Re-land SDCH filter support experiment

Fix up solution files for webkit and net

Add one line keyword change to help linux build

r=hunanr,openvcdiff,nsylvain
Review URL: http://codereview.chromium.org/4026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@2469 0039d316-1c4b-4281-b951-d872f2087c98

4 files changed, 85 insertions, 22 deletions

diff --git a/net/url_request/url_request_http_job.cc b/net/url_request/url_request_http_job.cc
index 8316e73..094acd8 100644
--- a/net/url_request/url_request_http_job.cc
+++ b/net/url_request/url_request_http_job.cc
@@ -5,12 +5,15 @@
 #include "net/url_request/url_request_http_job.h"
 
 #include "base/compiler_specific.h"
+#include "base/file_util.h"
+#include "base/file_version_info.h"
 #include "base/message_loop.h"
 #include "base/string_util.h"
 #include "net/base/cookie_monster.h"
 #include "net/base/load_flags.h"
 #include "net/base/net_errors.h"
 #include "net/base/net_util.h"
+#include "net/base/sdch_manager.h"
 #include "net/http/http_response_info.h"
 #include "net/http/http_transaction.h"
 #include "net/http/http_transaction_factory.h"
@@ -159,15 +162,20 @@ int URLRequestHttpJob::GetResponseCode() {
   return response_info_->headers->response_code();
 }
 
-bool URLRequestHttpJob::GetContentEncoding(std::string* encoding_type) {
+bool URLRequestHttpJob::GetContentEncodings(
+    std::vector<std::string>* encoding_types) {
   DCHECK(transaction_);
 
   if (!response_info_)
     return false;
 
-  // TODO(darin): what if there are multiple content encodings?
-  return response_info_->headers->EnumerateHeader(NULL, "Content-Encoding",
-                                                  encoding_type);
+  std::string encoding_type;
+  void* iter = NULL;
+  while (response_info_->headers->EnumerateHeader(&iter, "Content-Encoding",
+                                                  &encoding_type)) {
+    encoding_types->push_back(encoding_type);
+  }
+  return !encoding_types->empty();
 }
 
 bool URLRequestHttpJob::IsRedirectResponse(GURL* location,
@@ -413,6 +421,23 @@ void URLRequestHttpJob::NotifyHeadersComplete() {
     }
   }
 
+  // Get list of SDCH dictionary requests, and schedule them to be loaded.
+  if (SdchManager::Global()->IsInSupportedDomain(request_->url())) {
+    static const std::string name = "Get-Dictionary";
+    std::string url_text;
+    void* iter = NULL;
+    // TODO(jar): We need to not fetch dictionaries the first time they are
+    // seen, but rather wait until we can justify their usefulness.
+    // For now, we will only fetch the first dictionary, which will at least
+    // require multiple suggestions before we get additional ones for this site.
+    // Eventually we should wait until a dictionary is requested several times
+    // before we even download it (so that we don't waste memory or bandwidth).
+    if (response_info_->headers->EnumerateHeader(&iter, name, &url_text)) {
+      GURL dictionary_url = request_->url().Resolve(url_text);
+      SdchManager::Global()->FetchDictionary(request_->url(), dictionary_url);
+    }
+  }
+
   URLRequestJob::NotifyHeadersComplete();
 }
 
@@ -476,8 +501,32 @@ void URLRequestHttpJob::AddExtraHeaders() {
           context->accept_charset() + "\r\n";
   }
 
+  if (!SdchManager::Global()->IsInSupportedDomain(request_->url())) {
+    // Tell the server what compression formats we support (other than SDCH).
+    request_info_.extra_headers += "Accept-Encoding: gzip,deflate,bzip2\r\n";
+    return;
+  }
+
+  // Supply SDCH related headers, as well as accepting that encoding.
+
+  // TODO(jar): See if it is worth optimizing away these bytes when the URL is
+  // probably an img or such. (and SDCH encoding is not likely).
+  std::string avail_dictionaries;
+  SdchManager::Global()->GetAvailDictionaryList(request_->url(),
+                                                &avail_dictionaries);
+  if (!avail_dictionaries.empty())
+    request_info_.extra_headers += "Avail-Dictionary: "
+        + avail_dictionaries + "\r\n";
+
+  scoped_ptr<FileVersionInfo> file_version_info(
+    FileVersionInfo::CreateFileVersionInfoForCurrentModule());
+  request_info_.extra_headers += "X-SDCH: Chrome ";
+  request_info_.extra_headers +=
+      WideToASCII(file_version_info->product_version());
+  request_info_.extra_headers += "\r\n";
+
   // Tell the server what compression formats we support.
-  request_info_.extra_headers += "Accept-Encoding: gzip,deflate,bzip2\r\n";
+  request_info_.extra_headers += "Accept-Encoding: gzip,deflate,bzip2,sdch\r\n";
 }
 
 void URLRequestHttpJob::FetchResponseCookies() {
diff --git a/net/url_request/url_request_http_job.h b/net/url_request/url_request_http_job.h
index 9312e35..908db2c 100644
--- a/net/url_request/url_request_http_job.h
+++ b/net/url_request/url_request_http_job.h
@@ -2,8 +2,11 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
-#ifndef NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H__
-#define NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H__
+#ifndef NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H_
+#define NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H_
+
+#include <string>
+#include <vector>
 
 #include "net/base/completion_callback.h"
 #include "net/http/http_request_info.h"
@@ -24,7 +27,7 @@ class URLRequestHttpJob : public URLRequestJob {
   virtual ~URLRequestHttpJob();
 
  protected:
-  URLRequestHttpJob(URLRequest* request);
+  explicit URLRequestHttpJob(URLRequest* request);
 
