Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(149)

Side by Side Diff: content/browser/download/mhtml_generation_manager.cc

Issue 2683493002: Get signals working in the EXTRA_DATA section of MHTML (Closed)
Patch Set: Attempt to clear trybots. Created 3 years, 9 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View unified diff | Download patch
OLDNEW
1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. 1 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be 2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file. 3 // found in the LICENSE file.
4 4
5 #include "content/browser/download/mhtml_generation_manager.h" 5 #include "content/browser/download/mhtml_generation_manager.h"
6 6
7 #include <map> 7 #include <map>
8 #include <queue> 8 #include <queue>
9 #include <utility> 9 #include <utility>
10 10
11 #include "base/bind.h" 11 #include "base/bind.h"
12 #include "base/files/file.h" 12 #include "base/files/file.h"
13 #include "base/guid.h" 13 #include "base/guid.h"
14 #include "base/macros.h" 14 #include "base/macros.h"
15 #include "base/memory/ptr_util.h" 15 #include "base/memory/ptr_util.h"
16 #include "base/metrics/histogram_macros.h" 16 #include "base/metrics/histogram_macros.h"
17 #include "base/scoped_observer.h" 17 #include "base/scoped_observer.h"
18 #include "base/stl_util.h" 18 #include "base/stl_util.h"
19 #include "base/strings/string_util.h" 19 #include "base/strings/string_util.h"
20 #include "base/strings/stringprintf.h" 20 #include "base/strings/stringprintf.h"
21 #include "base/time/time.h" 21 #include "base/time/time.h"
22 #include "base/trace_event/trace_event.h" 22 #include "base/trace_event/trace_event.h"
23 #include "content/browser/bad_message.h" 23 #include "content/browser/bad_message.h"
24 #include "content/browser/download/mhtml_extra_data_impl.h"
24 #include "content/browser/frame_host/frame_tree_node.h" 25 #include "content/browser/frame_host/frame_tree_node.h"
25 #include "content/browser/frame_host/render_frame_host_impl.h" 26 #include "content/browser/frame_host/render_frame_host_impl.h"
26 #include "content/common/frame_messages.h" 27 #include "content/common/frame_messages.h"
27 #include "content/public/browser/browser_thread.h" 28 #include "content/public/browser/browser_thread.h"
29 #include "content/public/browser/mhtml_extra_data.h"
28 #include "content/public/browser/render_frame_host.h" 30 #include "content/public/browser/render_frame_host.h"
29 #include "content/public/browser/render_process_host.h" 31 #include "content/public/browser/render_process_host.h"
30 #include "content/public/browser/render_process_host_observer.h" 32 #include "content/public/browser/render_process_host_observer.h"
31 #include "content/public/browser/web_contents.h" 33 #include "content/public/browser/web_contents.h"
32 #include "content/public/common/mhtml_generation_params.h" 34 #include "content/public/common/mhtml_generation_params.h"
33 #include "net/base/mime_util.h" 35 #include "net/base/mime_util.h"
34 36
37 namespace {
38 const char kContentLocation[] = "Content-Location: ";
39 const char kContentType[] = "Content-Type: ";
40 } // namespace
41
35 namespace content { 42 namespace content {
36 43
37 // The class and all of its members live on the UI thread. Only static methods 44 // The class and all of its members live on the UI thread. Only static methods
38 // are executed on other threads. 45 // are executed on other threads.
39 class MHTMLGenerationManager::Job : public RenderProcessHostObserver { 46 class MHTMLGenerationManager::Job : public RenderProcessHostObserver {
40 public: 47 public:
41 Job(int job_id, 48 Job(int job_id,
42 WebContents* web_contents, 49 WebContents* web_contents,
43 const MHTMLGenerationParams& params, 50 const MHTMLGenerationParams& params,
44 const GenerateMHTMLCallback& callback); 51 const GenerateMHTMLCallback& callback);
(...skipping 49 matching lines...) Expand 10 before | Expand all | Expand 10 after
94 101
95 private: 102 private:
96 // Writes the MHTML footer to the file and closes it. 103 // Writes the MHTML footer to the file and closes it.
