1 // Copyright (c) 2010 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
6 #include "base/command_line.h"
7 #include "base/compiler_specific.h"
8 #include "base/containers/hash_tables.h"
9 #include "base/file_util.h"
10 #include "base/files/file_path.h"
11 #include "base/strings/string_util.h"
12 #include "base/strings/utf_string_conversions.h"
13 #include "content/public/common/content_switches.h"
14 #include "content/public/renderer/render_view.h"
15 #include "content/public/renderer/render_view_observer.h"
16 #include "content/public/test/test_utils.h"
17 #include "content/renderer/savable_resources.h"
18 #include "content/shell/browser/shell.h"
19 #include "content/test/content_browser_test.h"
20 #include "content/test/content_browser_test_utils.h"
21 #include "net/base/net_util.h"
22 #include "net/url_request/url_request_context.h"
23 #include "third_party/WebKit/public/platform/WebCString.h"
24 #include "third_party/WebKit/public/platform/WebData.h"
25 #include "third_party/WebKit/public/platform/WebString.h"
26 #include "third_party/WebKit/public/platform/WebURL.h"
27 #include "third_party/WebKit/public/platform/WebVector.h"
28 #include "third_party/WebKit/public/web/WebDocument.h"
29 #include "third_party/WebKit/public/web/WebElement.h"
30 #include "third_party/WebKit/public/web/WebFrame.h"
31 #include "third_party/WebKit/public/web/WebNode.h"
32 #include "third_party/WebKit/public/web/WebNodeCollection.h"
33 #include "third_party/WebKit/public/web/WebNodeList.h"
34 #include "third_party/WebKit/public/web/WebPageSerializer.h"
35 #include "third_party/WebKit/public/web/WebPageSerializerClient.h"
36 #include "third_party/WebKit/public/web/WebView.h"
38 using blink::WebCString
;
40 using blink::WebDocument
;
41 using blink::WebElement
;
42 using blink::WebFrame
;
44 using blink::WebNodeCollection
;
45 using blink::WebNodeList
;
46 using blink::WebPageSerializer
;
47 using blink::WebPageSerializerClient
;
49 using blink::WebString
;
52 using blink::WebVector
;
56 // The first RenderFrame is routing ID 1, and the first RenderView is 2.
57 const int kRenderViewRoutingId
= 2;
63 // Iterate recursively over sub-frames to find one with with a given url.
64 WebFrame
* FindSubFrameByURL(WebView
* web_view
, const GURL
& url
) {
65 if (!web_view
->mainFrame())
68 std::vector
<WebFrame
*> stack
;
69 stack
.push_back(web_view
->mainFrame());
71 while (!stack
.empty()) {
72 WebFrame
* current_frame
= stack
.back();
74 if (GURL(current_frame
->document().url()) == url
)
76 WebNodeCollection all
= current_frame
->document().all();
77 for (WebNode node
= all
.firstItem();
78 !node
.isNull(); node
= all
.nextItem()) {
79 if (!node
.isElementNode())
81 // Check frame tag and iframe tag
82 WebElement element
= node
.to
<WebElement
>();
83 if (!element
.hasTagName("frame") && !element
.hasTagName("iframe"))
85 WebFrame
* sub_frame
= WebFrame::fromFrameOwnerElement(element
);
87 stack
.push_back(sub_frame
);
93 // Helper function that test whether the first node in the doc is a doc type
95 bool HasDocType(const WebDocument
& doc
) {
96 WebNode node
= doc
.firstChild();
99 return node
.nodeType() == WebNode::DocumentTypeNode
;
102 // Helper function for checking whether input node is META tag. Return true
103 // means it is META element, otherwise return false. The parameter charset_info
104 // return actual charset info if the META tag has charset declaration.
105 bool IsMetaElement(const WebNode
& node
, std::string
& charset_info
) {
106 if (!node
.isElementNode())
108 const WebElement meta
= node
.toConst
<WebElement
>();
109 if (!meta
.hasTagName("meta"))
111 charset_info
.erase(0, charset_info
.length());
112 // Check the META charset declaration.
