Add pause/resume functionality to precache

components/precache/core/precache_fetcher.cc

 // Copyright 2013 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.
 
 #include "components/precache/core/precache_fetcher.h"
 
 #include <algorithm>
 #include <limits>
 #include <string>
 #include <utility>
 #include <vector>
 
 #include "base/bind.h"
+#include "base/bind_helpers.h"
 #include "base/callback.h"
 #include "base/command_line.h"
 #include "base/compiler_specific.h"
 #include "base/containers/hash_tables.h"
+#include "base/location.h"
 #include "base/logging.h"
 #include "base/memory/ptr_util.h"
+#include "base/memory/ref_counted.h"
 #include "base/metrics/histogram_macros.h"
 #include "components/precache/core/precache_switches.h"
 #include "components/precache/core/proto/precache.pb.h"
+#include "components/precache/core/proto/unfinished_work.pb.h"
 #include "net/base/completion_callback.h"
 #include "net/base/escape.h"
 #include "net/base/io_buffer.h"
 #include "net/base/load_flags.h"
 #include "net/base/net_errors.h"
 #include "net/http/http_response_headers.h"
 #include "net/url_request/url_fetcher_response_writer.h"
 #include "net/url_request/url_request_context_getter.h"
 #include "net/url_request/url_request_status.h"
 
@@ skipping 217 matching lines @@
   // - The request was for a config or manifest.
   // - The resource was a cache hit without validators.
   // - The response came from the network.
   // Then Fetcher is done with this URL and can return control to the caller.
   response_bytes_ = source->GetReceivedResponseContentLength();
   network_response_bytes_ = source->GetTotalReceivedBytes();
   callback_.Run(*this);
 }
 
 PrecacheFetcher::PrecacheFetcher(
-    const std::vector<std::string>& starting_hosts,
     net::URLRequestContextGetter* request_context,
     const GURL& config_url,
     const std::string& manifest_url_prefix,
+    std::unique_ptr<PrecacheUnfinishedWork> unfinished_work,
     PrecacheFetcher::PrecacheDelegate* precache_delegate)
-    : starting_hosts_(starting_hosts),
-      request_context_(request_context),
+    : request_context_(request_context),
       config_url_(config_url),
       manifest_url_prefix_(manifest_url_prefix),
       precache_delegate_(precache_delegate),
       config_(new PrecacheConfigurationSettings),
       total_response_bytes_(0),
       network_response_bytes_(0),
       num_manifest_urls_to_fetch_(0),
       pool_(kMaxParallelFetches) {
   DCHECK(request_context_.get());  // Request context must be non-NULL.
   DCHECK(precache_delegate_);  // Precache delegate must be non-NULL.
 
   DCHECK_NE(GURL(), GetDefaultConfigURL())
       << "Could not determine the precache config settings URL.";
   DCHECK_NE(std::string(), GetDefaultManifestURLPrefix())
       << "Could not determine the default precache manifest URL prefix.";
+  DCHECK(unfinished_work);
+
+  for (const auto& manifest : unfinished_work->manifest()) {
+    if (manifest.has_url())
+      manifest_urls_to_fetch_.push_back(GURL(manifest.url()));
+  }
+  for (const auto& resource : unfinished_work->resource()) {
+    if (resource.has_url())
+      resource_urls_to_fetch_.push_back(GURL(resource.url()));
+  }
+  if (unfinished_work->has_total_bytes())
+    total_response_bytes_ = unfinished_work->total_bytes();
+  if (unfinished_work->has_network_bytes())
+    network_response_bytes_ = unfinished_work->network_bytes();
+  if (unfinished_work->has_num_manifest_urls())
+    num_manifest_urls_to_fetch_ = unfinished_work->num_manifest_urls();
+  if (unfinished_work->has_start_time())
+    start_time_ = base::Time::FromInternalValue(unfinished_work->start_time());
+
+  unfinished_work_ = std::move(unfinished_work);
+
 }
 