97 // 104 //
98 // Note: The same |boundary| marker must be used for all "boundaries" -- in 105 // Note: The same |boundary| marker must be used for all "boundaries" -- in
99 // the header, parts and footer -- that belong to the same MHTML document (see 106 // the header, parts and footer -- that belong to the same MHTML document (see
100 // also rfc1341, section 7.2.1, "boundary" description). 107 // also rfc1341, section 7.2.1, "boundary" description).
101 static std::tuple<MhtmlSaveStatus, int64_t> CloseFileOnFileThread( 108 static std::tuple<MhtmlSaveStatus, int64_t> CloseFileOnFileThread(
102 MhtmlSaveStatus save_status, 109 MhtmlSaveStatus save_status,
103 const std::string& boundary, 110 const std::string& boundary,
104 base::File file); 111 base::File file,
112 const MHTMLExtraDataImpl* extra_data);
105 void AddFrame(RenderFrameHost* render_frame_host); 113 void AddFrame(RenderFrameHost* render_frame_host);
106 114
115 // If we have any extra MHTML parts to write out, write them into the file
116 // while on the file thread. Returns true for success.
fgorski 2017/03/28 17:01:12 true for success or no data to write.
Pete Williamson 2017/03/28 18:01:58 Done.
117 static bool WriteExtraDataPartsOnFileThread(
118 const std::string& boundary,
119 base::File& file,
120 const MHTMLExtraDataImpl* extra_data);
121
107 // Creates a new map with values (content ids) the same as in 122 // Creates a new map with values (content ids) the same as in
108 // |frame_tree_node_to_content_id_| map, but with the keys translated from 123 // |frame_tree_node_to_content_id_| map, but with the keys translated from
109 // frame_tree_node_id into a |site_instance|-specific routing_id. 124 // frame_tree_node_id into a |site_instance|-specific routing_id.
110 std::map<int, std::string> CreateFrameRoutingIdToContentId( 125 std::map<int, std::string> CreateFrameRoutingIdToContentId(
111 SiteInstance* site_instance); 126 SiteInstance* site_instance);
112 127
113 // Id used to map renderer responses to jobs. 128 // Id used to map renderer responses to jobs.
114 // See also MHTMLGenerationManager::id_to_job_ map. 129 // See also MHTMLGenerationManager::id_to_job_ map.
115 const int job_id_; 130 const int job_id_;
116 131
(...skipping 32 matching lines...) Expand 10 before | Expand all | Expand 10 after
149 std::string salt_; 164 std::string salt_;
150 165
151 // The callback to call once generation is complete. 166 // The callback to call once generation is complete.
152 const GenerateMHTMLCallback callback_; 167 const GenerateMHTMLCallback callback_;
153 168
154 // Whether the job is finished (set to true only for the short duration of 169 // Whether the job is finished (set to true only for the short duration of
155 // time between MHTMLGenerationManager::JobFinished is called and the job is 170 // time between MHTMLGenerationManager::JobFinished is called and the job is
156 // destroyed by MHTMLGenerationManager::OnFileClosed). 171 // destroyed by MHTMLGenerationManager::OnFileClosed).
157 bool is_finished_; 172 bool is_finished_;
158 173
174 // Any extra data parts that should be emitted into the output MHTML.
175 MHTMLExtraDataImpl* extra_data_;
176
159 // RAII helper for registering this Job as a RenderProcessHost observer. 177 // RAII helper for registering this Job as a RenderProcessHost observer.
