Changeset 270526 in webkit

Timestamp:

Dec 7, 2020 3:43:43 PM (3 years ago)

Author:

commit-queue@webkit.org

Message:

Parse content after # in data URLs with HLS mime types
​https://bugs.webkit.org/show_bug.cgi?id=219612
<rdar://problem/71039282>

Patch by Alex Christensen <​achristensen@webkit.org> on 2020-12-07
Reviewed by Darin Adler.

Source/WebCore:

r267995 made us conform to the behavior of Chrome and Firefox by removing fragments from data URLs before parsing.
While this is desirable, there is content in Safari-specific HLS data URLs that needs the content including and after #.
So, to fix this, wait until after we know the mime type and remove the fragment if it's not HLS.

Test: fast/url/data-url-mediatype.html

• loader/ResourceLoader.cpp:

(WebCore::ResourceLoader::loadDataURL):
(WebCore::shouldStripFragmentIdentifier): Deleted.

• platform/network/DataURLDecoder.cpp:

(WebCore::DataURLDecoder::shouldRemoveFragmentIdentifier):
(WebCore::DataURLDecoder::DecodeTask::DecodeTask):
(WebCore::DataURLDecoder::DecodeTask::process):
(WebCore::DataURLDecoder::createDecodeTask):

LayoutTests:

• fast/url/data-url-mediatype-expected.txt: Added.
• fast/url/data-url-mediatype.html: Added.

Location:

trunk

Files:

• LayoutTests/ChangeLog (modified) (1 diff)
• LayoutTests/fast/url/data-url-mediatype-expected.txt (added)
• LayoutTests/fast/url/data-url-mediatype.html (added)
• Source/WebCore/ChangeLog (modified) (1 diff)
• Source/WebCore/loader/ResourceLoader.cpp (modified) (3 diffs)
• Source/WebCore/platform/network/DataURLDecoder.cpp (modified) (5 diffs)

trunk/LayoutTests/ChangeLog

@@ +1 @@
+2020-12-07  Alex Christensen  <achristensen@webkit.org>
+
+        Parse content after # in data URLs with HLS mime types
+        https://bugs.webkit.org/show_bug.cgi?id=219612
+        <rdar://problem/71039282>
+
+        Reviewed by Darin Adler.
+
+        * fast/url/data-url-mediatype-expected.txt: Added.
+        * fast/url/data-url-mediatype.html: Added.
+
 2020-12-07  Lauro Moura  <lmoura@igalia.com>
 

trunk/Source/WebCore/ChangeLog

@@ +1 @@
+2020-12-07  Alex Christensen  <achristensen@webkit.org>
+
+        Parse content after # in data URLs with HLS mime types
+        https://bugs.webkit.org/show_bug.cgi?id=219612
+        <rdar://problem/71039282>
+
+        Reviewed by Darin Adler.
+
+        r267995 made us conform to the behavior of Chrome and Firefox by removing fragments from data URLs before parsing.
+        While this is desirable, there is content in Safari-specific HLS data URLs that needs the content including and after #.
+        So, to fix this, wait until after we know the mime type and remove the fragment if it's not HLS.
+
+        Test: fast/url/data-url-mediatype.html
+
+        * loader/ResourceLoader.cpp:
+        (WebCore::ResourceLoader::loadDataURL):
+        (WebCore::shouldStripFragmentIdentifier): Deleted.
+        * platform/network/DataURLDecoder.cpp:
+        (WebCore::DataURLDecoder::shouldRemoveFragmentIdentifier):
+        (WebCore::DataURLDecoder::DecodeTask::DecodeTask):
+        (WebCore::DataURLDecoder::DecodeTask::process):
+        (WebCore::DataURLDecoder::createDecodeTask):
+
 2020-12-07  Zalan Bujtas  <zalan@apple.com>
 

trunk/Source/WebCore/loader/ResourceLoader.cpp

@@ -55 +55 @@
 #include <wtf/Ref.h>
 
-#if PLATFORM(COCOA)
-#include "VersionChecks.h"
-#endif
-
 #if ENABLE(CONTENT_EXTENSIONS)
 #include "UserContentController.h"
@@ -253 +249 @@
 }
 