113 WebString httpEquiv
= meta
.getAttribute("http-equiv");
114 if (LowerCaseEqualsASCII(httpEquiv
, "content-type")) {
115 std::string content
= meta
.getAttribute("content").utf8();
116 int pos
= content
.find("charset", 0);
118 // Add a dummy charset declaration to charset_info, which indicates this
119 // META tag has charset declaration although we do not get correct value
121 charset_info
.append("has-charset-declaration");
122 int remaining_length
= content
.length() - pos
- 7;
123 if (!remaining_length
)
125 int start_pos
= pos
+ 7;
127 while (remaining_length
--)
128 if (content
[start_pos
++] == L
'=')
130 // Skip beginning space.
131 while (remaining_length
) {
132 if (content
[start_pos
] > 0x0020)
137 if (!remaining_length
)
139 int end_pos
= start_pos
;
140 // Now we find out the start point of charset info. Search the end point.
141 while (remaining_length
--) {
142 if (content
[end_pos
] <= 0x0020 || content
[end_pos
] == L
';')
146 // Get actual charset info.
147 charset_info
= content
.substr(start_pos
, end_pos
- start_pos
);
154 class LoadObserver
: public RenderViewObserver
{
156 LoadObserver(RenderView
* render_view
, const base::Closure
& quit_closure
)
157 : RenderViewObserver(render_view
),
158 quit_closure_(quit_closure
) {}
160 virtual void DidFinishLoad(blink::WebFrame
* frame
) OVERRIDE
{
161 if (frame
== render_view()->GetWebView()->mainFrame())
166 base::Closure quit_closure_
;
169 class DomSerializerTests
: public ContentBrowserTest
,
170 public WebPageSerializerClient
{
173 : serialized_(false),
174 local_directory_name_(FILE_PATH_LITERAL("./dummy_files/")) {}
176 virtual void SetUpCommandLine(CommandLine
* command_line
) OVERRIDE
{
177 command_line
->AppendSwitch(switches::kSingleProcess
);
178 #if defined(OS_WIN) && defined(USE_AURA)
179 // Don't want to try to create a GPU process.
180 command_line
->AppendSwitch(switches::kDisableAcceleratedCompositing
);
184 // DomSerializerDelegate.
185 virtual void didSerializeDataForFrame(const WebURL
& frame_web_url
,
186 const WebCString
& data
,
187 PageSerializationStatus status
) {
189 GURL
frame_url(frame_web_url
);
190 // If the all frames are finished saving, check all finish status
191 if (status
== WebPageSerializerClient::AllFramesAreFinished
) {
192 SerializationFinishStatusMap::iterator it
=
193 serialization_finish_status_
.begin();
194 for (; it
!= serialization_finish_status_
.end(); ++it
)
195 ASSERT_TRUE(it
->second
);
200 // Check finish status of current frame.
201 SerializationFinishStatusMap::iterator it
=
202 serialization_finish_status_
.find(frame_url
.spec());
203 // New frame, set initial status as false.
204 if (it
== serialization_finish_status_
.end())
205 serialization_finish_status_
[frame_url
.spec()] = false;
207 it
= serialization_finish_status_
.find(frame_url
.spec());
208 ASSERT_TRUE(it
!= serialization_finish_status_
.end());
209 // In process frame, finish status should be false.
210 ASSERT_FALSE(it
->second
);
212 // Add data to corresponding frame's content.
213 serialized_frame_map_
[frame_url
.spec()] += data
.data();
215 // Current frame is completed saving, change the finish status.
216 if (status
== WebPageSerializerClient::CurrentFrameIsFinished
)
220 bool HasSerializedFrame(const GURL
& frame_url
) {
221 return serialized_frame_map_
.find(frame_url
.spec()) !=
222 serialized_frame_map_
.end();
225 const std::string
& GetSerializedContentForFrame(
226 const GURL
& frame_url
) {
227 return serialized_frame_map_
[frame_url
.spec()];
230 RenderView
* GetRenderView() {
231 // We could have the test on the UI thread get the WebContent's routing ID,
232 // but we know this will be the first RV so skip that and just hardcode it.
233 return RenderView::FromRoutingID(kRenderViewRoutingId
);
236 WebView
* GetWebView() {
237 return GetRenderView()->GetWebView();
240 WebFrame
* GetMainFrame() {
241 return GetWebView()->mainFrame();
244 // Load web page according to input content and relative URLs within
246 void LoadContents(const std::string
& contents
,
247 const GURL
& base_url
,
248 const WebString encoding_info
) {
249 scoped_refptr
<MessageLoopRunner
> runner
= new MessageLoopRunner
;
250 LoadObserver
observer(GetRenderView(), runner
->QuitClosure());
252 // If input encoding is empty, use UTF-8 as default encoding.