160 ScopedObserver<RenderProcessHost, MHTMLGenerationManager::Job> 178 ScopedObserver<RenderProcessHost, MHTMLGenerationManager::Job>
161 observed_renderer_process_host_; 179 observed_renderer_process_host_;
162 180
163 DISALLOW_COPY_AND_ASSIGN(Job); 181 DISALLOW_COPY_AND_ASSIGN(Job);
164 }; 182 };
165 183
166 MHTMLGenerationManager::Job::Job(int job_id, 184 MHTMLGenerationManager::Job::Job(int job_id,
167 WebContents* web_contents, 185 WebContents* web_contents,
168 const MHTMLGenerationParams& params, 186 const MHTMLGenerationParams& params,
169 const GenerateMHTMLCallback& callback) 187 const GenerateMHTMLCallback& callback)
170 : job_id_(job_id), 188 : job_id_(job_id),
171 creation_time_(base::TimeTicks::Now()), 189 creation_time_(base::TimeTicks::Now()),
172 params_(params), 190 params_(params),
173 frame_tree_node_id_of_busy_frame_(FrameTreeNode::kFrameTreeNodeInvalidId), 191 frame_tree_node_id_of_busy_frame_(FrameTreeNode::kFrameTreeNodeInvalidId),
174 mhtml_boundary_marker_(net::GenerateMimeMultipartBoundary()), 192 mhtml_boundary_marker_(net::GenerateMimeMultipartBoundary()),
175 salt_(base::GenerateGUID()), 193 salt_(base::GenerateGUID()),
176 callback_(callback), 194 callback_(callback),
177 is_finished_(false), 195 is_finished_(false),
178 observed_renderer_process_host_(this) { 196 observed_renderer_process_host_(this) {
179 DCHECK_CURRENTLY_ON(BrowserThread::UI); 197 DCHECK_CURRENTLY_ON(BrowserThread::UI);
180 web_contents->ForEachFrame(base::Bind( 198 web_contents->ForEachFrame(base::Bind(
181 &MHTMLGenerationManager::Job::AddFrame, 199 &MHTMLGenerationManager::Job::AddFrame,
182 base::Unretained(this))); // Safe because ForEachFrame is synchronous. 200 base::Unretained(this))); // Safe because ForEachFrame is synchronous.
183 201
184 // Main frame needs to be processed first. 202 // Main frame needs to be processed first.
185 DCHECK(!pending_frame_tree_node_ids_.empty()); 203 DCHECK(!pending_frame_tree_node_ids_.empty());
186 DCHECK(FrameTreeNode::GloballyFindByID(pending_frame_tree_node_ids_.front()) 204 DCHECK(FrameTreeNode::GloballyFindByID(pending_frame_tree_node_ids_.front())
187 ->parent() == nullptr); 205 ->parent() == nullptr);
206
207 // Save off any extra data.
208 extra_data_ = static_cast<MHTMLExtraDataImpl*>(
209 MHTMLExtraData::FromWebContents(web_contents));
188 } 210 }
189 211
190 MHTMLGenerationManager::Job::~Job() { 212 MHTMLGenerationManager::Job::~Job() {
191 DCHECK_CURRENTLY_ON(BrowserThread::UI); 213 DCHECK_CURRENTLY_ON(BrowserThread::UI);
192 } 214 }
193 215
194 std::map<int, std::string> 216 std::map<int, std::string>
195 MHTMLGenerationManager::Job::CreateFrameRoutingIdToContentId( 217 MHTMLGenerationManager::Job::CreateFrameRoutingIdToContentId(
196 SiteInstance* site_instance) { 218 SiteInstance* site_instance) {
197 std::map<int, std::string> result; 219 std::map<int, std::string> result;
(...skipping 150 matching lines...) Expand 10 before | Expand all | Expand 10 after
348 return; 370 return;
349 } 371 }
350 372
351 // If no previous error occurred the boundary should be sent. 373 // If no previous error occurred the boundary should be sent.
