| File: | Volumes/Data/worker/Apple-iOS-26-Safer-CPP-Checks-EWS/build/Source/WebCore/loader/archive/cf/LegacyWebArchive.cpp |
| Warning: | line 776, column 31 Local variable 'cachedResource' is uncounted and unsafe |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
| 1 | /* |
| 2 | * Copyright (C) 2008-2025 Apple Inc. All rights reserved. |
| 3 | * |
| 4 | * Redistribution and use in source and binary forms, with or without |
| 5 | * modification, are permitted provided that the following conditions |
| 6 | * are met: |
| 7 | * |
| 8 | * 1. Redistributions of source code must retain the above copyright |
| 9 | * notice, this list of conditions and the following disclaimer. |
| 10 | * 2. Redistributions in binary form must reproduce the above copyright |
| 11 | * notice, this list of conditions and the following disclaimer in the |
| 12 | * documentation and/or other materials provided with the distribution. |
| 13 | * 3. Neither the name of Apple Inc. ("Apple") nor the names of |
| 14 | * its contributors may be used to endorse or promote products derived |
| 15 | * from this software without specific prior written permission. |
| 16 | * |
| 17 | * THIS SOFTWARE IS PROVIDED BY APPLE AND ITS CONTRIBUTORS "AS IS" AND ANY |
| 18 | * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED |
| 19 | * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
| 20 | * DISCLAIMED. IN NO EVENT SHALL APPLE OR ITS CONTRIBUTORS BE LIABLE FOR ANY |
| 21 | * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES |
| 22 | * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
| 23 | * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND |
| 24 | * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 25 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF |
| 26 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 27 | */ |
| 28 | |
| 29 | #include "config.h" |
| 30 | #include "LegacyWebArchive.h" |
| 31 | |
| 32 | #include "BoundaryPointInlines.h" |
| 33 | #include "CSSImportRule.h" |
| 34 | #include "CSSSerializationContext.h" |
| 35 | #include "CachedResource.h" |
| 36 | #include "DeprecatedGlobalSettings.h" |
| 37 | #include "DocumentLoader.h" |
| 38 | #include "DocumentPage.h" |
| 39 | #include "Editor.h" |
| 40 | #include "EditorClient.h" |
| 41 | #include "FrameDestructionObserverInlines.h" |
| 42 | #include "FrameLoader.h" |
| 43 | #include "FrameSelection.h" |
| 44 | #include "FrameTree.h" |
| 45 | #include "HTMLAttachmentElement.h" |
| 46 | #include "HTMLFrameElement.h" |
| 47 | #include "HTMLFrameOwnerElement.h" |
| 48 | #include "HTMLIFrameElement.h" |
| 49 | #include "HTMLNames.h" |
| 50 | #include "HTMLObjectElement.h" |
| 51 | #include "Image.h" |
| 52 | #include "LocalFrame.h" |
| 53 | #include "LocalFrameInlines.h" |
| 54 | #include "Logging.h" |
| 55 | #include "MIMETypeRegistry.h" |
| 56 | #include "MemoryCache.h" |
| 57 | #include "SerializedAttachmentData.h" |
| 58 | #include "Settings.h" |
| 59 | #include "SharedBuffer.h" |
| 60 | #include "StyleSheet.h" |
| 61 | #include "StyleSheetList.h" |
| 62 | #include "markup.h" |
| 63 | #include <algorithm> |
| 64 | #include <wtf/ListHashSet.h> |
| 65 | #include <wtf/RetainPtr.h> |
| 66 | #include <wtf/URLHash.h> |
| 67 | #include <wtf/URLParser.h> |
| 68 | #include <wtf/text/CString.h> |
| 69 | #include <wtf/text/MakeString.h> |
| 70 | #include <wtf/text/StringBuilder.h> |
| 71 | |
| 72 | namespace WebCore { |
| 73 | |
| 74 | static constexpr unsigned maxFileNameSizeInBytes = 255; |
| 75 | static constexpr char defaultFileName[] = "file"; |
| 76 | static const CFStringRef LegacyWebArchiveMainResourceKey = CFSTR("WebMainResource")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebMainResource" "")); |
| 77 | static const CFStringRef LegacyWebArchiveSubresourcesKey = CFSTR("WebSubresources")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebSubresources" "")); |
| 78 | static const CFStringRef LegacyWebArchiveSubframeArchivesKey = CFSTR("WebSubframeArchives")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebSubframeArchives" "")); |
| 79 | static const CFStringRef LegacyWebArchiveResourceDataKey = CFSTR("WebResourceData")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceData" "")); |
| 80 | static const CFStringRef LegacyWebArchiveResourceFrameNameKey = CFSTR("WebResourceFrameName")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceFrameName" "")); |
| 81 | static const CFStringRef LegacyWebArchiveResourceMIMETypeKey = CFSTR("WebResourceMIMEType")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceMIMEType" "")); |
| 82 | static const CFStringRef LegacyWebArchiveResourceURLKey = CFSTR("WebResourceURL")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceURL" "")); |
| 83 | static const CFStringRef LegacyWebArchiveResourceFilePathKey = CFSTR("WebResourceFilePath")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceFilePath" "")); |
| 84 | static const CFStringRef LegacyWebArchiveResourceTextEncodingNameKey = CFSTR("WebResourceTextEncodingName")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceTextEncodingName" "")); |
| 85 | static const CFStringRef LegacyWebArchiveResourceResponseKey = CFSTR("WebResourceResponse")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceResponse" "")); |
| 86 | static const CFStringRef LegacyWebArchiveResourceResponseVersionKey = CFSTR("WebResourceResponseVersion")((CFStringRef) __builtin___CFStringMakeConstantString ("" "WebResourceResponseVersion" "")); |
| 87 | |
| 88 | static bool NODELETE[[clang::annotate_type("webkit.nodelete")]] isUnreservedURICharacter(char16_t character) |
| 89 | { |
| 90 | return isASCIIAlphanumeric(character) || character == '-' || character == '.' || character == '_' || character == '~'; |
| 91 | } |
| 92 | |
| 93 | static String getFileNameFromURIComponent(StringView input) |
| 94 | { |