253 if (encoding_info
.isEmpty()) {
254 GetMainFrame()->loadHTMLString(contents
, base_url
);
256 WebData
data(contents
.data(), contents
.length());
258 // Do not use WebFrame.LoadHTMLString because it assumes that input
259 // html contents use UTF-8 encoding.
260 // TODO(darin): This should use WebFrame::loadData.
261 WebFrame
* web_frame
= GetMainFrame();
263 ASSERT_TRUE(web_frame
!= NULL
);
265 web_frame
->loadData(data
, "text/html", encoding_info
, base_url
);
271 // Serialize page DOM according to specific page URL. The parameter
272 // recursive_serialization indicates whether we will serialize all
274 void SerializeDomForURL(const GURL
& page_url
,
275 bool recursive_serialization
) {
276 // Find corresponding WebFrame according to page_url.
277 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), page_url
);
278 ASSERT_TRUE(web_frame
!= NULL
);
279 WebVector
<WebURL
> links
;
280 links
.assign(&page_url
, 1);
281 WebString file_path
=
282 base::FilePath(FILE_PATH_LITERAL("c:\\dummy.htm")).AsUTF16Unsafe();
283 WebVector
<WebString
> local_paths
;
284 local_paths
.assign(&file_path
, 1);
285 // Start serializing DOM.
286 bool result
= WebPageSerializer::serialize(web_frame
,
287 recursive_serialization
,
288 static_cast<WebPageSerializerClient
*>(this),
291 local_directory_name_
.AsUTF16Unsafe());
293 ASSERT_TRUE(serialized_
);
296 void SerializeHTMLDOMWithDocTypeOnRenderer(const GURL
& file_url
) {
297 // Make sure original contents have document type.
298 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
299 ASSERT_TRUE(web_frame
!= NULL
);
300 WebDocument doc
= web_frame
->document();
301 ASSERT_TRUE(HasDocType(doc
));
303 SerializeDomForURL(file_url
, false);
304 // Load the serialized contents.
305 ASSERT_TRUE(HasSerializedFrame(file_url
));
306 const std::string
& serialized_contents
=
307 GetSerializedContentForFrame(file_url
);
308 LoadContents(serialized_contents
, file_url
,
309 web_frame
->document().encoding());
310 // Make sure serialized contents still have document type.
311 web_frame
= GetMainFrame();
312 doc
= web_frame
->document();
313 ASSERT_TRUE(HasDocType(doc
));
316 void SerializeHTMLDOMWithoutDocTypeOnRenderer(const GURL
& file_url
) {
317 // Make sure original contents do not have document type.
318 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
319 ASSERT_TRUE(web_frame
!= NULL
);
320 WebDocument doc
= web_frame
->document();
321 ASSERT_TRUE(!HasDocType(doc
));
323 SerializeDomForURL(file_url
, false);
324 // Load the serialized contents.
325 ASSERT_TRUE(HasSerializedFrame(file_url
));
326 const std::string
& serialized_contents
=
327 GetSerializedContentForFrame(file_url
);
328 LoadContents(serialized_contents
, file_url
,
329 web_frame
->document().encoding());
330 // Make sure serialized contents do not have document type.
331 web_frame
= GetMainFrame();
332 doc
= web_frame
->document();
333 ASSERT_TRUE(!HasDocType(doc
));
336 void SerializeXMLDocWithBuiltInEntitiesOnRenderer(
337 const GURL
& xml_file_url
, const std::string
& original_contents
) {
339 SerializeDomForURL(xml_file_url
, false);
340 // Compare the serialized contents with original contents.
341 ASSERT_TRUE(HasSerializedFrame(xml_file_url
));
342 const std::string
& serialized_contents
=
343 GetSerializedContentForFrame(xml_file_url
);
344 ASSERT_EQ(original_contents
, serialized_contents
);
347 void SerializeHTMLDOMWithAddingMOTWOnRenderer(
348 const GURL
& file_url
, const std::string
& original_contents
) {
349 // Make sure original contents does not have MOTW;
350 std::string motw_declaration
=
351 WebPageSerializer::generateMarkOfTheWebDeclaration(file_url
).utf8();
352 ASSERT_FALSE(motw_declaration
.empty());
353 // The encoding of original contents is ISO-8859-1, so we convert the MOTW
354 // declaration to ASCII and search whether original contents has it or not.
