Improve linearized pdf load/show time.

pdf/pdfium/pdfium_engine.cc

 // Copyright (c) 2012 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 "pdf/pdfium/pdfium_engine.h"
 
 #include <math.h>
 #include <stddef.h>
 #include <stdint.h>
 
@@ skipping 12 matching lines @@
 #include "base/strings/string_piece.h"
 #include "base/strings/string_util.h"
 #include "base/strings/utf_string_conversions.h"
 #include "gin/array_buffer.h"
 #include "gin/public/gin_embedders.h"
 #include "gin/public/isolate_holder.h"
 #include "pdf/draw_utils.h"
 #include "pdf/pdfium/pdfium_api_string_buffer_adapter.h"
 #include "pdf/pdfium/pdfium_mem_buffer_file_read.h"
 #include "pdf/pdfium/pdfium_mem_buffer_file_write.h"
+#include "pdf/url_loader_wrapper_impl.h"
 #include "ppapi/c/pp_errors.h"
 #include "ppapi/c/pp_input_event.h"
 #include "ppapi/c/ppb_core.h"
 #include "ppapi/c/private/ppb_pdf.h"
 #include "ppapi/cpp/dev/memory_dev.h"
 #include "ppapi/cpp/input_event.h"
 #include "ppapi/cpp/instance.h"
 #include "ppapi/cpp/module.h"
 #include "ppapi/cpp/private/pdf.h"
 #include "ppapi/cpp/trusted/browser_font_trusted.h"
@@ skipping 482 matching lines @@
 }
 
 PDFEngine* PDFEngine::Create(PDFEngine::Client* client) {
   return new PDFiumEngine(client);
 }
 
 PDFiumEngine::PDFiumEngine(PDFEngine::Client* client)
     : client_(client),
       current_zoom_(1.0),
       current_rotation_(0),
-      doc_loader_(this),
       password_tries_remaining_(0),
       doc_(nullptr),
       form_(nullptr),
       defer_page_unload_(false),
       selecting_(false),
       mouse_down_state_(PDFiumPage::NONSELECTABLE_AREA,
                         PDFiumPage::LinkTarget()),
       next_page_to_search_(-1),
       last_page_to_search_(-1),
       last_character_index_to_search_(-1),
       permissions_(0),
       permissions_handler_revision_(-1),
       fpdf_availability_(nullptr),
       next_timer_id_(0),
       last_page_mouse_down_(-1),
       most_visible_page_(-1),
       called_do_document_action_(false),
       render_grayscale_(false),
       render_annots_(true),
       progressive_paint_timeout_(0),
       getting_password_(false) {
   find_factory_.Initialize(this);
   password_factory_.Initialize(this);
 
   file_access_.m_FileLen = 0;
   file_access_.m_GetBlock = &GetBlock;
-  file_access_.m_Param = &doc_loader_;
+  file_access_.m_Param = this;
 
   file_availability_.version = 1;
   file_availability_.IsDataAvail = &IsDataAvail;
-  file_availability_.loader = &doc_loader_;
+  file_availability_.engine = this;
 
   download_hints_.version = 1;
   download_hints_.AddSegment = &AddSegment;
-  download_hints_.loader = &doc_loader_;
+  download_hints_.engine = this;
 
   // Initialize FPDF_FORMFILLINFO member variables.  Deriving from this struct
   // allows the static callbacks to be able to cast the FPDF_FORMFILLINFO in
   // callbacks to ourself instead of maintaining a map of them to
   // PDFiumEngine.
   FPDF_FORMFILLINFO::version = 1;
   FPDF_FORMFILLINFO::m_pJsPlatform = this;
   FPDF_FORMFILLINFO::Release = nullptr;
   FPDF_FORMFILLINFO::FFI_Invalidate = Form_Invalidate;
   FPDF_FORMFILLINFO::FFI_OutputSelectedRect = Form_OutputSelectedRect;
@@ skipping 251 matching lines @@
 int PDFiumEngine::Form_GetLanguage(FPDF_FORMFILLINFO* param,
                                    void* language,
                                    int length) {
   return 0;
 }
 
 #endif  // defined(PDF_ENABLE_XFA)
 
 int PDFiumEngine::GetBlock(void* param, unsigned long position,
                            unsigned char* buffer, unsigned long size) {
-  DocumentLoader* loader = static_cast<DocumentLoader*>(param);
-  return loader->GetBlock(position, size, buffer);
+  PDFiumEngine* p_this = static_cast<PDFiumEngine*>(param);

[Lei Zhang, 2016/10/05 07:18:08]

s/p_this/engine/ ?