| 95 | auto decodedInput = WTF::URLParser::formURLDecode(input); |
| 96 | if (!decodedInput) |
| 97 | return { }; |
| 98 | |
| 99 | unsigned length = decodedInput->length(); |
| 100 | if (!length) |
| 101 | return { }; |
| 102 | |
| 103 | StringBuilder result; |
| 104 | result.reserveCapacity(length); |
| 105 | for (unsigned index = 0; index < length; ++index) { |
| 106 | char16_t character = decodedInput->characterAt(index); |
| 107 | if (isUnreservedURICharacter(character)) { |
| 108 | result.append(character); |
| 109 | continue; |
| 110 | } |
| 111 | result.append('-'); |
| 112 | } |
| 113 | |
| 114 | return result.toString(); |
| 115 | } |
| 116 | |
| 117 | static String generateValidFileName(const URL& url, const HashSet<String>& existingFileNames, const String& extension = { }) |
| 118 | { |
| 119 | String suffix = extension.isEmpty() ? emptyString() : makeString('.', extension); |
| 120 | auto extractedFileName = getFileNameFromURIComponent(url.lastPathComponent()); |
| 121 | if (extractedFileName.endsWith(suffix)) |
| 122 | extractedFileName = extractedFileName.left(extractedFileName.length() - suffix.length()); |
| 123 | auto fileName = extractedFileName.isEmpty() ? String::fromLatin1(defaultFileName) : extractedFileName; |
| 124 | |
| 125 | RELEASE_ASSERT(suffix.length() < maxFileNameSizeInBytes)do { if (__builtin_expect(!!(!(suffix.length() < maxFileNameSizeInBytes )), 0)) do { WTF::isIntegralOrPointerType(); compilerFenceForCrash (); WTFCrashWithInfo(125, "/Volumes/Data/worker/Apple-iOS-26-Safer-CPP-Checks-EWS/build/Source/WebCore/loader/archive/cf/LegacyWebArchive.cpp" , __PRETTY_FUNCTION__ ); } while (false); } while (0); |
| 126 | unsigned maxUniqueFileNameLength = maxFileNameSizeInBytes - suffix.length(); |
| 127 | String uniqueFileName; |
| 128 | |
| 129 | unsigned count = 0; |
| 130 | do { |
| 131 | uniqueFileName = fileName; |
| 132 | if (count) |
| 133 | uniqueFileName = makeString(fileName, '-', count); |
| 134 | if (uniqueFileName.sizeInBytes() > maxUniqueFileNameLength) |
| 135 | uniqueFileName = uniqueFileName.right(maxUniqueFileNameLength); |
| 136 | uniqueFileName = makeString(uniqueFileName, suffix); |
| 137 | ++count; |
| 138 | } while (existingFileNames.contains(uniqueFileName)); |
| 139 | |
| 140 | return uniqueFileName; |
| 141 | } |
| 142 | |
| 143 | RetainPtr<CFDictionaryRef> LegacyWebArchive::createPropertyListRepresentation(ArchiveResource* resource, MainResourceStatus isMainResource) |
| 144 | { |
| 145 | if (!resource) { |
| 146 | // The property list representation of a null/empty WebResource has the following 3 objects stored as nil. |
| 147 | // FIXME: 0 is not serializable. Presumably we need to use kCFNull here instead for compatibility. |
| 148 | // FIXME: But why do we need to support a resource of 0? Who relies on that? |
| 149 | RetainPtr<CFMutableDictionaryRef> propertyList = adoptCF(CFDictionaryCreateMutable(0, 3, 0, 0)); |
| 150 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceDataKey, 0); |
| 151 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceURLKey, 0); |
| 152 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceMIMETypeKey, 0); |
| 153 | return propertyList; |
| 154 | } |
| 155 | |
| 156 | auto propertyList = adoptCF(CFDictionaryCreateMutable(0, 6, 0, &kCFTypeDictionaryValueCallBacks)); |
| 157 | |
| 158 | // Resource data can be empty, but must be represented by an empty CFDataRef |
| 159 | Ref data = resource->data(); |
| 160 | |
| 161 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceDataKey, data->makeContiguous()->createCFData().get()); |
| 162 | |
| 163 | // Resource URL cannot be null |
| 164 | if (auto cfURL = resource->url().string().createCFString()) |
| 165 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceURLKey, cfURL.get()); |
| 166 | else { |
| 167 | LOG(Archives, "LegacyWebArchive - NULL resource URL is invalid - returning null property list")((void)0); |
| 168 | return nullptr; |
| 169 | } |
| 170 | |
| 171 | auto& filePath = resource->relativeFilePath(); |
| 172 | if (!filePath.isEmpty()) |
| 173 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceFilePathKey, filePath.createCFString().get()); |
| 174 | |
| 175 | // FrameName should be left out if empty for subresources, but always included for main resources |
| 176 | auto& frameName = resource->frameName(); |
| 177 | if (!frameName.isEmpty() || isMainResource) |
| 178 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceFrameNameKey, frameName.createCFString().get()); |
| 179 | |
| 180 | // Set MIMEType, TextEncodingName, and ResourceResponse only if they actually exist |
| 181 | auto& mimeType = resource->mimeType(); |
| 182 | if (!mimeType.isEmpty()) |
| 183 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceMIMETypeKey, mimeType.createCFString().get()); |
| 184 | |
| 185 | auto& textEncoding = resource->textEncoding(); |
| 186 | if (!textEncoding.isEmpty()) |
| 187 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceTextEncodingNameKey, textEncoding.createCFString().get()); |
| 188 | |
| 189 | // Don't include the resource response for the main resource |
| 190 | if (!isMainResource) { |
| 191 | if (auto resourceResponseData = createPropertyListRepresentation(resource->response())) |
| 192 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveResourceResponseKey, resourceResponseData.get()); |
| 193 | } |
| 194 | |
| 195 | return propertyList; |
| 196 | } |
| 197 | |
| 198 | RetainPtr<CFDictionaryRef> LegacyWebArchive::createPropertyListRepresentation(Archive& archive) |
| 199 | { |
| 200 | auto propertyList = adoptCF(CFDictionaryCreateMutable(0, 3, 0, &kCFTypeDictionaryValueCallBacks)); |
| 201 | |
| 202 | auto mainResourceDict = createPropertyListRepresentation(archive.mainResource(), MainResource); |
| 203 | ASSERT(mainResourceDict)((void)0); |
| 204 | if (!mainResourceDict) |
| 205 | return nullptr; |