355 ASSERT_TRUE(std::string::npos
== original_contents
.find(motw_declaration
));
358 SerializeDomForURL(file_url
, false);
359 // Make sure the serialized contents have MOTW ;
360 ASSERT_TRUE(HasSerializedFrame(file_url
));
361 const std::string
& serialized_contents
=
362 GetSerializedContentForFrame(file_url
);
363 ASSERT_FALSE(std::string::npos
==
364 serialized_contents
.find(motw_declaration
));
367 void SerializeHTMLDOMWithNoMetaCharsetInOriginalDocOnRenderer(
368 const GURL
& file_url
) {
369 // Make sure there is no META charset declaration in original document.
370 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
371 ASSERT_TRUE(web_frame
!= NULL
);
372 WebDocument doc
= web_frame
->document();
373 ASSERT_TRUE(doc
.isHTMLDocument());
374 WebElement head_element
= doc
.head();
375 ASSERT_TRUE(!head_element
.isNull());
376 // Go through all children of HEAD element.
377 for (WebNode child
= head_element
.firstChild(); !child
.isNull();
378 child
= child
.nextSibling()) {
379 std::string charset_info
;
380 if (IsMetaElement(child
, charset_info
))
381 ASSERT_TRUE(charset_info
.empty());
384 SerializeDomForURL(file_url
, false);
386 // Load the serialized contents.
387 ASSERT_TRUE(HasSerializedFrame(file_url
));
388 const std::string
& serialized_contents
=
389 GetSerializedContentForFrame(file_url
);
390 LoadContents(serialized_contents
, file_url
,
391 web_frame
->document().encoding());
392 // Make sure the first child of HEAD element is META which has charset
393 // declaration in serialized contents.
394 web_frame
= GetMainFrame();
395 ASSERT_TRUE(web_frame
!= NULL
);
396 doc
= web_frame
->document();
397 ASSERT_TRUE(doc
.isHTMLDocument());
398 head_element
= doc
.head();
399 ASSERT_TRUE(!head_element
.isNull());
400 WebNode meta_node
= head_element
.firstChild();
401 ASSERT_TRUE(!meta_node
.isNull());
402 // Get meta charset info.
403 std::string charset_info2
;
404 ASSERT_TRUE(IsMetaElement(meta_node
, charset_info2
));
405 ASSERT_TRUE(!charset_info2
.empty());
406 ASSERT_EQ(charset_info2
,
407 std::string(web_frame
->document().encoding().utf8()));
409 // Make sure no more additional META tags which have charset declaration.
410 for (WebNode child
= meta_node
.nextSibling(); !child
.isNull();
411 child
= child
.nextSibling()) {
412 std::string charset_info
;
413 if (IsMetaElement(child
, charset_info
))
414 ASSERT_TRUE(charset_info
.empty());
418 void SerializeHTMLDOMWithMultipleMetaCharsetInOriginalDocOnRenderer(
419 const GURL
& file_url
) {
420 // Make sure there are multiple META charset declarations in original
422 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
423 ASSERT_TRUE(web_frame
!= NULL
);
424 WebDocument doc
= web_frame
->document();
425 ASSERT_TRUE(doc
.isHTMLDocument());
426 WebElement head_ele
= doc
.head();
427 ASSERT_TRUE(!head_ele
.isNull());
428 // Go through all children of HEAD element.
429 int charset_declaration_count
= 0;
430 for (WebNode child
= head_ele
.firstChild(); !child
.isNull();
431 child
= child
.nextSibling()) {
432 std::string charset_info
;
433 if (IsMetaElement(child
, charset_info
) && !charset_info
.empty())
434 charset_declaration_count
++;
436 // The original doc has more than META tags which have charset declaration.
437 ASSERT_TRUE(charset_declaration_count
> 1);
440 SerializeDomForURL(file_url
, false);
442 // Load the serialized contents.
443 ASSERT_TRUE(HasSerializedFrame(file_url
));
444 const std::string
& serialized_contents
=
445 GetSerializedContentForFrame(file_url
);
446 LoadContents(serialized_contents
, file_url
,
447 web_frame
->document().encoding());
448 // Make sure only first child of HEAD element is META which has charset
449 // declaration in serialized contents.