 PrecacheFetcher::~PrecacheFetcher() {
-  base::TimeDelta time_to_fetch = base::TimeTicks::Now() - start_time_;
+  base::TimeDelta time_to_fetch = base::Time::Now() - start_time_;
   UMA_HISTOGRAM_CUSTOM_TIMES("Precache.Fetch.TimeToComplete", time_to_fetch,
                              base::TimeDelta::FromSeconds(1),
                              base::TimeDelta::FromHours(4), 50);
 
   // Number of manifests for which we have downloaded all resources.
   int manifests_completed =
       num_manifest_urls_to_fetch_ - manifest_urls_to_fetch_.size();
 
   // If there are resource URLs left to fetch, the last manifest is not yet
   // completed.
   if (!resource_urls_to_fetch_.empty())
     --manifests_completed;
 
   DCHECK_GE(manifests_completed, 0);
   int percent_completed = num_manifest_urls_to_fetch_ == 0
                               ? 0
                               : (static_cast<double>(manifests_completed) /
                                  num_manifest_urls_to_fetch_ * 100);
   UMA_HISTOGRAM_PERCENTAGE("Precache.Fetch.PercentCompleted",

[sclittle, 2016/05/20 21:57:07]

Can you clean up this UMA like we discussed offline?

[bengr, 2016/05/21 00:16:32]

Done.

                            percent_completed);
   UMA_HISTOGRAM_CUSTOM_COUNTS("Precache.Fetch.ResponseBytes.Total",
                               total_response_bytes_, 1, kMaxResponseBytes, 100);
   UMA_HISTOGRAM_CUSTOM_COUNTS("Precache.Fetch.ResponseBytes.Network",
                               network_response_bytes_, 1, kMaxResponseBytes,
                               100);
 }
 
+std::unique_ptr<PrecacheUnfinishedWork> PrecacheFetcher::CancelPrecaching() {
+  for (const auto& manifest : manifest_urls_to_fetch_)
+    unfinished_work_->add_manifest()->set_url(manifest.spec());
+  for (const auto& resource : resource_urls_to_fetch_)
+    unfinished_work_->add_resource()->set_url(resource.spec());
+  unfinished_work_->set_total_bytes(total_response_bytes_);
+  unfinished_work_->set_network_bytes(network_response_bytes_);
+  unfinished_work_->set_num_manifest_urls(num_manifest_urls_to_fetch_);
+  unfinished_work_->set_start_time(start_time_.ToInternalValue());
+  return std::move(unfinished_work_);
+}
+
 void PrecacheFetcher::Start() {
+  if (unfinished_work_->has_config_settings()) {
+    DetermineManifests();
+    return;
+  }
+
   GURL config_url =
       config_url_.is_empty() ? GetDefaultConfigURL() : config_url_;
 
   DCHECK(config_url.is_valid()) << "Config URL not valid: "
                                 << config_url.possibly_invalid_spec();
 
-  start_time_ = base::TimeTicks::Now();
+  start_time_ = base::Time::Now();
 
   // Fetch the precache configuration settings from the server.
   DCHECK(pool_.IsEmpty()) << "All parallel requests should be available";
   VLOG(3) << "Fetching " << config_url;
   pool_.Add(base::WrapUnique(new Fetcher(
       request_context_.get(), config_url,
       base::Bind(&PrecacheFetcher::OnConfigFetchComplete,
                  base::Unretained(this)),
       false /* is_resource_request */, std::numeric_limits<int32_t>::max())));
 }
 