352 BrowserThread::PostTaskAndReplyWithResult( 374 BrowserThread::PostTaskAndReplyWithResult(
353 BrowserThread::FILE, FROM_HERE, 375 BrowserThread::FILE, FROM_HERE,
354 base::Bind( 376 base::Bind(
355 &MHTMLGenerationManager::Job::CloseFileOnFileThread, save_status, 377 &MHTMLGenerationManager::Job::CloseFileOnFileThread, save_status,
356 (save_status == MhtmlSaveStatus::SUCCESS ? mhtml_boundary_marker_ 378 (save_status == MhtmlSaveStatus::SUCCESS ? mhtml_boundary_marker_
357 : std::string()), 379 : std::string()),
358 base::Passed(&browser_file_)), 380 base::Passed(&browser_file_), extra_data_),
359 callback); 381 callback);
360 } 382 }
361 383
362 bool MHTMLGenerationManager::Job::IsMessageFromFrameExpected( 384 bool MHTMLGenerationManager::Job::IsMessageFromFrameExpected(
363 RenderFrameHostImpl* sender) { 385 RenderFrameHostImpl* sender) {
364 int sender_id = sender->frame_tree_node()->frame_tree_node_id(); 386 int sender_id = sender->frame_tree_node()->frame_tree_node_id();
365 if (sender_id != frame_tree_node_id_of_busy_frame_) 387 if (sender_id != frame_tree_node_id_of_busy_frame_)
366 return false; 388 return false;
367 389
368 // We only expect one message per frame - let's make sure subsequent messages 390 // We only expect one message per frame - let's make sure subsequent messages
(...skipping 25 matching lines...) Expand all
394 416
395 // Report success if all frames have been processed. 417 // Report success if all frames have been processed.
396 if (pending_frame_tree_node_ids_.empty()) 418 if (pending_frame_tree_node_ids_.empty())
397 return MhtmlSaveStatus::SUCCESS; 419 return MhtmlSaveStatus::SUCCESS;
398 420
399 return SendToNextRenderFrame(); 421 return SendToNextRenderFrame();
400 } 422 }
401 423
402 // static 424 // static
403 std::tuple<MhtmlSaveStatus, int64_t> 425 std::tuple<MhtmlSaveStatus, int64_t>
404 MHTMLGenerationManager::Job::CloseFileOnFileThread(MhtmlSaveStatus save_status, 426 MHTMLGenerationManager::Job::CloseFileOnFileThread(
405 const std::string& boundary, 427 MhtmlSaveStatus save_status,
406 base::File file) { 428 const std::string& boundary,
429 base::File file,
430 const MHTMLExtraDataImpl* extra_data) {
407 DCHECK_CURRENTLY_ON(BrowserThread::FILE); 431 DCHECK_CURRENTLY_ON(BrowserThread::FILE);
432 int64_t file_size = -1;
408 433
409 // If no previous error occurred the boundary should have been provided. 434 // If no previous error occurred the boundary should have been provided.
410 if (save_status == MhtmlSaveStatus::SUCCESS) { 435 if (save_status == MhtmlSaveStatus::SUCCESS) {
411 TRACE_EVENT0("page-serialization", 436 TRACE_EVENT0("page-serialization",
412 "MHTMLGenerationManager::Job MHTML footer writing"); 437 "MHTMLGenerationManager::Job MHTML footer writing");
413 DCHECK(!boundary.empty()); 438 DCHECK(!boundary.empty());
439
440 // Write the extra data into a part of its own, if we have any.
441 if (!WriteExtraDataPartsOnFileThread(boundary, file, extra_data)) {
442 save_status = MhtmlSaveStatus::FILE_WRITTING_ERROR;
443 return std::make_tuple(save_status, file_size);
fgorski 2017/03/28 17:01:12 -1 would work better here. (Or some constant). Thi
Pete Williamson 2017/03/28 18:01:58 Done.
444 }
445
414 std::string footer = base::StringPrintf("--%s--\r\n", boundary.c_str()); 446 std::string footer = base::StringPrintf("--%s--\r\n", boundary.c_str());
415 DCHECK(base::IsStringASCII(footer)); 447 DCHECK(base::IsStringASCII(footer));
416 if (file.WriteAtCurrentPos(footer.data(), footer.size()) < 0) 448 if (file.WriteAtCurrentPos(footer.data(), footer.size()) < 0) {
417 save_status = MhtmlSaveStatus::FILE_WRITTING_ERROR; 449 save_status = MhtmlSaveStatus::FILE_WRITTING_ERROR;
450 return std::make_tuple(save_status, file_size);
451 }
418 } 452 }
419 453
420 // If the file is still valid try to close it. Only update the status if that 454 // If the file is still valid try to close it. Only update the status if that
421 // won't hide an earlier error. 455 // won't hide an earlier error.