450 web_frame
= GetMainFrame();
451 ASSERT_TRUE(web_frame
!= NULL
);
452 doc
= web_frame
->document();
453 ASSERT_TRUE(doc
.isHTMLDocument());
454 head_ele
= doc
.head();
455 ASSERT_TRUE(!head_ele
.isNull());
456 WebNode meta_node
= head_ele
.firstChild();
457 ASSERT_TRUE(!meta_node
.isNull());
458 // Get meta charset info.
459 std::string charset_info2
;
460 ASSERT_TRUE(IsMetaElement(meta_node
, charset_info2
));
461 ASSERT_TRUE(!charset_info2
.empty());
462 ASSERT_EQ(charset_info2
,
463 std::string(web_frame
->document().encoding().utf8()));
465 // Make sure no more additional META tags which have charset declaration.
466 for (WebNode child
= meta_node
.nextSibling(); !child
.isNull();
467 child
= child
.nextSibling()) {
468 std::string charset_info
;
469 if (IsMetaElement(child
, charset_info
))
470 ASSERT_TRUE(charset_info
.empty());
474 void SerializeHTMLDOMWithEntitiesInTextOnRenderer() {
475 base::FilePath page_file_path
= GetTestFilePath(
476 "dom_serializer", "dom_serializer/htmlentities_in_text.htm");
477 // Get file URL. The URL is dummy URL to identify the following loading
478 // actions. The test content is in constant:original_contents.
479 GURL file_url
= net::FilePathToFileURL(page_file_path
);
480 ASSERT_TRUE(file_url
.SchemeIsFile());
482 static const char* const original_contents
=
483 "<html><body>&<>\"\'</body></html>";
484 // Load the test contents.
485 LoadContents(original_contents
, file_url
, WebString());
487 // Get BODY's text content in DOM.
488 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
489 ASSERT_TRUE(web_frame
!= NULL
);
490 WebDocument doc
= web_frame
->document();
491 ASSERT_TRUE(doc
.isHTMLDocument());
492 WebElement body_ele
= doc
.body();
493 ASSERT_TRUE(!body_ele
.isNull());
494 WebNode text_node
= body_ele
.firstChild();
495 ASSERT_TRUE(text_node
.isTextNode());
496 ASSERT_TRUE(std::string(text_node
.createMarkup().utf8()) ==
497 "&<>\"\'");
499 SerializeDomForURL(file_url
, false);
500 // Compare the serialized contents with original contents.
501 ASSERT_TRUE(HasSerializedFrame(file_url
));
502 const std::string
& serialized_contents
=
503 GetSerializedContentForFrame(file_url
);
504 // Compare the serialized contents with original contents to make sure
506 // Because we add MOTW when serializing DOM, so before comparison, we also
507 // need to add MOTW to original_contents.
508 std::string original_str
=
509 WebPageSerializer::generateMarkOfTheWebDeclaration(file_url
).utf8();
510 original_str
+= original_contents
;
511 // Since WebCore now inserts a new HEAD element if there is no HEAD element
512 // when creating BODY element. (Please see
513 // HTMLParser::bodyCreateErrorCheck.) We need to append the HEAD content and
514 // corresponding META content if we find WebCore-generated HEAD element.
515 if (!doc
.head().isNull()) {
516 WebString encoding
= web_frame
->document().encoding();
517 std::string
htmlTag("<html>");
518 std::string::size_type pos
= original_str
.find(htmlTag
);
519 ASSERT_NE(std::string::npos
, pos
);
520 pos
+= htmlTag
.length();
521 std::string
head_part("<head>");
523 WebPageSerializer::generateMetaCharsetDeclaration(encoding
).utf8();
524 head_part
+= "</head>";
525 original_str
.insert(pos
, head_part
);
527 ASSERT_EQ(original_str
, serialized_contents
);
530 void SerializeHTMLDOMWithEntitiesInAttributeValueOnRenderer() {
531 base::FilePath page_file_path
= GetTestFilePath(
532 "dom_serializer", "dom_serializer/htmlentities_in_attribute_value.htm");
533 // Get file URL. The URL is dummy URL to identify the following loading
534 // actions. The test content is in constant:original_contents.
535 GURL file_url
= net::FilePathToFileURL(page_file_path
);
536 ASSERT_TRUE(file_url
.SchemeIsFile());
538 static const char* const original_contents
=
539 "<html><body title=\"&<>"'\"></body></html>";
540 // Load the test contents.