| 206 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveMainResourceKey, mainResourceDict.get()); |
| 207 | |
| 208 | auto subresourcesArray = adoptCF(CFArrayCreateMutable(0, archive.subresources().size(), &kCFTypeArrayCallBacks)); |
| 209 | for (auto& resource : archive.subresources()) { |
| 210 | if (auto subresource = createPropertyListRepresentation(resource.ptr(), Subresource)) |
| 211 | CFArrayAppendValue(subresourcesArray.get(), subresource.get()); |
| 212 | else |
| 213 | LOG(Archives, "LegacyWebArchive - Failed to create property list for subresource")((void)0); |
| 214 | } |
| 215 | if (CFArrayGetCount(subresourcesArray.get())) |
| 216 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveSubresourcesKey, subresourcesArray.get()); |
| 217 | |
| 218 | auto subframesArray = adoptCF(CFArrayCreateMutable(0, archive.subframeArchives().size(), &kCFTypeArrayCallBacks)); |
| 219 | for (auto& subframe : archive.subframeArchives()) { |
| 220 | if (auto subframeArchive = createPropertyListRepresentation(subframe.get())) |
| 221 | CFArrayAppendValue(subframesArray.get(), subframeArchive.get()); |
| 222 | else |
| 223 | LOG(Archives, "LegacyWebArchive - Failed to create property list for subframe archive")((void)0); |
| 224 | } |
| 225 | if (CFArrayGetCount(subframesArray.get())) |
| 226 | CFDictionarySetValue(propertyList.get(), LegacyWebArchiveSubframeArchivesKey, subframesArray.get()); |
| 227 | |
| 228 | return propertyList; |
| 229 | } |
| 230 | |
| 231 | ResourceResponse LegacyWebArchive::createResourceResponseFromPropertyListData(CFDataRef data, CFStringRef responseDataType) |
| 232 | { |
| 233 | ASSERT(data)((void)0); |
| 234 | if (!data) |
| 235 | return ResourceResponse(); |
| 236 | |
| 237 | // If the ResourceResponseVersion (passed in as responseDataType) exists at all, this is a "new" web archive that we |
| 238 | // can parse well in a cross platform manner If it doesn't exist, we will assume this is an "old" web archive with, |
| 239 | // NSURLResponse objects in it and parse the ResourceResponse as such. |
| 240 | if (!responseDataType) |
| 241 | return createResourceResponseFromMacArchivedData(data); |
| 242 | |
| 243 | // FIXME: Parse the "new" format that the above comment references here. This format doesn't exist yet. |
| 244 | return ResourceResponse(); |
| 245 | } |
| 246 | |
| 247 | RefPtr<ArchiveResource> LegacyWebArchive::createResource(CFDictionaryRef dictionary) |
| 248 | { |
| 249 | ASSERT(dictionary)((void)0); |
| 250 | if (!dictionary) |
| 251 | return nullptr; |
| 252 | |
| 253 | RetainPtr resourceData = static_cast<CFDataRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceDataKey)); |
| 254 | if (resourceData && CFGetTypeID(resourceData.get()) != CFDataGetTypeID()) { |
| 255 | LOG(Archives, "LegacyWebArchive - Resource data is not of type CFData, cannot create invalid resource")((void)0); |
| 256 | return nullptr; |
| 257 | } |
| 258 | |
| 259 | RetainPtr frameName = static_cast<CFStringRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceFrameNameKey)); |
| 260 | if (frameName && CFGetTypeID(frameName.get()) != CFStringGetTypeID()) { |
| 261 | LOG(Archives, "LegacyWebArchive - Frame name is not of type CFString, cannot create invalid resource")((void)0); |
| 262 | return nullptr; |
| 263 | } |
| 264 | |
| 265 | RetainPtr mimeType = static_cast<CFStringRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceMIMETypeKey)); |
| 266 | if (mimeType && CFGetTypeID(mimeType.get()) != CFStringGetTypeID()) { |
| 267 | LOG(Archives, "LegacyWebArchive - MIME type is not of type CFString, cannot create invalid resource")((void)0); |
| 268 | return nullptr; |
| 269 | } |
| 270 | |
| 271 | RetainPtr url = static_cast<CFStringRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceURLKey)); |
| 272 | if (url && CFGetTypeID(url.get()) != CFStringGetTypeID()) { |
| 273 | LOG(Archives, "LegacyWebArchive - URL is not of type CFString, cannot create invalid resource")((void)0); |
| 274 | return nullptr; |
| 275 | } |
| 276 | |
| 277 | RetainPtr textEncoding = static_cast<CFStringRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceTextEncodingNameKey)); |
| 278 | if (textEncoding && CFGetTypeID(textEncoding.get()) != CFStringGetTypeID()) { |
| 279 | LOG(Archives, "LegacyWebArchive - Text encoding is not of type CFString, cannot create invalid resource")((void)0); |
| 280 | return nullptr; |
| 281 | } |
| 282 | |
| 283 | ResourceResponse response; |
| 284 | |
| 285 | if (RetainPtr resourceResponseData = static_cast<CFDataRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceResponseKey))) { |
| 286 | if (CFGetTypeID(resourceResponseData.get()) != CFDataGetTypeID()) { |
| 287 | LOG(Archives, "LegacyWebArchive - Resource response data is not of type CFData, cannot create invalid resource")((void)0); |
| 288 | return nullptr; |
| 289 | } |
| 290 | |
| 291 | RetainPtr resourceResponseVersion = static_cast<CFStringRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceResponseVersionKey)); |
| 292 | if (resourceResponseVersion && CFGetTypeID(resourceResponseVersion.get()) != CFStringGetTypeID()) { |
| 293 | LOG(Archives, "LegacyWebArchive - Resource response version is not of type CFString, cannot create invalid resource")((void)0); |
| 294 | return nullptr; |
| 295 | } |
| 296 | |
| 297 | response = createResourceResponseFromPropertyListData(resourceResponseData.get(), resourceResponseVersion.get()); |
| 298 | } |
| 299 | |
| 300 | auto filePathValue = CFDictionaryGetValue(dictionary, LegacyWebArchiveResourceFilePathKey); |
| 301 | RetainPtr filePath = dynamic_cf_cast<CFStringRef>(filePathValue); |
| 302 | if (filePathValue && !filePath) { |
| 303 | LOG(Archives, "LegacyWebArchive - File path is not of type CFString, cannot create invalid resource")((void)0); |
| 304 | return nullptr; |
| 305 | } |
| 306 | |