[snake, 2016/10/05 14:14:10]

Done.

+  return p_this->doc_loader_->GetBlock(position, size, buffer);
 }
 
 FPDF_BOOL PDFiumEngine::IsDataAvail(FX_FILEAVAIL* param,
                                     size_t offset, size_t size) {
   PDFiumEngine::FileAvail* file_avail =
       static_cast<PDFiumEngine::FileAvail*>(param);
-  return file_avail->loader->IsDataAvailable(offset, size);
+  return file_avail->engine->doc_loader_->IsDataAvailable(offset, size);
 }
 
 void PDFiumEngine::AddSegment(FX_DOWNLOADHINTS* param,
                               size_t offset, size_t size) {
   PDFiumEngine::DownloadHints* download_hints =
       static_cast<PDFiumEngine::DownloadHints*>(param);
-  return download_hints->loader->RequestData(offset, size);
+  return download_hints->engine->doc_loader_->RequestData(offset, size);
 }
 
 bool PDFiumEngine::New(const char* url,
                        const char* headers) {
   url_ = url;
   if (headers)
     headers_ = headers;
   else
     headers_.clear();
   return true;
@@ skipping 114 matching lines @@
     FPDF_RenderPage_Close(
         pages_[progressive_paints_[i].page_index]->GetPage());
     FPDFBitmap_Destroy(progressive_paints_[i].bitmap);
     progressive_paints_.erase(progressive_paints_.begin() + i);
     --i;
   }
 }
 
 bool PDFiumEngine::HandleDocumentLoad(const pp::URLLoader& loader) {
   password_tries_remaining_ = kMaxPasswordTries;
-  return doc_loader_.Init(loader, url_, headers_);
+  should_process_pending_request_complete_ = true;
+  URLLoaderWrapperImpl* loader_wrapper_ptr =

[Lei Zhang, 2016/10/05 07:18:07]

Why bother with the raw pointer? Just:

auto loader_wrapper = base::MakeUnique<
URLLoaderWrapperImpl>(GetPluginInstance(), loader);
loader_wrapper->SetResponseHeaders(headers_);
...

[snake, 2016/10/05 14:14:10]

Done.

+      new URLLoaderWrapperImpl(GetPluginInstance(), loader);
+  loader_wrapper_ptr->SetResponseHeaders(headers_);
+  std::unique_ptr<URLLoaderWrapper> loader_wrapper(loader_wrapper_ptr);
+  doc_loader_.reset(new DocumentLoader(this));

[Lei Zhang, 2016/10/05 07:18:08]

More base::MakeUnique?

[snake, 2016/10/05 14:14:10]

Done.

+  if (doc_loader_->Init(std::move(loader_wrapper), url_)) {
+    // request initial data.
+    doc_loader_->RequestData(0, 1);
+    return true;
+  }
+  return false;
 }
 
 pp::Instance* PDFiumEngine::GetPluginInstance() {
   return client_->GetPluginInstance();
 }
 
-pp::URLLoader PDFiumEngine::CreateURLLoader() {
-  return client_->CreateURLLoader();
+std::unique_ptr<URLLoaderWrapper> PDFiumEngine::CreateURLLoader() {
+  std::unique_ptr<URLLoaderWrapper> result;
+  result.reset(new URLLoaderWrapperImpl(GetPluginInstance(),
+                                        client_->CreateURLLoader()));
+
+  return result;
 }
 
 void PDFiumEngine::AppendPage(PDFEngine* engine, int index) {
   // Unload and delete the blank page before appending.
   pages_[index]->Unload();
   pages_[index]->set_calculated_links(false);
   pp::Size curr_page_size = GetPageSize(index);
   FPDFPage_Delete(doc_, index);
   FPDF_ImportPages(doc_,
                    static_cast<PDFiumEngine*>(engine)->doc(),
                    "1",
                    index);
   pp::Size new_page_size = GetPageSize(index);
   if (curr_page_size != new_page_size)
     LoadPageInfo(true);
   client_->Invalidate(GetPageScreenRect(index));
 }
 
 #if defined(PDF_ENABLE_XFA)
 void PDFiumEngine::SetScrollPosition(const pp::Point& position) {
   position_ = position;
 }
 #endif
 
