://bugs.webkit.org/show_bug.cgi?id=217289
Modified: trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-expected.txt (267963 => 267964)
--- trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-expected.txt 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-expected.txt 2020-10-05 14:43:04 UTC (rev 267964)
@@ -447,7 +447,8 @@
PASS Parsing: <file://> against <file://ape/>
PASS Parsing: </rooibos> against <file://tea/>
PASS Parsing: </?chai> against <file://tea/>
-FAIL Parsing: <C|> against <file://host/dir/file> assert_equals: href expected "file://host/C:" but got "file://host/dir/C|"
+PASS Parsing: <C|> against <file://host/dir/file>
+PASS Parsing: <C|> against <file://host/D:/dir1/dir2/file>
PASS Parsing: <C|#> against <file://host/dir/file>
PASS Parsing: <C|?> against <file://host/dir/file>
PASS Parsing: <C|/> against <file://host/dir/file>
Modified: trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-xhtml-expected.txt (267963 => 267964)
--- trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-xhtml-expected.txt 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/LayoutTests/imported/w3c/web-platform-tests/url/a-element-xhtml-expected.txt 2020-10-05 14:43:04 UTC (rev 267964)
@@ -447,7 +447,8 @@
PASS Parsing: <file://> against <file://ape/>
PASS Parsing: </rooibos> against <file://tea/>
PASS Parsing: </?chai> against <file://tea/>
-FAIL Parsing: <C|> against <file://host/dir/file> assert_equals: href expected "file://host/C:" but got "file://host/dir/C|"
+PASS Parsing: <C|> against <file://host/dir/file>
+PASS Parsing: <C|> against <file://host/D:/dir1/dir2/file>
PASS Parsing: <C|#> against <file://host/dir/file>
PASS Parsing: <C|?> against <file://host/dir/file>
PASS Parsing: <C|/> against <file://host/dir/file>
Modified: trunk/LayoutTests/imported/w3c/web-platform-tests/url/resources/urltestdata.json (267963 => 267964)
--- trunk/LayoutTests/imported/w3c/web-platform-tests/url/resources/urltestdata.json 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/LayoutTests/imported/w3c/web-platform-tests/url/resources/urltestdata.json 2020-10-05 14:43:04 UTC (rev 267964)
@@ -5684,6 +5684,20 @@
"hash": ""
},
{
+ "input": "C|",
+ "base": "file://host/D:/dir1/dir2/file",
+ "href": "file://host/C:",
+ "protocol": "file:",
+ "username": "",
+ "password": "",
+ "host": "host",
+ "hostname": "host",
+ "port": "",
+ "pathname": "/C:",
+ "search": "",
+ "hash": ""
+ },
+ {
"input": "C|#",
"base": "file://host/dir/file",
"href": "file://host/C:#",
Modified: trunk/LayoutTests/imported/w3c/web-platform-tests/url/url-constructor-expected.txt (267963 => 267964)
--- trunk/LayoutTests/imported/w3c/web-platform-tests/url/url-constructor-expected.txt 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/LayoutTests/imported/w3c/web-platform-tests/url/url-constructor-expected.txt 2020-10-05 14:43:04 UTC (rev 267964)
@@ -448,7 +448,8 @@
PASS Parsing: <file://> against <file://ape/>
PASS Parsing: </rooibos> against <file://tea/>
PASS Parsing: </?chai> against <file://tea/>
-FAIL Parsing: <C|> against <file://host/dir/file> assert_equals: href expected "file://host/C:" but got "file://host/dir/C|"
+PASS Parsing: <C|> against <file://host/dir/file>
+PASS Parsing: <C|> against <file://host/D:/dir1/dir2/file>
PASS Parsing: <C|#> against <file://host/dir/file>
PASS Parsing: <C|?> against <file://host/dir/file>
PASS Parsing: <C|/> against <file://host/dir/file>
Modified: trunk/Source/WTF/ChangeLog (267963 => 267964)
--- trunk/Source/WTF/ChangeLog 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/Source/WTF/ChangeLog 2020-10-05 14:43:04 UTC (rev 267964)
@@ -1,5 +1,20 @@
2020-10-05 Alex Christensen <achristen...@webkit.org>
+ When appending a Windows drive letter to a file URL, remove any path we may have copied from the base URL
+ https://bugs.webkit.org/show_bug.cgi?id=217284
+
+ Reviewed by Darin Adler.
+
+ This fixes a web platform test and adds another to verify that we remove the entire path, not just the last segment like we do with ..
+ Edge has implemented this, and in order to have a platform-independent URL we should do it too.
+
+ * wtf/URLParser.cpp:
+ (WTF::URLParser::isWindowsDriveLetter):
+ (WTF::URLParser::appendWindowsDriveLetter):
+ (WTF::URLParser::parse):
+
+2020-10-05 Alex Christensen <achristen...@webkit.org>
+
Fix UTF-8 encoding in URL parsing
https://bugs.webkit.org/show_bug.cgi?id=217289
Modified: trunk/Source/WTF/wtf/URLParser.cpp (267963 => 267964)
--- trunk/Source/WTF/wtf/URLParser.cpp 2020-10-05 14:40:55 UTC (rev 267963)
+++ trunk/Source/WTF/wtf/URLParser.cpp 2020-10-05 14:43:04 UTC (rev 267964)
@@ -371,16 +371,16 @@
template<typename CharacterType>
ALWAYS_INLINE bool URLParser::isWindowsDriveLetter(CodePointIterator<CharacterType> iterator)
{
+ // https://url.spec.whatwg.org/#start-with-a-windows-drive-letter
if (iterator.atEnd() || !isASCIIAlpha(*iterator))
return false;
advance<CharacterType, ReportSyntaxViolation::No>(iterator);
if (iterator.atEnd())
return false;
- if (*iterator == ':')
- return true;
- if (UNLIKELY(*iterator == '|'))
- return true;
- return false;
+ if (*iterator != ':' && *iterator != '|')
+ return false;
+ advance<CharacterType, ReportSyntaxViolation::No>(iterator);
+ return iterator.atEnd() || *iterator == '/' || *iterator == '\\' || *iterator == '?' || *iterator == '#';
}
ALWAYS_INLINE void URLParser::appendToASCIIBuffer(UChar32 codePoint)
@@ -399,6 +399,12 @@
template<typename CharacterType>
void URLParser::appendWindowsDriveLetter(CodePointIterator<CharacterType>& iterator)
{
+ auto lengthWithOnlyOneSlashInPath = m_url.m_hostEnd + m_url.m_portLength + 1;
+ if (m_url.m_pathAfterLastSlash > lengthWithOnlyOneSlashInPath) {
+ syntaxViolation(iterator);
+ m_url.m_pathAfterLastSlash = lengthWithOnlyOneSlashInPath;
+ m_asciiBuffer.resize(lengthWithOnlyOneSlashInPath);
+ }
ASSERT(isWindowsDriveLetter(iterator));
appendToASCIIBuffer(*iterator);
advance(iterator);
@@ -1548,9 +1554,9 @@
m_url.m_portLength = 0;
}
m_url.m_pathAfterLastSlash = m_url.m_hostEnd + 1;
- if (isWindowsDriveLetter(c))
- appendWindowsDriveLetter(c);
}
+ if (isWindowsDriveLetter(c))
+ appendWindowsDriveLetter(c);
state = State::Path;
break;
}