541 LoadContents(original_contents
, file_url
, WebString());
542 // Get value of BODY's title attribute in DOM.
543 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
544 ASSERT_TRUE(web_frame
!= NULL
);
545 WebDocument doc
= web_frame
->document();
546 ASSERT_TRUE(doc
.isHTMLDocument());
547 WebElement body_ele
= doc
.body();
548 ASSERT_TRUE(!body_ele
.isNull());
549 WebString value
= body_ele
.getAttribute("title");
550 ASSERT_TRUE(std::string(value
.utf8()) == "&<>\"\'");
552 SerializeDomForURL(file_url
, false);
553 // Compare the serialized contents with original contents.
554 ASSERT_TRUE(HasSerializedFrame(file_url
));
555 const std::string
& serialized_contents
=
556 GetSerializedContentForFrame(file_url
);
557 // Compare the serialized contents with original contents to make sure
559 std::string original_str
=
560 WebPageSerializer::generateMarkOfTheWebDeclaration(file_url
).utf8();
561 original_str
+= original_contents
;
563 WebString encoding
= web_frame
->document().encoding();
564 std::string
htmlTag("<html>");
565 std::string::size_type pos
= original_str
.find(htmlTag
);
566 ASSERT_NE(std::string::npos
, pos
);
567 pos
+= htmlTag
.length();
568 std::string
head_part("<head>");
570 WebPageSerializer::generateMetaCharsetDeclaration(encoding
).utf8();
571 head_part
+= "</head>";
572 original_str
.insert(pos
, head_part
);
574 ASSERT_EQ(original_str
, serialized_contents
);
577 void SerializeHTMLDOMWithNonStandardEntitiesOnRenderer(const GURL
& file_url
) {
578 // Get value of BODY's title attribute in DOM.
579 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
580 WebDocument doc
= web_frame
->document();
581 ASSERT_TRUE(doc
.isHTMLDocument());
582 WebElement body_element
= doc
.body();
583 // Unescaped string for "%⊅¹'".
584 static const wchar_t parsed_value
[] = {
585 '%', 0x2285, 0x00b9, '\'', 0
587 WebString value
= body_element
.getAttribute("title");
588 ASSERT_TRUE(base::UTF16ToWide(value
) == parsed_value
);
589 ASSERT_TRUE(base::UTF16ToWide(body_element
.innerText()) == parsed_value
);
592 SerializeDomForURL(file_url
, false);
593 // Check the serialized string.
594 ASSERT_TRUE(HasSerializedFrame(file_url
));
595 const std::string
& serialized_contents
=
596 GetSerializedContentForFrame(file_url
);
597 // Confirm that the serialized string has no non-standard HTML entities.
598 ASSERT_EQ(std::string::npos
, serialized_contents
.find("%"));
599 ASSERT_EQ(std::string::npos
, serialized_contents
.find("⊅"));
600 ASSERT_EQ(std::string::npos
, serialized_contents
.find("¹"));
601 ASSERT_EQ(std::string::npos
, serialized_contents
.find("'"));
604 void SerializeHTMLDOMWithBaseTagOnRenderer(const GURL
& file_url
,
605 const GURL
& path_dir_url
) {
606 // There are total 2 available base tags in this test file.
607 const int kTotalBaseTagCountInTestFile
= 2;
609 // Since for this test, we assume there is no savable sub-resource links for
610 // this test file, also all links are relative URLs in this test file, so we
611 // need to check those relative URLs and make sure document has BASE tag.
612 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
613 ASSERT_TRUE(web_frame
!= NULL
);
614 WebDocument doc
= web_frame
->document();
615 ASSERT_TRUE(doc
.isHTMLDocument());
616 // Go through all descent nodes.
617 WebNodeCollection all
= doc
.all();
618 int original_base_tag_count
= 0;
619 for (WebNode node
= all
.firstItem(); !node
.isNull();
620 node
= all
.nextItem()) {
621 if (!node
.isElementNode())
623 WebElement element
= node
.to
<WebElement
>();
624 if (element
.hasTagName("base")) {
625 original_base_tag_count
++;
628 WebString value
= GetSubResourceLinkFromElement(element
);
629 if (value
.isNull() && element
.hasTagName("a")) {
630 value
= element
.getAttribute("href");
634 // Each link is relative link.