 void PrecacheFetcher::StartNextResourceFetch() {
+  DCHECK(unfinished_work_->has_config_settings());
   while (!resource_urls_to_fetch_.empty() && pool_.IsAvailable()) {
     const size_t max_bytes =
-        std::min(config_->max_bytes_per_resource(),
-                 config_->max_bytes_total() - total_response_bytes_);
+        std::min(unfinished_work_->config_settings().max_bytes_per_resource(),
+                 unfinished_work_->config_settings().max_bytes_total() -
+                     total_response_bytes_);
     VLOG(3) << "Fetching " << resource_urls_to_fetch_.front();
     pool_.Add(base::WrapUnique(
         new Fetcher(request_context_.get(), resource_urls_to_fetch_.front(),
                     base::Bind(&PrecacheFetcher::OnResourceFetchComplete,
                                base::Unretained(this)),
                     true /* is_resource_request */, max_bytes)));
 
     resource_urls_to_fetch_.pop_front();

[sclittle, 2016/05/20 21:57:07]

If precaching is cancelled, what happens to any resource fetches currently in
progress? From this, it looks like they'd be cancelled then never continued when
precaching is resumed.

[bengr, 2016/05/21 00:16:32]

Done.

   }
 }
 
 void PrecacheFetcher::StartNextManifestFetch() {
   if (manifest_urls_to_fetch_.empty())
     return;
 
   // We only fetch one manifest at a time to keep the size of
   // resource_urls_to_fetch_ as small as possible.
   DCHECK(pool_.IsAvailable())
       << "There are no available parallel requests to fetch the next manifest. "
          "Did you forget to call Delete?";
   VLOG(3) << "Fetching " << manifest_urls_to_fetch_.front();
   pool_.Add(base::WrapUnique(new Fetcher(
       request_context_.get(), manifest_urls_to_fetch_.front(),
       base::Bind(&PrecacheFetcher::OnManifestFetchComplete,
                  base::Unretained(this)),
       false /* is_resource_request */, std::numeric_limits<int32_t>::max())));
 
   manifest_urls_to_fetch_.pop_front();
 }
 
 void PrecacheFetcher::StartNextFetch() {
+  DCHECK(unfinished_work_->has_config_settings());
   // If over the precache total size cap, then stop prefetching.
-  if (total_response_bytes_ > config_->max_bytes_total()) {
+  if (total_response_bytes_ >
+      unfinished_work_->config_settings().max_bytes_total()) {
     precache_delegate_->OnDone();
     return;
   }
 
   StartNextResourceFetch();
   StartNextManifestFetch();
-
   if (pool_.IsEmpty()) {
     // There are no more URLs to fetch, so end the precache cycle.
     precache_delegate_->OnDone();
     // OnDone may have deleted this PrecacheFetcher, so don't do anything after
     // it is called.
   }
 }
 
 void PrecacheFetcher::OnConfigFetchComplete(const Fetcher& source) {
   UpdateStats(source.response_bytes(), source.network_response_bytes());
   if (source.network_url_fetcher() == nullptr) {
     pool_.DeleteAll();  // Cancel any other ongoing request.
   } else {
     // Attempt to parse the config proto. On failure, continue on with the
     // default configuration.
-    ParseProtoFromFetchResponse(*source.network_url_fetcher(), config_.get());
+    ParseProtoFromFetchResponse(
+        *source.network_url_fetcher(),
+        unfinished_work_->mutable_config_settings());
+    pool_.Delete(source);
+    DetermineManifests();
+  }
+}
 
+void PrecacheFetcher::DetermineManifests() {
+  DCHECK(unfinished_work_->has_config_settings());
     std::string prefix = manifest_url_prefix_.empty()
                              ? GetDefaultManifestURLPrefix()
                              : manifest_url_prefix_;
     DCHECK_NE(std::string(), prefix)
         << "Could not determine the precache manifest URL prefix.";
 
     // Keep track of manifest URLs that are being fetched, in order to elide
     // duplicates.
     base::hash_set<std::string> seen_manifest_urls;
 