-static bool shouldStripFragmentIdentifier()
-{
-#if PLATFORM(COCOA)
-    return linkedOnOrAfter(SDKVersion::FirstWithDataURLFragmentRemoval);
-#else
-    return true;
-#endif
-}
-
 void ResourceLoader::loadDataURL()
 {
@@ -275 +262 @@
     if (m_request.requester() == ResourceRequest::Requester::Fetch)
         mode = DataURLDecoder::Mode::ForgivingBase64;
-    if (shouldStripFragmentIdentifier())
-        url.removeFragmentIdentifier();
     DataURLDecoder::decode(url, scheduleContext, mode, [this, protectedThis = makeRef(*this), url](auto decodeResult) mutable {
         if (this->reachedTerminalState())

trunk/Source/WebCore/platform/network/DataURLDecoder.cpp

@@ -39 +39 @@
 #include <wtf/text/Base64.h>
 
+#if PLATFORM(COCOA)
+#include "VersionChecks.h"
+#endif
+
 namespace WebCore {
 namespace DataURLDecoder {
+
+static bool shouldRemoveFragmentIdentifier(const String& mediaType)
+{
+#if PLATFORM(COCOA)
+    if (!linkedOnOrAfter(SDKVersion::FirstWithDataURLFragmentRemoval))
+        return false;
+
+    // HLS uses # in the middle of the manifests.
+    return !equalLettersIgnoringASCIICase(mediaType, "video/mpegurl")
+        && !equalLettersIgnoringASCIICase(mediaType, "audio/mpegurl")
+        && !equalLettersIgnoringASCIICase(mediaType, "application/x-mpegurl")
+        && !equalLettersIgnoringASCIICase(mediaType, "vnd.apple.mpegurl");
+#else
+    return false;
+#endif
+}
 
 static WorkQueue& decodeQueue()
@@ -58 +78 @@
     WTF_MAKE_FAST_ALLOCATED;
 public:
-    DecodeTask(const String& urlString, const ScheduleContext& scheduleContext, DecodeCompletionHandler&& completionHandler)
-        : urlString(urlString.isolatedCopy())
+    DecodeTask(const URL& url, const ScheduleContext& scheduleContext, DecodeCompletionHandler&& completionHandler)
+        : url(url.isolatedCopy())
         , scheduleContext(scheduleContext)
         , completionHandler(WTFMove(completionHandler))
@@ -72 +92 @@
         //  mediatype := [<mimetype>][;charset=<charsettype>]
 
-        if (urlString.find(',') == notFound)
+        const char dataString[] = "data:";
+        ASSERT(url.string().startsWith(dataString));
+
+        size_t headerStart = strlen(dataString);
+        size_t headerEnd = url.string().find(',', headerStart);
+        if (headerEnd == notFound)
             return false;
-
-        const char dataString[] = "data:";
-        ASSERT(urlString.startsWith(dataString));
-
-        size_t headerStart = strlen(dataString);
-        size_t headerEnd = urlString.find(',', headerStart);
+        if (size_t fragmentInHeader = url.string().reverseFind('#', headerEnd); fragmentInHeader != notFound)
+            return false;
         size_t encodedDataStart = headerEnd == notFound ? headerEnd : headerEnd + 1;
 
-        encodedData = StringView(urlString).substring(encodedDataStart);
-        auto header = StringView(urlString).substring(headerStart, headerEnd - headerStart);
+        auto header = StringView(url.string()).substring(headerStart, headerEnd - headerStart);
         
         // There might one or two semicolons in the header, find the last one.
@@ -102 +122 @@
             mediaType.insert("text/plain", 0);
 
+        if (shouldRemoveFragmentIdentifier(mediaType))
+            url.removeFragmentIdentifier();
+
+        encodedData = StringView(url.string()).substring(encodedDataStart);
+
         result = parseMediaType(mediaType);
         return true;
     }
 
-    const String urlString;
+    URL url;
     StringView encodedData;
     bool isBase64 { false };
@@ -118 +143 @@
 {
     return makeUnique<DecodeTask>(
-        url.string(),
+        url,
         scheduleContext,
         WTFMove(completionHandler)