   // URLRequestJob methods:
   virtual void SetUpload(net::UploadData* upload);
@@ -38,7 +41,7 @@ class URLRequestHttpJob : public URLRequestJob {
   virtual void GetResponseInfo(net::HttpResponseInfo* info);
   virtual bool GetResponseCookies(std::vector<std::string>* cookies);
   virtual int GetResponseCode();
-  virtual bool GetContentEncoding(std::string* encoding_type);
+  virtual bool GetContentEncodings(std::vector<std::string>* encoding_type);
   virtual bool IsRedirectResponse(GURL* location, int* http_status_code);
   virtual bool IsSafeRedirect(const GURL& location);
   virtual bool NeedsAuth();
@@ -77,12 +80,11 @@ class URLRequestHttpJob : public URLRequestJob {
 
   bool read_in_progress_;
 
-  // Keep a reference to the url request context to be sure it's not
-  // deleted before us.
+  // Keep a reference to the url request context to be sure it's not deleted
+  // before us.
   scoped_refptr<URLRequestContext> context_;
 
-  DISALLOW_EVIL_CONSTRUCTORS(URLRequestHttpJob);
+  DISALLOW_COPY_AND_ASSIGN(URLRequestHttpJob);
 };
 
-#endif  // NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H__
-
+#endif  // NET_URL_REQUEST_URL_REQUEST_HTTP_JOB_H_
diff --git a/net/url_request/url_request_job.cc b/net/url_request/url_request_job.cc
index b0517ba..0501286 100644
--- a/net/url_request/url_request_job.cc
+++ b/net/url_request/url_request_job.cc
@@ -47,11 +47,13 @@ void URLRequestJob::DetachRequest() {
 }
 
 void URLRequestJob::SetupFilter() {
-  std::string encoding_type;
-  if (GetContentEncoding(&encoding_type)) {
+  std::vector<std::string> encoding_types;
+  if (GetContentEncodings(&encoding_types)) {
     std::string mime_type;
     GetMimeType(&mime_type);
-    filter_.reset(Filter::Factory(encoding_type, mime_type, kFilterBufSize));
+    filter_.reset(Filter::Factory(encoding_types, mime_type, kFilterBufSize));
+    if (filter_.get())
+      filter_->SetURL(request_->url());
   }
 }
 
@@ -174,7 +176,7 @@ bool URLRequestJob::ReadFilteredData(int *bytes_read) {
     // Get filtered data
     int filtered_data_len = read_buffer_len_;
     Filter::FilterStatus status;
-    status = filter_->ReadFilteredData(read_buffer_, &filtered_data_len);
+    status = filter_->ReadData(read_buffer_, &filtered_data_len);
     switch (status) {
       case Filter::FILTER_DONE: {
         *bytes_read = filtered_data_len;
diff --git a/net/url_request/url_request_job.h b/net/url_request/url_request_job.h
index b70bbc7..1e4a089 100644
--- a/net/url_request/url_request_job.h
+++ b/net/url_request/url_request_job.h
@@ -5,6 +5,7 @@
 #ifndef NET_URL_REQUEST_URL_REQUEST_JOB_H_
 #define NET_URL_REQUEST_URL_REQUEST_JOB_H_
 
+#include <string>
 #include <vector>
 
 #include "base/basictypes.h"
@@ -28,7 +29,7 @@ class URLRequestJobMetrics;
 // UrlRequestFileJob.
 class URLRequestJob : public base::RefCountedThreadSafe<URLRequestJob> {
  public:
-  URLRequestJob(URLRequest* request);
+  explicit URLRequestJob(URLRequest* request);
   virtual ~URLRequestJob();
 
   // Returns the request that owns this job. THIS POINTER MAY BE NULL if the
@@ -44,7 +45,8 @@ class URLRequestJob : public base::RefCountedThreadSafe<URLRequestJob> {
   // Sets extra request headers for Job types that support request headers.
   virtual void SetExtraRequestHeaders(const std::string& headers) { }
 
-  // If any error occurs while starting the Job, NotifyStartError should be called.
+  // If any error occurs while starting the Job, NotifyStartError should be
+  // called.
   // This helps ensure that all errors follow more similar notification code
   // paths, which should simplify testing.
   virtual void Start() = 0;
@@ -108,10 +110,18 @@ class URLRequestJob : public base::RefCountedThreadSafe<URLRequestJob> {
   // Returns the HTTP response code for the request.
   virtual int GetResponseCode() { return -1; }
 
-  // Called to fetch the encoding type for this request. Only makes sense for
+  // Called to fetch the encoding types for this request. Only makes sense for
   // some types of requests. Returns true on success. Calling this on a request
   // that doesn't have or specify an encoding type will return false.
-  virtual bool GetContentEncoding(std::string* encoding_type) { return false; }
+  // Returns a array of strings showing the sequential encodings used on the
+  // content.  For example, types[0] = "sdch" and types[1] = gzip, means the
+  // content was first encoded by sdch, and then encoded by gzip.  To decode,
+  // a series of filters must be applied in the reverse order (in the above
+  // example, ungzip first, and then sdch expand).
+  // TODO(jar): Cleaner API would return an array of enums.
+  virtual bool GetContentEncodings(std::vector<std::string>* encoding_types) {
+    return false;
+  }
 
   // Called to setup stream filter for this request. An example of filter is
   // content encoding/decoding.