| 307 | return ArchiveResource::create(SharedBuffer::create(resourceData.get()), URL { url.get() }, mimeType.get(), textEncoding.get(), frameName.get(), response, filePath.get()); |
| 308 | } |
| 309 | |
| 310 | LegacyWebArchive::LegacyWebArchive(std::optional<FrameIdentifier> frameIdentifier, Vector<FrameIdentifier>&& subframeIdentifiers) |
| 311 | : m_frameIdentifier(frameIdentifier) |
| 312 | , m_subframeIdentifiers(WTF::move(subframeIdentifiers)) |
| 313 | { |
| 314 | } |
| 315 | |
| 316 | Ref<LegacyWebArchive> LegacyWebArchive::create(Ref<ArchiveResource>&& mainResource, Vector<Ref<ArchiveResource>>&& subresources, Vector<FrameIdentifier>&& subframeIdentifiers, std::optional<FrameIdentifier> mainFrameIdentifier) |
| 317 | { |
| 318 | auto archive = adoptRef(*new LegacyWebArchive(mainFrameIdentifier, WTF::move(subframeIdentifiers))); |
| 319 | archive->setMainResource(WTF::move(mainResource)); |
| 320 | |
| 321 | for (auto& subresource : subresources) |
| 322 | archive->addSubresource(WTF::move(subresource)); |
| 323 | |
| 324 | return archive; |
| 325 | } |
| 326 | |
| 327 | Ref<LegacyWebArchive> LegacyWebArchive::create(Ref<ArchiveResource>&& mainResource, Vector<Ref<ArchiveResource>>&& subresources, Vector<Ref<LegacyWebArchive>>&& subframeArchives, std::optional<FrameIdentifier> mainFrameIdentifier) |
| 328 | { |
| 329 | auto archive = adoptRef(*new LegacyWebArchive(mainFrameIdentifier, { })); |
| 330 | archive->setMainResource(WTF::move(mainResource)); |
| 331 | |
| 332 | for (auto& subresource : subresources) |
| 333 | archive->addSubresource(WTF::move(subresource)); |
| 334 | |
| 335 | for (auto& subframeArchive : subframeArchives) |
| 336 | archive->addSubframeArchive(WTF::move(subframeArchive)); |
| 337 | |
| 338 | return archive; |
| 339 | } |
| 340 | |
| 341 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(FragmentedSharedBuffer& data) |
| 342 | { |
| 343 | return create(URL(), data); |
| 344 | } |
| 345 | |
| 346 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(const URL&, FragmentedSharedBuffer& data) |
| 347 | { |
| 348 | LOG(Archives, "LegacyWebArchive - Creating from raw data")((void)0); |
| 349 | |
| 350 | RetainPtr<CFDataRef> cfData = data.makeContiguous()->createCFData(); |
| 351 | if (!cfData) |
| 352 | return nullptr; |
| 353 | |
| 354 | CFErrorRef error = nullptr; |
| 355 | |
| 356 | RetainPtr<CFDictionaryRef> plist = adoptCF(static_cast<CFDictionaryRef>(CFPropertyListCreateWithData(0, cfData.get(), kCFPropertyListImmutable, 0, &error))); |
| 357 | if (!plist) { |
| 358 | #if !LOG_DISABLED!0 |
| 359 | RetainPtr<CFStringRef> errorString = error ? adoptCF(CFErrorCopyDescription(error)) : 0; |
| 360 | const char* cError = errorString ? CFStringGetCStringPtr(errorString.get(), kCFStringEncodingUTF8) : "unknown error"; |
| 361 | LOG(Archives, "LegacyWebArchive - Error parsing PropertyList from archive data - %s", cError)((void)0); |
| 362 | #endif |
| 363 | if (error) |
| 364 | CFRelease(error); |
| 365 | return nullptr; |
| 366 | } |
| 367 | |
| 368 | if (CFGetTypeID(plist.get()) != CFDictionaryGetTypeID()) { |
| 369 | LOG(Archives, "LegacyWebArchive - Archive property list is not the expected CFDictionary, aborting invalid WebArchive")((void)0); |
| 370 | return nullptr; |
| 371 | } |
| 372 | |
| 373 | return create(plist.get()); |
| 374 | } |
| 375 | |
| 376 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(CFDictionaryRef dictionary) |
| 377 | { |
| 378 | ASSERT(dictionary)((void)0); |
| 379 | if (!dictionary) { |
| 380 | LOG(Archives, "LegacyWebArchive - Null root CFDictionary, aborting invalid WebArchive")((void)0); |
| 381 | return nullptr; |
| 382 | } |
| 383 | |
| 384 | RetainPtr mainResourceDict = static_cast<CFDictionaryRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveMainResourceKey)); |
| 385 | if (!mainResourceDict) { |
| 386 | LOG(Archives, "LegacyWebArchive - No main resource in archive, aborting invalid WebArchive")((void)0); |
| 387 | return nullptr; |
| 388 | } |
| 389 | if (CFGetTypeID(mainResourceDict.get()) != CFDictionaryGetTypeID()) { |
| 390 | LOG(Archives, "LegacyWebArchive - Main resource is not the expected CFDictionary, aborting invalid WebArchive")((void)0); |
| 391 | return nullptr; |
| 392 | } |
| 393 | |
| 394 | RefPtr mainResource = createResource(mainResourceDict.get()); |
| 395 | if (!mainResource) { |
| 396 | LOG(Archives, "LegacyWebArchive - Failed to parse main resource from CFDictionary or main resource does not exist, aborting invalid WebArchive")((void)0); |
| 397 | return nullptr; |
| 398 | } |
| 399 | |
| 400 | if (mainResource->mimeType().isNull()) { |
| 401 | LOG(Archives, "LegacyWebArchive - Main resource MIME type is required, but was null.")((void)0); |
| 402 | return nullptr; |
| 403 | } |
| 404 | |
| 405 | auto subresourceArray = static_cast<CFArrayRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveSubresourcesKey)); |
| 406 | if (subresourceArray && CFGetTypeID(subresourceArray) != CFArrayGetTypeID()) { |
| 407 | LOG(Archives, "LegacyWebArchive - Subresources is not the expected Array, aborting invalid WebArchive")((void)0); |
| 408 | return nullptr; |
| 409 | } |
| 410 | |
| 411 | Vector<Ref<ArchiveResource>> subresources; |
| 412 | if (subresourceArray) { |
| 413 | auto count = CFArrayGetCount(subresourceArray); |
| 414 | for (CFIndex i = 0; i < count; ++i) { |
| 415 | auto subresourceDict = static_cast<CFDictionaryRef>(CFArrayGetValueAtIndex(subresourceArray, i)); |
| 416 | if (CFGetTypeID(subresourceDict) != CFDictionaryGetTypeID()) { |
| 417 | LOG(Archives, "LegacyWebArchive - Subresource is not expected CFDictionary, aborting invalid WebArchive")((void)0); |
| 418 | return nullptr; |
| 419 | } |
| 420 | |
| 421 | if (auto subresource = createResource(subresourceDict)) |
| 422 | subresources.append(subresource.releaseNonNull()); |
| 423 | } |