422 int64_t file_size = -1;
423 if (file.IsValid()) { 456 if (file.IsValid()) {
424 file_size = file.GetLength(); 457 file_size = file.GetLength();
425 file.Close(); 458 file.Close();
426 } else if (save_status == MhtmlSaveStatus::SUCCESS) { 459 } else if (save_status == MhtmlSaveStatus::SUCCESS) {
427 save_status = MhtmlSaveStatus::FILE_CLOSING_ERROR; 460 save_status = MhtmlSaveStatus::FILE_CLOSING_ERROR;
428 } 461 }
429 462
430 return std::make_tuple(save_status, file_size); 463 return std::make_tuple(save_status, file_size);
431 } 464 }
432 465
466 // static
467 bool MHTMLGenerationManager::Job::WriteExtraDataPartsOnFileThread(
468 const std::string& boundary,
469 base::File& file,
470 const MHTMLExtraDataImpl* extra_data) {
fgorski 2017/03/28 17:01:12 DCHECK thread.
Pete Williamson 2017/03/28 18:01:58 Done.
471 // Don't write an extra data part if there is none.
472 if (extra_data == nullptr)
473 return true;
474
475 const std::vector<MHTMLExtraDataPart>& extra_parts(extra_data->parts());
476 if (extra_parts.empty())
477 return true;
478
479 std::string serialized_extra_data_parts;
480
481 // For each extra part, serialize that part and add to our accumulator
482 // string.
483 for (auto part : extra_parts) {
484 // Write a newline, then a boundary, another newline, then the content
485 // location, another newline, the content type, another newline, the extra
486 // data string, and end with a newline.
487 std::string serialized_extra_data_part = base::StringPrintf(
488 "--%s--\r\n%s%s\r\n%s%s\r\n%s\r\n", boundary.c_str(), kContentLocation,
489 part.content_location.c_str(), kContentType, part.content_type.c_str(),
490 part.body.c_str());
491 DCHECK(base::IsStringASCII(serialized_extra_data_part));
492
493 serialized_extra_data_parts += serialized_extra_data_part;
494 }
495
496 // Write the string into the file. Returns false if we failed the write.
497 return (file.WriteAtCurrentPos(serialized_extra_data_parts.data(),
498 serialized_extra_data_parts.size()) >= 0);
499 }
500
433 MHTMLGenerationManager* MHTMLGenerationManager::GetInstance() { 501 MHTMLGenerationManager* MHTMLGenerationManager::GetInstance() {
434 return base::Singleton<MHTMLGenerationManager>::get(); 502 return base::Singleton<MHTMLGenerationManager>::get();
435 } 503 }
436 504
437 MHTMLGenerationManager::MHTMLGenerationManager() : next_job_id_(0) {} 505 MHTMLGenerationManager::MHTMLGenerationManager() : next_job_id_(0) {}
438 506
439 MHTMLGenerationManager::~MHTMLGenerationManager() { 507 MHTMLGenerationManager::~MHTMLGenerationManager() {
440 } 508 }
441 509
442 void MHTMLGenerationManager::SaveMHTML(WebContents* web_contents, 510 void MHTMLGenerationManager::SaveMHTML(WebContents* web_contents,
(...skipping 149 matching lines...) Expand 10 before | Expand all | Expand 10 after
592 return iter->second.get(); 660 return iter->second.get();
593 } 661 }
594 662
595 void MHTMLGenerationManager::RenderProcessExited(Job* job) { 663 void MHTMLGenerationManager::RenderProcessExited(Job* job) {
596 DCHECK_CURRENTLY_ON(BrowserThread::UI); 664 DCHECK_CURRENTLY_ON(BrowserThread::UI);
597 DCHECK(job); 665 DCHECK(job);
598 JobFinished(job, MhtmlSaveStatus::RENDER_PROCESS_EXITED); 666 JobFinished(job, MhtmlSaveStatus::RENDER_PROCESS_EXITED);
599 } 667 }
600 668
601 } // namespace content 669 } // namespace content
OLDNEW

Powered by Google App Engine
This is Rietveld 408576698