AppCache INTERCEPT namespace.

webkit/appcache/manifest_parser.cc

 // Copyright (c) 2011 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 //
 // This is a port of ManifestParser.cc from WebKit/WebCore/loader/appcache.
 
 /*
  * Copyright (C) 2008 Apple Inc. All Rights Reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ skipping 22 matching lines @@
 
 #include "base/i18n/icu_string_conversions.h"
 #include "base/logging.h"
 #include "base/utf_string_conversions.h"
 #include "googleurl/src/gurl.h"
 
 namespace appcache {
 
 enum Mode {
   EXPLICIT,
+  INTERCEPT,
   FALLBACK,
   ONLINE_WHITELIST,
   UNKNOWN,
 };
 
 Manifest::Manifest() : online_whitelist_all(false) {}
 
 Manifest::~Manifest() {}
 
 bool ParseManifest(const GURL& manifest_url, const char* data, int length,
                    Manifest& manifest) {
   // This is an implementation of the parsing algorithm specified in
   // the HTML5 offline web application docs:
   //   http://www.w3.org/TR/html5/offline.html
   // Do not modify it without consulting those docs.
   // Though you might be tempted to convert these wstrings to UTF-8 or
   // string16, this implementation seems simpler given the constraints.
 
   const wchar_t kSignature[] = L"CACHE MANIFEST";
   const size_t kSignatureLength = arraysize(kSignature) - 1;
+  const wchar_t kChromiumSignature[] = L"CHROMIUM CACHE MANIFEST";
+  const size_t kChromiumSignatureLength = arraysize(kChromiumSignature) - 1;
 
   DCHECK(manifest.explicit_urls.empty());
   DCHECK(manifest.fallback_namespaces.empty());
   DCHECK(manifest.online_whitelist_namespaces.empty());
   DCHECK(!manifest.online_whitelist_all);
 
   Mode mode = EXPLICIT;
 
   std::wstring data_string;
   // TODO(jennb): cannot do UTF8ToWide(data, length, &data_string);
   // until UTF8ToWide uses 0xFFFD Unicode replacement character.
   base::CodepageToWide(std::string(data, length), base::kCodepageUTF8,
                        base::OnStringConversionError::SUBSTITUTE, &data_string);
   const wchar_t* p = data_string.c_str();
   const wchar_t* end = p + data_string.length();
 
   // Look for the magic signature: "^\xFEFF?CACHE MANIFEST[ \t]?"
   // Example: "CACHE MANIFEST #comment" is a valid signature.
   // Example: "CACHE MANIFEST;V2" is not.
 
   // When the input data starts with a UTF-8 Byte-Order-Mark
   // (0xEF, 0xBB, 0xBF), the UTF8ToWide() function converts it to a
   // Unicode BOM (U+FEFF). Skip a converted Unicode BOM if it exists.
   int bom_offset = 0;
   if (!data_string.empty() && data_string[0] == 0xFEFF) {
     bom_offset = 1;
     ++p;
   }
 
-  if (p >= end ||
-      data_string.compare(bom_offset, kSignatureLength, kSignature)) {
+  if (p >= end)
+    return false;
+
+  // Check for a supported signature and skip p past it.
+  if (0 == data_string.compare(bom_offset, kSignatureLength,
+                               kSignature)) {
+    p += kSignatureLength;
+  } else if (0 == data_string.compare(bom_offset, kChromiumSignatureLength,
+                                      kChromiumSignature)) {
+    p += kChromiumSignatureLength;
+  } else {
     return false;
   }
 
-  p += kSignatureLength;  // Skip past "CACHE MANIFEST"
-
   // Character after "CACHE MANIFEST" must be whitespace.
   if (p < end && *p != ' ' && *p != '\t' && *p != '\n' && *p != '\r')
     return false;
 
   // Skip to the end of the line.
   while (p < end && *p != '\r' && *p != '\n')
     ++p;
 
   while (1) {
     // Skip whitespace
@@ skipping 19 matching lines @@
       --tmp;
 
     std::wstring line(line_start, tmp - line_start + 1);
 
     if (line == L"CACHE:") {
       mode = EXPLICIT;
     } else if (line == L"FALLBACK:") {
       mode = FALLBACK;
     } else if (line == L"NETWORK:") {
       mode = ONLINE_WHITELIST;
+    } else if (line == L"CHROMIUM-INTERCEPT:") {
+      mode = INTERCEPT;
     } else if (*(line.end() - 1) == ':') {
       mode = UNKNOWN;
     } else if (mode == UNKNOWN) {
       continue;
     } else if (line == L"*" && mode == ONLINE_WHITELIST) {
       manifest.online_whitelist_all = true;
       continue;
     } else if (mode == EXPLICIT || mode == ONLINE_WHITELIST) {
       const wchar_t *line_p = line.c_str();
       const wchar_t *line_end = line_p + line.length();
@@ skipping 24 matching lines @@
       // Per the spec, EXPLICIT cross-origin HTTS resources should be
       // ignored here. We've opted for a milder constraint and allow
       // caching unless the resource has a "no-store" header. That
       // condition is enforced in AppCacheUpdateJob.
 