| 424 | } |
| 425 | |
| 426 | auto subframeArray = static_cast<CFArrayRef>(CFDictionaryGetValue(dictionary, LegacyWebArchiveSubframeArchivesKey)); |
| 427 | if (subframeArray && CFGetTypeID(subframeArray) != CFArrayGetTypeID()) { |
| 428 | LOG(Archives, "LegacyWebArchive - Subframe archives is not the expected Array, aborting invalid WebArchive")((void)0); |
| 429 | return nullptr; |
| 430 | } |
| 431 | |
| 432 | Vector<Ref<LegacyWebArchive>> subframeArchives; |
| 433 | if (subframeArray) { |
| 434 | auto count = CFArrayGetCount(subframeArray); |
| 435 | for (CFIndex i = 0; i < count; ++i) { |
| 436 | auto subframeDict = static_cast<CFDictionaryRef>(CFArrayGetValueAtIndex(subframeArray, i)); |
| 437 | if (CFGetTypeID(subframeDict) != CFDictionaryGetTypeID()) { |
| 438 | LOG(Archives, "LegacyWebArchive - Subframe array is not expected CFDictionary, aborting invalid WebArchive")((void)0); |
| 439 | return nullptr; |
| 440 | } |
| 441 | |
| 442 | if (RefPtr subFrameArchive = create(subframeDict)) |
| 443 | subframeArchives.append(subFrameArchive.releaseNonNull()); |
| 444 | } |
| 445 | } |
| 446 | |
| 447 | return create(mainResource.releaseNonNull(), WTF::move(subresources), WTF::move(subframeArchives), std::nullopt); |
| 448 | } |
| 449 | |
| 450 | RetainPtr<CFDataRef> LegacyWebArchive::rawDataRepresentation() |
| 451 | { |
| 452 | auto propertyList = createPropertyListRepresentation(*this); |
| 453 | ASSERT(propertyList)((void)0); |
| 454 | if (!propertyList) { |
| 455 | LOG(Archives, "LegacyWebArchive - Failed to create property list for archive, returning no data")((void)0); |
| 456 | return nullptr; |
| 457 | } |
| 458 | |
| 459 | auto stream = adoptCF(CFWriteStreamCreateWithAllocatedBuffers(0, 0)); |
| 460 | |
| 461 | CFWriteStreamOpen(stream.get()); |
| 462 | CFPropertyListWrite(propertyList.get(), stream.get(), kCFPropertyListBinaryFormat_v1_0, 0, 0); |
| 463 | |
| 464 | auto plistData = adoptCF(static_cast<CFDataRef>(CFWriteStreamCopyProperty(stream.get(), kCFStreamPropertyDataWritten))); |
| 465 | ASSERT(plistData)((void)0); |
| 466 | |
| 467 | CFWriteStreamClose(stream.get()); |
| 468 | |
| 469 | if (!plistData) { |
| 470 | LOG(Archives, "LegacyWebArchive - Failed to convert property list into raw data, returning no data")((void)0); |
| 471 | return nullptr; |
| 472 | } |
| 473 | |
| 474 | return plistData; |
| 475 | } |
| 476 | |
| 477 | #if !PLATFORM(COCOA)(defined 1 && 1) |
| 478 | |
| 479 | ResourceResponse LegacyWebArchive::createResourceResponseFromMacArchivedData(CFDataRef responseData) |
| 480 | { |
| 481 | // FIXME: If is is possible to parse in a serialized NSURLResponse manually, without using |
| 482 | // NSKeyedUnarchiver, manipulating plists directly, then we want to do that here. |
| 483 | // Until then, this can be done on Mac only. |
| 484 | return ResourceResponse(); |
| 485 | } |
| 486 | |
| 487 | RetainPtr<CFDataRef> LegacyWebArchive::createPropertyListRepresentation(const ResourceResponse& response) |
| 488 | { |
| 489 | // FIXME: Write out the "new" format described in createResourceResponseFromPropertyListData once we invent it. |
| 490 | return nullptr; |
| 491 | } |
| 492 | |
| 493 | #endif |
| 494 | |
| 495 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(Node& node) |
| 496 | { |
| 497 | return create(node, { }); |
| 498 | } |
| 499 | |
| 500 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(Node& node, ArchiveOptions&& options, NOESCAPE[[clang::noescape]] const Function<bool(LocalFrame&)>& frameFilter) |
| 501 | { |
| 502 | RefPtr frame = node.document().frame(); |
| 503 | if (!frame) |
| 504 | return nullptr; |
| 505 | |
| 506 | auto currentOptions = WTF::move(options); |
| 507 | // If the page was loaded with JavaScript enabled, we don't want to archive <noscript> tags |
| 508 | // In practice we don't actually know whether scripting was enabled when the page was originally loaded |
| 509 | // but we can approximate that by checking if scripting is enabled right now. |
| 510 | if (frame->page() && frame->page()->settings().isScriptEnabled()) |
| 511 | currentOptions.markupExclusionRules.append(MarkupExclusionRule { AtomString { "noscript"_s }, { } }); |
| 512 | |
| 513 | // This archive is created for saving, and all subresources URLs will be rewritten to relative file paths |
| 514 | // based on the main resource file. |
| 515 | if (!currentOptions.mainResourceFileName.isEmpty()) |
| 516 | currentOptions.markupExclusionRules.append(MarkupExclusionRule { AtomString { "base"_s }, { } }); |
| 517 | |
| 518 | return createInternal(node, currentOptions, frameFilter); |
| 519 | } |
| 520 | |
| 521 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(LocalFrame& frame) |
| 522 | { |
| 523 | return create(frame, { }); |
| 524 | } |
| 525 | |
| 526 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(LocalFrame& frame, ArchiveOptions&& options) |
| 527 | { |
| 528 | RefPtr documentLoader = frame.loader().documentLoader(); |
| 529 | if (!documentLoader) |
| 530 | return nullptr; |
| 531 | |
| 532 | auto mainResource = documentLoader->mainResource(); |
| 533 | if (!mainResource) |
| 534 | return nullptr; |
| 535 | |
| 536 | Vector<Ref<LegacyWebArchive>> subframeArchives; |
| 537 | Vector<FrameIdentifier> subframeIdentifiers; |
| 538 | for (RefPtr child = frame.tree().firstChild(); child; child = child->tree().nextSibling()) { |
| 539 | if (options.shouldArchiveSubframes == ShouldArchiveSubframes::No) { |
| 540 | subframeIdentifiers.append(child->frameID()); |
| 541 | continue; |
| 542 | } |
| 543 | |
| 544 | if (auto localChild = dynamicDowncast<LocalFrame>(child.get())) { |
| 545 | if (auto childFrameArchive = create(*localChild, { })) |
| 546 | subframeArchives.append(childFrameArchive.releaseNonNull()); |
| 547 | } |
| 548 | } |
| 549 | |
| 550 | if (!subframeIdentifiers.isEmpty()) { |