-bool PDFiumEngine::IsProgressiveLoad() {
-  return doc_loader_.is_partial_document();
-}
-
 std::string PDFiumEngine::GetMetadata(const std::string& key) {
   return GetDocumentMetadata(doc(), key);
 }
 
-void PDFiumEngine::OnPartialDocumentLoaded() {
-  file_access_.m_FileLen = doc_loader_.document_size();
+void PDFiumEngine::OnPendingRequestComplete() {
+  if (!should_process_pending_request_complete_)
+    return;
   if (!fpdf_availability_) {
+    file_access_.m_FileLen = doc_loader_->GetDocumentSize();
     fpdf_availability_ = FPDFAvail_Create(&file_availability_, &file_access_);
     DCHECK(fpdf_availability_);
+    // Currently engine does not deal efficiently with some non-linearized
+    // files.
+    // See http://code.google.com/p/chromium/issues/detail?id=59400
+    // To improve user experience we download entire file for non-linearized
+    // PDF.
+    if (FPDFAvail_IsLinearized(fpdf_availability_) != PDF_LINEARIZED) {
+      // Wait complete document.
+      should_process_pending_request_complete_ = false;
+      FPDFAvail_Destroy(fpdf_availability_);
+      fpdf_availability_ = nullptr;
+      return;
+    }
   }
 
-  // Currently engine does not deal efficiently with some non-linearized files.
-  // See http://code.google.com/p/chromium/issues/detail?id=59400
-  // To improve user experience we download entire file for non-linearized PDF.
-  if (!FPDFAvail_IsLinearized(fpdf_availability_)) {
-    doc_loader_.RequestData(0, doc_loader_.document_size());
-    return;
-  }
-
-  LoadDocument();
-}
-
-void PDFiumEngine::OnPendingRequestComplete() {
-  if (!doc_ || !form_) {
-    DCHECK(fpdf_availability_);
+  if (!doc_) {
     LoadDocument();
     return;
   }
 
   // LoadDocument() will result in |pending_pages_| being reset so there's no
   // need to run the code below in that case.
   bool update_pages = false;
   std::vector<int> still_pending;
   for (int pending_page : pending_pages_) {
     if (CheckPageAvailable(pending_page, &still_pending)) {
       update_pages = true;
       if (IsPageVisible(pending_page))
         client_->Invalidate(GetPageScreenRect(pending_page));
     }
   }
   pending_pages_.swap(still_pending);
   if (update_pages)
     LoadPageInfo(true);
 }
 
 void PDFiumEngine::OnNewDataAvailable() {
-  client_->DocumentLoadProgress(doc_loader_.GetAvailableData(),
-                                doc_loader_.document_size());
+  const float progress = doc_loader_->GetProgress();
+  if (progress < 0.001) {
+    client_->DocumentLoadProgress(0, 0);
+  } else {
+    client_->DocumentLoadProgress(progress * 10000, 10000);
+  }
 }
 
 void PDFiumEngine::OnDocumentComplete() {
-  if (!doc_ || !form_) {
-    file_access_.m_FileLen = doc_loader_.document_size();
+  if (!doc_) {

[Lei Zhang, 2016/10/05 07:18:08]

In general, if (!foo) { bar(); } else { qux(); } may make more sense when
flipped back to if (foo) { qux(); } else { bar(); };

[snake, 2016/10/05 14:14:10]

Done.

+    file_access_.m_FileLen = doc_loader_->GetDocumentSize();
     if (!fpdf_availability_) {
       fpdf_availability_ = FPDFAvail_Create(&file_availability_, &file_access_);
       DCHECK(fpdf_availability_);
     }
     LoadDocument();
-    return;
+  } else {
+    FinishLoadingDocument();

[Lei Zhang, 2016/10/05 07:18:07]

Or how about just:

if (doc_) {
  FinishLoadingDocument();
  return;
}

....

[snake, 2016/10/05 14:14:10]

Done.