       if (mode == EXPLICIT) {
         manifest.explicit_urls.insert(url.spec());
       } else {
         manifest.online_whitelist_namespaces.push_back(url);
       }
+    } else if (mode == INTERCEPT) {
+      // Lines of the form,
+      // <urlnamespace> <intercept_type> <targeturl>
+      const wchar_t* line_p = line.c_str();
+      const wchar_t* line_end = line_p + line.length();
+
+      // Look for first whitespace separating the url namespace from
+      // the intercept type.
+      while (line_p < line_end && *line_p != '\t' && *line_p != ' ')
+        ++line_p;
+
+      if (line_p == line_end)
+        continue;  // There was no whitespace separating the URLs.
+
+      string16 namespace_url16;
+      WideToUTF16(line.c_str(), line_p - line.c_str(), &namespace_url16);
+      GURL namespace_url = manifest_url.Resolve(namespace_url16);
+      if (!namespace_url.is_valid())
+        continue;
+      if (namespace_url.has_ref()) {
+        GURL::Replacements replacements;
+        replacements.ClearRef();
+        namespace_url = namespace_url.ReplaceComponents(replacements);
+      }
+
+      // The namespace URL must have the same scheme, host and port
+      // as the manifest's URL.
+      if (manifest_url.GetOrigin() != namespace_url.GetOrigin())
+        continue;
+
+      // Skip whitespace separating namespace from the type.
+      while (line_p < line_end && (*line_p == '\t' || *line_p == ' '))
+        ++line_p;
+
+      // Look for whitespace separating the type from the target url.
+      const wchar_t* type_start = line_p;
+      while (line_p < line_end && *line_p != '\t' && *line_p != ' ')
+        ++line_p;
+
+      // Look for a type value we understand, otherwise skip the line.
+      std::wstring type(type_start, line_p - type_start);
+      if (type != L"return")
+        continue;
+
+      // Skip whitespace separating type from the target_url.
+      while (line_p < line_end && (*line_p == '\t' || *line_p == ' '))
+        ++line_p;
+
+      // Look for whitespace separating the URL from subsequent ignored tokens.
+      const wchar_t* target_url_start = line_p;
+      while (line_p < line_end && *line_p != '\t' && *line_p != ' ')
+        ++line_p;
+
+      string16 target_url16;
+      WideToUTF16(target_url_start, line_p - target_url_start, &target_url16);
+      GURL target_url = manifest_url.Resolve(target_url16);
+      if (!target_url.is_valid())
+        continue;
+
+      if (target_url.has_ref()) {
+        GURL::Replacements replacements;
+        replacements.ClearRef();
+        target_url = target_url.ReplaceComponents(replacements);
+      }
+      if (manifest_url.GetOrigin() != target_url.GetOrigin())
+        continue;
+
+      manifest.intercept_namespaces.push_back(
+          Namespace(INTERCEPT_NAMESPACE, namespace_url, target_url));
     } else if (mode == FALLBACK) {
       const wchar_t* line_p = line.c_str();
       const wchar_t* line_end = line_p + line.length();
 
       // Look for whitespace separating the two URLs
       while (line_p < line_end && *line_p != '\t' && *line_p != ' ')
         ++line_p;
 
       if (line_p == line_end) {
         // There was no whitespace separating the URLs.
@@ skipping 39 matching lines @@
 
       // Fallback entry URL must have the same scheme, host and port
       // as the manifest's URL.
       if (manifest_url.GetOrigin() != fallback_url.GetOrigin()) {
         continue;
       }
 
       // Store regardless of duplicate namespace URL. Only first match
       // will ever be used.
       manifest.fallback_namespaces.push_back(
-          FallbackNamespace(namespace_url, fallback_url));
+          Namespace(FALLBACK_NAMESPACE, namespace_url, fallback_url));
     } else {
       NOTREACHED();
     }
   }
 
   return true;
 }
 
 }  // namespace appcache