| 551 | ASSERT(subframeArchives.isEmpty())((void)0); |
| 552 | return create(mainResource.releaseNonNull(), documentLoader->subresources(), WTF::move(subframeIdentifiers), frame.frameID()); |
| 553 | } |
| 554 | |
| 555 | return create(mainResource.releaseNonNull(), documentLoader->subresources(), WTF::move(subframeArchives), frame.frameID()); |
| 556 | } |
| 557 | |
| 558 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(const SimpleRange& range) |
| 559 | { |
| 560 | return LegacyWebArchive::create(range, { }); |
| 561 | } |
| 562 | |
| 563 | RefPtr<LegacyWebArchive> LegacyWebArchive::create(const SimpleRange& range, ArchiveOptions&& options) |
| 564 | { |
| 565 | Ref document = range.start.document(); |
| 566 | RefPtr frame = document->frame(); |
| 567 | if (!frame) |
| 568 | return nullptr; |
| 569 | |
| 570 | // FIXME: This is always "for interchange". Is that right? |
| 571 | Vector<Ref<Node>> nodeList; |
| 572 | auto markupString = makeString(documentTypeString(document), serializePreservingVisualAppearance(range, &nodeList, AnnotateForInterchange::Yes)); |
| 573 | return createInternal(markupString, WTF::move(options), *frame, WTF::move(nodeList), nullptr); |
| 574 | } |
| 575 | |
| 576 | #if ENABLE(ATTACHMENT_ELEMENT)(defined 1 && 1) |
| 577 | |
| 578 | static void addSubresourcesForAttachmentElementsIfNecessary(LocalFrame& frame, const Vector<Ref<Node>>& nodes, Vector<Ref<ArchiveResource>>& subresources) |
| 579 | { |
| 580 | if (!DeprecatedGlobalSettings::attachmentElementEnabled()) |
| 581 | return; |
| 582 | |
| 583 | Vector<String> identifiers; |
| 584 | for (Ref node : nodes) { |
| 585 | RefPtr attachment = dynamicDowncast<HTMLAttachmentElement>(node); |
| 586 | if (!attachment) |
| 587 | continue; |
| 588 | |
| 589 | auto uniqueIdentifier = attachment->uniqueIdentifier(); |
| 590 | if (uniqueIdentifier.isEmpty()) |
| 591 | continue; |
| 592 | |
| 593 | identifiers.append(WTF::move(uniqueIdentifier)); |
| 594 | } |
| 595 | |
| 596 | if (identifiers.isEmpty()) |
| 597 | return; |
| 598 | |
| 599 | CheckedPtr editorClient = frame.editor().client(); |
| 600 | if (!editorClient) |
| 601 | return; |
| 602 | |
| 603 | auto frameName = frame.tree().uniqueName(); |
| 604 | for (auto& data : editorClient->serializedAttachmentDataForIdentifiers(WTF::move(identifiers))) { |
| 605 | auto resourceURL = HTMLAttachmentElement::archiveResourceURL(data.identifier); |
| 606 | if (auto resource = ArchiveResource::create(data.data.ptr(), WTF::move(resourceURL), data.mimeType, { }, frameName)) |
| 607 | subresources.append(resource.releaseNonNull()); |
| 608 | } |
| 609 | } |
| 610 | |
| 611 | #endif |
| 612 | |
| 613 | static HashMap<Ref<CSSStyleSheet>, String> addSubresourcesForCSSStyleSheetsIfNecessary(LocalFrame& frame, const String& subresourcesDirectoryName, HashSet<String>& uniqueFileNames, HashMap<String, String>& uniqueSubresources, Vector<Ref<ArchiveResource>>& subresources) |
| 614 | { |
| 615 | if (subresourcesDirectoryName.isEmpty()) |
| 616 | return { }; |
| 617 | |
| 618 | RefPtr document = frame.document(); |
| 619 | if (!document) |
| 620 | return { }; |
| 621 | |
| 622 | CSS::SerializationContext serializationContext; |
| 623 | |
| 624 | HashMap<Ref<CSSStyleSheet>, String> uniqueCSSStyleSheets; |
| 625 | Ref documentStyleSheets = document->styleSheets(); |
| 626 | for (unsigned index = 0; index < documentStyleSheets->length(); ++index) { |
| 627 | RefPtr cssStyleSheet = dynamicDowncast<CSSStyleSheet>(documentStyleSheets->item(index)); |
| 628 | if (!cssStyleSheet) |
| 629 | continue; |
| 630 | |
| 631 | if (uniqueCSSStyleSheets.contains(*cssStyleSheet)) |
| 632 | continue; |
| 633 | |
| 634 | HashSet<Ref<CSSStyleSheet>> cssStyleSheets; |
| 635 | cssStyleSheets.add(*cssStyleSheet); |
| 636 | cssStyleSheet->getChildStyleSheets(cssStyleSheets); |
| 637 | for (auto& currentCSSStyleSheet : cssStyleSheets) { |
| 638 | bool isExternalStyleSheet = !currentCSSStyleSheet->href().isEmpty() || currentCSSStyleSheet->ownerRule(); |
| 639 | if (!isExternalStyleSheet) |
| 640 | continue; |
| 641 | |
| 642 | auto url = currentCSSStyleSheet->baseURL(); |
| 643 | if (url.isNull() || url.isEmpty()) |
| 644 | continue; |
| 645 | |
| 646 | auto addResult = uniqueCSSStyleSheets.add(currentCSSStyleSheet.copyRef(), emptyString()); |
| 647 | if (!addResult.isNewEntry) |
| 648 | continue; |
| 649 | |
| 650 | // Delete cached resource for this style sheet. |
| 651 | auto index = subresources.findIf([&](auto& subresource) { |
| 652 | return subresource->url() == url; |
| 653 | }); |
| 654 | if (index != notFound) { |
| 655 | auto fileName = FileSystem::lastComponentOfPathIgnoringTrailingSlash(subresources[index]->relativeFilePath()); |
| 656 | uniqueFileNames.remove(fileName); |
| 657 | uniqueSubresources.remove(url.string()); |
| 658 | subresources.removeAt(index); |
| 659 | } |
| 660 | |
| 661 | auto extension = MIMETypeRegistry::preferredExtensionForMIMEType(cssContentTypeAtom()); |
| 662 | String subresourceFileName = generateValidFileName(url, uniqueFileNames, extension); |
| 663 | uniqueFileNames.add(subresourceFileName); |
| 664 | addResult.iterator->value = FileSystem::pathByAppendingComponent(subresourcesDirectoryName, subresourceFileName); |
| 665 | serializationContext.replacementURLStringsForCSSStyleSheet.add(currentCSSStyleSheet.copyRef(), subresourceFileName); |
| 666 | } |
| 667 | } |
| 668 | |
| 669 | auto frameName = frame.tree().uniqueName(); |
| 670 | for (auto& [urlString, path] : uniqueSubresources) { |
| 671 | // The style sheet files are stored in the same directory as other subresources. |
| 672 | serializationContext.replacementURLStrings.add(urlString, FileSystem::lastComponentOfPathIgnoringTrailingSlash(path)); |
| 673 | } |
| 674 | |
| 675 | for (auto& [cssStyleSheet, path] : uniqueCSSStyleSheets) { |
| 676 | auto contentString = cssStyleSheet->cssText(serializationContext); |