   }
+}
 
-  FinishLoadingDocument();
+void PDFiumEngine::OnDocumentCanceled() {
+  OnDocumentComplete();
+}
+
+void PDFiumEngine::CancelBrowserDownload() {
+  client_->CancelBrowserDownload();
 }
 
 void PDFiumEngine::FinishLoadingDocument() {
-  DCHECK(doc_loader_.IsDocumentComplete() && doc_);
+  DCHECK(doc_loader_->IsDocumentComplete() && doc_);
+
+  if (!form_) {
+    int form_status =
+        FPDFAvail_IsFormAvail(fpdf_availability_, &download_hints_);
+    if (form_status != PDF_FORM_NOTAVAIL) {
+      form_ = FPDFDOC_InitFormFillEnvironment(
+          doc_, static_cast<FPDF_FORMFILLINFO*>(this));
+#if defined(PDF_ENABLE_XFA)
+      FPDF_LoadXFA(doc_);
+#endif
+
+      FPDF_SetFormFieldHighlightColor(form_, 0, kFormHighlightColor);
+      FPDF_SetFormFieldHighlightAlpha(form_, kFormHighlightAlpha);
+    }
+  }
 
   bool need_update = false;
   for (size_t i = 0; i < pages_.size(); ++i) {
     if (pages_[i]->available())
       continue;
 
     pages_[i]->set_available(true);
     // We still need to call IsPageAvail() even if the whole document is
     // already downloaded.
     FPDFAvail_IsPageAvail(fpdf_availability_, i, &download_hints_);
@@ skipping 186 matching lines @@
   return temp_doc;
 }
 
 pp::Buffer_Dev PDFiumEngine::PrintPagesAsRasterPDF(
     const PP_PrintPageNumberRange_Dev* page_ranges, uint32_t page_range_count,
     const PP_PrintSettings_Dev& print_settings) {
   if (!page_range_count)
     return pp::Buffer_Dev();
 
   // If document is not downloaded yet, disable printing.
-  if (doc_ && !doc_loader_.IsDocumentComplete())
+  if (doc_ && !doc_loader_->IsDocumentComplete())
     return pp::Buffer_Dev();
 
   FPDF_DOCUMENT output_doc = FPDF_CreateNewDocument();
   if (!output_doc)
     return pp::Buffer_Dev();
 
   SaveSelectedFormForPrint();
 
   std::vector<PDFiumPage> pages_to_print;
   // width and height of source PDF pages.
@@ skipping 1117 matching lines @@
   }
 
   CalculateVisiblePages();
   if (document_size_ != old_document_size)
     client_->DocumentSizeUpdated(document_size_);
 }
 
 void PDFiumEngine::LoadDocument() {
   // Check if the document is ready for loading. If it isn't just bail for now,
   // we will call LoadDocument() again later.
-  if (!doc_ && !doc_loader_.IsDocumentComplete() &&
+  if (!doc_ && !doc_loader_->IsDocumentComplete() &&
       !FPDFAvail_IsDocAvail(fpdf_availability_, &download_hints_)) {
     return;
   }
 
   // If we're in the middle of getting a password, just return. We will retry
   // loading the document after we get the password anyway.
   if (getting_password_)
     return;
 
   ScopedUnsupportedFeature scoped_unsupported_feature(this);
@@ skipping 17 matching lines @@
     // call may have failed, so call it again for good measure.
     FPDFAvail_IsDocAvail(fpdf_availability_, &download_hints_);
     return true;
   }
 
   const char* password_cstr = nullptr;
   if (!password.empty()) {
     password_cstr = password.c_str();
     password_tries_remaining_--;
   }
-  if (doc_loader_.IsDocumentComplete() &&
+  if (doc_loader_->IsDocumentComplete() &&
       !FPDFAvail_IsLinearized(fpdf_availability_)) {
     doc_ = FPDF_LoadCustomDocument(&file_access_, password_cstr);
   } else {
     doc_ = FPDFAvail_GetDocument(fpdf_availability_, password_cstr);
   }
   if (!doc_) {
     if (FPDF_GetLastError() == FPDF_ERR_PASSWORD)
       *needs_password = true;
     return false;
   }
@@ skipping 36 matching lines @@
     client_->DocumentLoadFailed();
     return;
   }
 
   if (FPDFDoc_GetPageMode(doc_) == PAGEMODE_USEOUTLINES)
     client_->DocumentHasUnsupportedFeature("Bookmarks");
 
   permissions_ = FPDF_GetDocPermissions(doc_);
   permissions_handler_revision_ = FPDF_GetSecurityHandlerRevision(doc_);
 