635 if (!value
.isNull()) {
636 GURL
link(value
.utf8());
637 ASSERT_TRUE(link
.scheme().empty());
641 ASSERT_EQ(original_base_tag_count
, kTotalBaseTagCountInTestFile
);
642 // Make sure in original document, the base URL is not equal with the
644 GURL
original_base_url(doc
.baseURL());
645 ASSERT_NE(original_base_url
, path_dir_url
);
648 SerializeDomForURL(file_url
, false);
650 // Load the serialized contents.
651 ASSERT_TRUE(HasSerializedFrame(file_url
));
652 const std::string
& serialized_contents
=
653 GetSerializedContentForFrame(file_url
);
654 LoadContents(serialized_contents
, file_url
,
655 web_frame
->document().encoding());
657 // Make sure all links are absolute URLs and doc there are some number of
658 // BASE tags in serialized HTML data. Each of those BASE tags have same base
659 // URL which is as same as URL of current test file.
660 web_frame
= GetMainFrame();
661 ASSERT_TRUE(web_frame
!= NULL
);
662 doc
= web_frame
->document();
663 ASSERT_TRUE(doc
.isHTMLDocument());
664 // Go through all descent nodes.
666 int new_base_tag_count
= 0;
667 for (WebNode node
= all
.firstItem(); !node
.isNull();
668 node
= all
.nextItem()) {
669 if (!node
.isElementNode())
671 WebElement element
= node
.to
<WebElement
>();
672 if (element
.hasTagName("base")) {
673 new_base_tag_count
++;
676 WebString value
= GetSubResourceLinkFromElement(element
);
677 if (value
.isNull() && element
.hasTagName("a")) {
678 value
= element
.getAttribute("href");
682 // Each link is absolute link.
683 if (!value
.isNull()) {
684 GURL
link(std::string(value
.utf8()));
685 ASSERT_FALSE(link
.scheme().empty());
689 // We have one more added BASE tag which is generated by JavaScript.
690 ASSERT_EQ(new_base_tag_count
, original_base_tag_count
+ 1);
691 // Make sure in new document, the base URL is equal with the |path_dir_url|.
692 GURL
new_base_url(doc
.baseURL());
693 ASSERT_EQ(new_base_url
, path_dir_url
);
696 void SerializeHTMLDOMWithEmptyHeadOnRenderer() {
697 base::FilePath page_file_path
= GetTestFilePath(
698 "dom_serializer", "empty_head.htm");
699 GURL file_url
= net::FilePathToFileURL(page_file_path
);
700 ASSERT_TRUE(file_url
.SchemeIsFile());
702 // Load the test html content.
703 static const char* const empty_head_contents
=
704 "<html><head></head><body>hello world</body></html>";
705 LoadContents(empty_head_contents
, file_url
, WebString());
707 // Make sure the head tag is empty.
708 WebFrame
* web_frame
= GetMainFrame();
709 ASSERT_TRUE(web_frame
!= NULL
);
710 WebDocument doc
= web_frame
->document();
711 ASSERT_TRUE(doc
.isHTMLDocument());
712 WebElement head_element
= doc
.head();
713 ASSERT_TRUE(!head_element
.isNull());
714 ASSERT_TRUE(!head_element
.hasChildNodes());
715 ASSERT_TRUE(head_element
.childNodes().length() == 0);
718 SerializeDomForURL(file_url
, false);
719 // Make sure the serialized contents have META ;
720 ASSERT_TRUE(HasSerializedFrame(file_url
));
721 const std::string
& serialized_contents
=
722 GetSerializedContentForFrame(file_url
);
724 // Reload serialized contents and make sure there is only one META tag.
725 LoadContents(serialized_contents
, file_url
,
726 web_frame
->document().encoding());
727 web_frame
= GetMainFrame();
728 ASSERT_TRUE(web_frame
!= NULL
);
729 doc
= web_frame
->document();
730 ASSERT_TRUE(doc
.isHTMLDocument());
731 head_element
= doc
.head();
732 ASSERT_TRUE(!head_element
.isNull());
733 ASSERT_TRUE(head_element
.hasChildNodes());
734 ASSERT_TRUE(head_element
.childNodes().length() == 1);
735 WebNode meta_node
= head_element
.firstChild();
736 ASSERT_TRUE(!meta_node
.isNull());
737 // Get meta charset info.