| 677 | if (auto newResource = ArchiveResource::create(utf8Buffer(contentString), URL { cssStyleSheet->href() }, "text/css"_s, "UTF-8"_s, frameName, ResourceResponse(), path)) |
| 678 | subresources.append(newResource.releaseNonNull()); |
| 679 | } |
| 680 | |
| 681 | return frame.isMainFrame() ? uniqueCSSStyleSheets : serializationContext.replacementURLStringsForCSSStyleSheet; |
| 682 | } |
| 683 | |
| 684 | RefPtr<LegacyWebArchive> LegacyWebArchive::createInternal(Node& node, const ArchiveOptions& options, NOESCAPE[[clang::noescape]] const Function<bool(LocalFrame&)>& frameFilter) |
| 685 | { |
| 686 | RefPtr frame = node.document().frame(); |
| 687 | if (!frame) |
| 688 | return nullptr; |
| 689 | |
| 690 | Vector<Ref<Node>> nodeList; |
| 691 | String markupString = serializeFragment(node, SerializedNodes::SubtreeIncludingNode, &nodeList, ResolveURLs::No, std::nullopt, SerializeShadowRoots::AllForInterchange, { }, options.markupExclusionRules); |
| 692 | auto nodeType = node.nodeType(); |
| 693 | if (nodeType != NodeType::Document && nodeType != NodeType::DocumentType) |
| 694 | markupString = makeString(documentTypeString(node.document()), markupString); |
| 695 | |
| 696 | return createInternal(markupString, options, *frame, WTF::move(nodeList), frameFilter); |
| 697 | } |
| 698 | |
| 699 | RefPtr<LegacyWebArchive> LegacyWebArchive::createInternal(const String& markupString, const ArchiveOptions& options, LocalFrame& frame, Vector<Ref<Node>>&& nodes, NOESCAPE[[clang::noescape]] const Function<bool(LocalFrame&)>& frameFilter) |
| 700 | { |
| 701 | auto& response = frame.loader().documentLoader()->response(); |
| 702 | URL responseURL = response.url(); |
| 703 | |
| 704 | // it's possible to have a response without a URL here |
| 705 | // <rdar://problem/5454935> |
| 706 | if (responseURL.isNull()) |
| 707 | responseURL = URL({ }, emptyString()); |
| 708 | |
| 709 | auto mainResource = ArchiveResource::create(utf8Buffer(markupString), responseURL, response.mimeType(), "UTF-8"_s, frame.tree().uniqueName()); |
| 710 | if (!mainResource) |
| 711 | return nullptr; |
| 712 | |
| 713 | Vector<Ref<LegacyWebArchive>> subframeArchives; |
| 714 | Vector<FrameIdentifier> subframeIdentifiers; |
| 715 | Vector<Ref<ArchiveResource>> subresources; |
| 716 | HashMap<String, String> uniqueSubresources; |
| 717 | HashSet<String> uniqueFileNames; |
| 718 | String subresourcesDirectoryName = options.mainResourceFileName.isNull() ? String { } : makeString(options.mainResourceFileName, "_files"_s); |
| 719 | |
| 720 | for (auto& node : nodes) { |
| 721 | RefPtr frameOwnerElement = dynamicDowncast<HTMLFrameOwnerElement>(node); |
| 722 | RefPtr childFrame = frameOwnerElement ? frameOwnerElement->contentFrame() : nullptr; |
| 723 | RefPtr localChildFrame = dynamicDowncast<LocalFrame>(childFrame.get()); |
| 724 | if (childFrame) { |
| 725 | if (frameFilter && localChildFrame && !frameFilter(*localChildFrame)) |
| 726 | continue; |
| 727 | |
| 728 | if (options.shouldArchiveSubframes == ShouldArchiveSubframes::No) { |
| 729 | subframeIdentifiers.append(childFrame->frameID()); |
| 730 | continue; |
| 731 | } |
| 732 | |
| 733 | if (!localChildFrame) |
| 734 | continue; |
| 735 | |
| 736 | if (auto subframeArchive = createInternal(*localChildFrame->document(), options, frameFilter)) { |
| 737 | RefPtr subframeMainResource = subframeArchive->mainResource(); |
| 738 | auto subframeMainResourceURL = subframeMainResource ? subframeMainResource->url() : URL { }; |
| 739 | if (!subframeMainResourceURL.isEmpty()) { |
| 740 | auto subframeMainResourceRelativePath = frame.isMainFrame() ? subframeMainResource->relativeFilePath() : FileSystem::lastComponentOfPathIgnoringTrailingSlash(subframeMainResource->relativeFilePath()); |
| 741 | uniqueSubresources.add(makeString(childFrame->frameID().toUInt64()), subframeMainResourceRelativePath); |
| 742 | } |
| 743 | subframeArchives.append(subframeArchive.releaseNonNull()); |
| 744 | } else |
| 745 | LOG_ERROR("Unabled to archive subframe %s", childFrame->tree().uniqueName().string().utf8().data())((void)0); |
| 746 | |
| 747 | } else { |
| 748 | ListHashSet<URL> subresourceURLs; |
| 749 | node->getSubresourceURLs(subresourceURLs); |
| 750 | node->getCandidateSubresourceURLs(subresourceURLs); |
| 751 | |
| 752 | if (options.shouldSaveScriptsFromMemoryCache == ShouldSaveScriptsFromMemoryCache::Yes && responseURL.protocolIsInHTTPFamily()) { |
| 753 | RegistrableDomain domain { responseURL }; |
| 754 | MemoryCache::singleton().forEachSessionResource(frame.page()->sessionID(), [&](auto& resource) { |
| 755 | if (domain.matches(resource.url()) && resource.hasClients() && (resource.type() == CachedResource::Type::Script || resource.type() == CachedResource::Type::JSON)) |
| 756 | subresourceURLs.add(resource.url()); |
| 757 | }); |
| 758 | } |
| 759 | |
| 760 | ASSERT(frame.loader().documentLoader())((void)0); |
| 761 | Ref documentLoader = *frame.loader().documentLoader(); |
| 762 | |
| 763 | for (auto& subresourceURL : subresourceURLs) { |
| 764 | if (uniqueSubresources.contains(subresourceURL.string())) |
| 765 | continue; |
| 766 | |
| 767 | // WebArchive is created for saving, and we don't need to store resources for data URLs. |
| 768 | if (!subresourcesDirectoryName.isNull() && subresourceURL.protocolIsData()) |
| 769 | continue; |
| 770 | |
| 771 | auto addResult = uniqueSubresources.add(subresourceURL.string(), emptyString()); |
| 772 | auto resource = documentLoader->subresource(subresourceURL); |
| 773 | if (!resource) { |
| 774 | ResourceRequest request(URL { subresourceURL }); |
| 775 | request.setDomainForCachePartition(frame.document()->domainForCachePartition()); |
| 776 | if (auto* cachedResource = MemoryCache::singleton().resourceForRequest(request, frame.page()->sessionID())) |