-  if (!form_) {
-    int form_status =
-        FPDFAvail_IsFormAvail(fpdf_availability_, &download_hints_);
-    bool doc_complete = doc_loader_.IsDocumentComplete();
-    // Try again if the data is not available and the document hasn't finished
-    // downloading.
-    if (form_status == PDF_FORM_NOTAVAIL && !doc_complete)
-      return;
-
-    form_ = FPDFDOC_InitFormFillEnvironment(
-        doc_, static_cast<FPDF_FORMFILLINFO*>(this));
-#if defined(PDF_ENABLE_XFA)
-    FPDF_LoadXFA(doc_);
-#endif
-
-    FPDF_SetFormFieldHighlightColor(form_, 0, kFormHighlightColor);
-    FPDF_SetFormFieldHighlightAlpha(form_, kFormHighlightAlpha);
-  }
-
-  if (!doc_loader_.IsDocumentComplete()) {
+  if (!doc_loader_->IsDocumentComplete()) {
     // Check if the first page is available.  In a linearized PDF, that is not
     // always page 0.  Doing this gives us the default page size, since when the
     // document is available, the first page is available as well.
     CheckPageAvailable(FPDFAvail_GetFirstPageNum(doc_), &pending_pages_);
   }
 
   LoadPageInfo(false);
 
-  if (doc_loader_.IsDocumentComplete())
+  if (doc_loader_->IsDocumentComplete())
     FinishLoadingDocument();
 }
 
 void PDFiumEngine::LoadPageInfo(bool reload) {
   pending_pages_.clear();
   pp::Size old_document_size = document_size_;
   document_size_ = pp::Size();
   std::vector<pp::Rect> page_rects;
   int page_count = FPDF_GetPageCount(doc_);
-  bool doc_complete = doc_loader_.IsDocumentComplete();
+  bool doc_complete = doc_loader_->IsDocumentComplete();
   bool is_linear = FPDFAvail_IsLinearized(fpdf_availability_) == PDF_LINEARIZED;
   for (int i = 0; i < page_count; ++i) {
     if (i != 0) {
       // Add space for horizontal separator.
       document_size_.Enlarge(0, kPageSeparatorThickness);
     }
 
     // Get page availability. If |reload| == true, then the document has been
     // constructed already. Get page availability flag from already existing
     // PDFiumPage class.
@@ skipping 36 matching lines @@
       pages_.push_back(new PDFiumPage(this, i, page_rect, false));
     }
   }
 
   CalculateVisiblePages();
   if (document_size_ != old_document_size)
     client_->DocumentSizeUpdated(document_size_);
 }
 
 void PDFiumEngine::CalculateVisiblePages() {
+  if (!doc_loader_)
+    return;
   // Clear pending requests queue, since it may contain requests to the pages
   // that are already invisible (after scrolling for example).
   pending_pages_.clear();
-  doc_loader_.ClearPendingRequests();
+  doc_loader_->ClearPendingRequests();
 
   visible_pages_.clear();
   pp::Rect visible_rect(plugin_size_);
   for (size_t i = 0; i < pages_.size(); ++i) {
     // Check an entire PageScreenRect, since we might need to repaint side
     // borders and shadows even if the page itself is not visible.
     // For example, when user use pdf with different page sizes and zoomed in
     // outside page area.
     if (visible_rect.Intersects(GetPageScreenRect(i))) {
       visible_pages_.push_back(i);
@@ skipping 37 matching lines @@
 bool PDFiumEngine::IsPageVisible(int index) const {
   return base::ContainsValue(visible_pages_, index);
 }
 
 void PDFiumEngine::ScrollToPage(int page) {
   in_flight_visible_page_ = page;
   client_->ScrollToPage(page);
 }
 
 bool PDFiumEngine::CheckPageAvailable(int index, std::vector<int>* pending) {
-  if (!doc_ || !form_)
+  if (!doc_)
     return false;
 
   const int num_pages = static_cast<int>(pages_.size());
   if (index < num_pages && pages_[index]->available())
     return true;
 
   if (!FPDFAvail_IsPageAvail(fpdf_availability_, index, &download_hints_)) {
     if (!base::ContainsValue(*pending, index))
       pending->push_back(index);
     return false;
@@ skipping 1274 matching lines @@
   FPDF_DOCUMENT doc =
       FPDF_LoadMemDocument(pdf_buffer, pdf_buffer_size, nullptr);
   if (!doc)
     return false;
   bool success = FPDF_GetPageSizeByIndex(doc, page_number, width, height) != 0;
   FPDF_CloseDocument(doc);
   return success;
 }
 
 }  // namespace chrome_pdf