     // Attempt to fetch manifests for starting hosts up to the maximum top sites
     // count. If a manifest does not exist for a particular starting host, then
-    // the fetch will fail, and that starting host will be ignored.
-    int64_t rank = 0;
-    for (const std::string& host : starting_hosts_) {
-      ++rank;
-      if (rank > config_->top_sites_count())
-        break;
-      AppendManifestURLIfNew(prefix, host, &seen_manifest_urls,
-                             &manifest_urls_to_fetch_);
+    // the fetch will fail, and that starting host will be ignored. Starting
+    // hosts are not added if this is a continuation from a previous precache
+    // session.
+    if (manifest_urls_to_fetch_.empty() &&
+        resource_urls_to_fetch_.empty()) {
+      int64_t rank = 0;
+      for (const auto& host : unfinished_work_->top_host()) {
+        ++rank;
+        if (rank > unfinished_work_->config_settings().top_sites_count())
+          break;
+        AppendManifestURLIfNew(prefix, host.hostname(), &seen_manifest_urls,
+                               &manifest_urls_to_fetch_);
+      }
+
+      for (const std::string& host
+          : unfinished_work_->config_settings().forced_site()) {
+        AppendManifestURLIfNew(prefix, host, &seen_manifest_urls,
+                               &manifest_urls_to_fetch_);
+      }
     }
-
-    for (const std::string& host : config_->forced_site())
-      AppendManifestURLIfNew(prefix, host, &seen_manifest_urls,
-                             &manifest_urls_to_fetch_);
-
     num_manifest_urls_to_fetch_ = manifest_urls_to_fetch_.size();
-  }
-  pool_.Delete(source);
-
-  StartNextFetch();
+    StartNextFetch();
 }
 
 void PrecacheFetcher::OnManifestFetchComplete(const Fetcher& source) {
+  DCHECK(unfinished_work_->has_config_settings());
   UpdateStats(source.response_bytes(), source.network_response_bytes());
   if (source.network_url_fetcher() == nullptr) {
     pool_.DeleteAll();  // Cancel any other ongoing request.
   } else {
     PrecacheManifest manifest;
 
     if (ParseProtoFromFetchResponse(*source.network_url_fetcher(), &manifest)) {
       const int32_t len =
-          std::min(manifest.resource_size(), config_->top_resources_count());
+          std::min(manifest.resource_size(),
+                   unfinished_work_->config_settings().top_resources_count());
       for (int i = 0; i < len; ++i) {
-        if (manifest.resource(i).has_url()) {
+        if (manifest.resource(i).has_url())
           resource_urls_to_fetch_.push_back(GURL(manifest.resource(i).url()));
-        }
       }
     }
   }
 
   pool_.Delete(source);
   StartNextFetch();
 }
 
 void PrecacheFetcher::OnResourceFetchComplete(const Fetcher& source) {
   UpdateStats(source.response_bytes(), source.network_response_bytes());
   pool_.Delete(source);
   // The resource has already been put in the cache during the fetch process, so
   // nothing more needs to be done for the resource.
   StartNextFetch();
 }
 
 void PrecacheFetcher::UpdateStats(int64_t response_bytes,
                                   int64_t network_response_bytes) {
   total_response_bytes_ += response_bytes;
   network_response_bytes_ += network_response_bytes;
 }
 
+void PrecacheFetcher::AssignWorkForTest(const std::list<GURL>& manifests,
+                                        const std::list<GURL>& resources,
+                                        size_t total_response_bytes,
+                                        size_t network_response_bytes,
+                                        size_t num_manifest_urls_to_fetch,
+                                        const base::Time& start_time) {
+  DCHECK(manifest_urls_to_fetch_.empty());
+  DCHECK(resource_urls_to_fetch_.empty());
+  manifest_urls_to_fetch_ = manifests;
+  resource_urls_to_fetch_ = resources;
+  total_response_bytes_ = total_response_bytes;
+  network_response_bytes_ = network_response_bytes;
+  num_manifest_urls_to_fetch_ = num_manifest_urls_to_fetch;
+  start_time_ = start_time;
+}
+
 }  // namespace precache