Local variable 'cachedResource' is uncounted and unsafe | |
| 777 | resource = ArchiveResource::create(cachedResource->resourceBuffer(), subresourceURL, cachedResource->response()); |
| 778 | } |
| 779 | |
| 780 | if (!resource) { |
| 781 | // FIXME: should do something better than spew to console here |
| 782 | LOG_ERROR("Failed to archive subresource for %s", subresourceURL.string().utf8().data())((void)0); |
| 783 | continue; |
| 784 | } |
| 785 | |
| 786 | if (!subresourcesDirectoryName.isNull()) { |
| 787 | String subresourceFileName = generateValidFileName(subresourceURL, uniqueFileNames); |
| 788 | uniqueFileNames.add(subresourceFileName); |
| 789 | String subresourceFilePath = FileSystem::pathByAppendingComponent(subresourcesDirectoryName, subresourceFileName); |
| 790 | resource->setRelativeFilePath(subresourceFilePath); |
| 791 | addResult.iterator->value = frame.isMainFrame() ? subresourceFilePath : subresourceFileName; |
| 792 | } |
| 793 | |
| 794 | subresources.append(resource.releaseNonNull()); |
| 795 | } |
| 796 | } |
| 797 | } |
| 798 | |
| 799 | auto uniqueCSSStyleSheets = addSubresourcesForCSSStyleSheetsIfNecessary(frame, subresourcesDirectoryName, uniqueFileNames, uniqueSubresources, subresources); |
| 800 | |
| 801 | #if ENABLE(ATTACHMENT_ELEMENT)(defined 1 && 1) |
| 802 | addSubresourcesForAttachmentElementsIfNecessary(frame, nodes, subresources); |
| 803 | #endif |
| 804 | |
| 805 | // If we are archiving the entire page, add any link icons that we have data for. |
| 806 | if (!nodes.isEmpty() && nodes[0]->isDocumentNode()) { |
| 807 | RefPtr documentLoader = frame.loader().documentLoader(); |
| 808 | ASSERT(documentLoader)((void)0); |
| 809 | for (auto& icon : documentLoader->linkIcons()) { |
| 810 | if (auto resource = documentLoader->subresource(icon.url)) |
| 811 | subresources.append(resource.releaseNonNull()); |
| 812 | } |
| 813 | } |
| 814 | |
| 815 | if (!options.mainResourceFileName.isNull()) { |
| 816 | RefPtr document = frame.document(); |
| 817 | if (!document) |
| 818 | return nullptr; |
| 819 | |
| 820 | if (responseURL.isEmpty()) |
| 821 | return nullptr; |
| 822 | |
| 823 | auto extension = MIMETypeRegistry::preferredExtensionForMIMEType(textHTMLContentTypeAtom()); |
| 824 | if (!extension.isEmpty()) |
| 825 | extension = makeString('.', extension); |
| 826 | auto mainFrameFileNameWithExtension = options.mainResourceFileName.endsWith(extension) ? options.mainResourceFileName : makeString(options.mainResourceFileName, extension); |
| 827 | auto fileNameWithExtension = frame.isMainFrame() ? mainFrameFileNameWithExtension : makeString(subresourcesDirectoryName, "/frame_"_s, frame.frameID().toUInt64(), extension); |
| 828 | |
| 829 | ResolveURLs resolveURLs = ResolveURLs::No; |
| 830 | // Base element is excluded, so all URLs should be replaced with absolute URL. |
| 831 | bool baseElementExcluded = std::ranges::any_of(options.markupExclusionRules, [&](auto& rule) { |
| 832 | return rule.elementLocalName == "base"_s; |
| 833 | }); |
| 834 | if (!document->baseElementURL().isEmpty() && baseElementExcluded) |
| 835 | resolveURLs = ResolveURLs::Yes; |
| 836 | |
| 837 | String updatedMarkupString = serializeFragmentWithURLReplacement(*document, SerializedNodes::SubtreeIncludingNode, nullptr, resolveURLs, std::nullopt, WTF::move(uniqueSubresources), WTF::move(uniqueCSSStyleSheets), SerializeShadowRoots::AllForInterchange, { }, options.markupExclusionRules); |
| 838 | mainResource = ArchiveResource::create(utf8Buffer(updatedMarkupString), responseURL, response.mimeType(), "UTF-8"_s, frame.tree().uniqueName(), ResourceResponse(), fileNameWithExtension); |
| 839 | } |
| 840 | |
| 841 | if (!subframeIdentifiers.isEmpty()) { |
| 842 | ASSERT(subframeArchives.isEmpty())((void)0); |
| 843 | return create(mainResource.releaseNonNull(), WTF::move(subresources), WTF::move(subframeIdentifiers), frame.frameID()); |
| 844 | } |
| 845 | |
| 846 | return create(mainResource.releaseNonNull(), WTF::move(subresources), WTF::move(subframeArchives), frame.frameID()); |
| 847 | } |
| 848 | |
| 849 | RefPtr<LegacyWebArchive> LegacyWebArchive::createFromSelection(LocalFrame* frame) |
| 850 | { |
| 851 | return createFromSelection(frame, { }); |
| 852 | } |
| 853 | |
| 854 | RefPtr<LegacyWebArchive> LegacyWebArchive::createFromSelection(LocalFrame* frame, ArchiveOptions&& options) |
| 855 | { |
| 856 | if (!frame) |
| 857 | return nullptr; |
| 858 | |
| 859 | RefPtr document = frame->document(); |
| 860 | if (!document) |
| 861 | return nullptr; |
| 862 | |
| 863 | StringBuilder builder; |
| 864 | builder.append(documentTypeString(*document)); |
| 865 | |
| 866 | Vector<Ref<Node>> nodeList; |
| 867 | builder.append(serializePreservingVisualAppearance(frame->selection().selection(), ResolveURLs::No, SerializeComposedTree::Yes, IgnoreUserSelectNone::Yes, PreserveBaseElement::Yes, PreserveDirectionForInlineText::Yes, &nodeList)); |
| 868 | |
| 869 | RefPtr archive = createInternal(builder.toString(), WTF::move(options), *frame, WTF::move(nodeList), { }); |
| 870 | if (!archive) |
| 871 | return nullptr; |
| 872 | |
| 873 | if (!document->isFrameSet()) |
| 874 | return archive; |
| 875 | |
| 876 | // Wrap the frameset document in an iframe so it can be pasted into |
| 877 | // another document (which will have a body or frameset of its own). |
| 878 | auto iframeMarkup = makeString("<iframe frameborder=\"no\" marginwidth=\"0\" marginheight=\"0\" width=\"98%%\" height=\"98%%\" src=\""_s, frame->loader().documentLoader()->response().url().string(), "\"></iframe>"_s); |
| 879 | auto iframeResource = ArchiveResource::create(utf8Buffer(iframeMarkup), aboutBlankURL(), textHTMLContentTypeAtom(), "UTF-8"_s, String()); |
| 880 | |
| 881 | return create(iframeResource.releaseNonNull(), { }, { archive.releaseNonNull() }, frame->frameID()); |
| 882 | } |
| 883 | |
| 884 | } |