738 std::string charset_info
;
739 ASSERT_TRUE(IsMetaElement(meta_node
, charset_info
));
740 ASSERT_TRUE(!charset_info
.empty());
741 ASSERT_EQ(charset_info
,
742 std::string(web_frame
->document().encoding().utf8()));
744 // Check the body's first node is text node and its contents are
746 WebElement body_element
= doc
.body();
747 ASSERT_TRUE(!body_element
.isNull());
748 WebNode text_node
= body_element
.firstChild();
749 ASSERT_TRUE(text_node
.isTextNode());
750 WebString text_node_contents
= text_node
.nodeValue();
751 ASSERT_TRUE(std::string(text_node_contents
.utf8()) == "hello world");
754 void SerializeDocumentWithDownloadedIFrameOnRenderer(const GURL
& file_url
) {
755 // Do a recursive serialization. We pass if we don't crash.
756 SerializeDomForURL(file_url
, true);
759 void SubResourceForElementsInNonHTMLNamespaceOnRenderer(
760 const GURL
& file_url
) {
761 WebFrame
* web_frame
= FindSubFrameByURL(GetWebView(), file_url
);
762 ASSERT_TRUE(web_frame
!= NULL
);
763 WebDocument doc
= web_frame
->document();
764 WebNode lastNodeInBody
= doc
.body().lastChild();
765 ASSERT_EQ(WebNode::ElementNode
, lastNodeInBody
.nodeType());
766 WebString uri
= GetSubResourceLinkFromElement(
767 lastNodeInBody
.to
<WebElement
>());
768 EXPECT_TRUE(uri
.isNull());
772 // Map frame_url to corresponding serialized_content.
773 typedef base::hash_map
<std::string
, std::string
> SerializedFrameContentMap
;
774 SerializedFrameContentMap serialized_frame_map_
;
775 // Map frame_url to corresponding status of serialization finish.
776 typedef base::hash_map
<std::string
, bool> SerializationFinishStatusMap
;
777 SerializationFinishStatusMap serialization_finish_status_
;
778 // Flag indicates whether the process of serializing DOM is finished or not.
780 // The local_directory_name_ is dummy relative path of directory which
781 // contain all saved auxiliary files included all sub frames and resources.
782 const base::FilePath local_directory_name_
;
785 // If original contents have document type, the serialized contents also have
787 IN_PROC_BROWSER_TEST_F(DomSerializerTests
, SerializeHTMLDOMWithDocType
) {
788 base::FilePath page_file_path
=
789 GetTestFilePath("dom_serializer", "youtube_1.htm");
790 GURL file_url
= net::FilePathToFileURL(page_file_path
);
791 ASSERT_TRUE(file_url
.SchemeIsFile());
792 // Load the test file.
793 NavigateToURL(shell(), file_url
);
795 PostTaskToInProcessRendererAndWait(
796 base::Bind(&DomSerializerTests::SerializeHTMLDOMWithDocTypeOnRenderer
,
797 base::Unretained(this), file_url
));
800 // If original contents do not have document type, the serialized contents
801 // also do not have document type.
802 IN_PROC_BROWSER_TEST_F(DomSerializerTests
, SerializeHTMLDOMWithoutDocType
) {
803 base::FilePath page_file_path
=
804 GetTestFilePath("dom_serializer", "youtube_2.htm");
805 GURL file_url
= net::FilePathToFileURL(page_file_path
);
806 ASSERT_TRUE(file_url
.SchemeIsFile());
807 // Load the test file.
808 NavigateToURL(shell(), file_url
);
810 PostTaskToInProcessRendererAndWait(
812 &DomSerializerTests::SerializeHTMLDOMWithoutDocTypeOnRenderer
,
813 base::Unretained(this), file_url
));
816 // Serialize XML document which has all 5 built-in entities. After
817 // finishing serialization, the serialized contents should be same
818 // with original XML document.
820 // TODO(tiger@opera.com): Disabled in preparation of page serializer merge --
821 // XML headers are handled differently in the merged serializer.
822 // Bug: http://crbug.com/328354
823 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
824 DISABLED_SerializeXMLDocWithBuiltInEntities
) {
825 base::FilePath page_file_path
=
826 GetTestFilePath("dom_serializer", "note.html");
827 base::FilePath xml_file_path
= GetTestFilePath("dom_serializer", "note.xml");
828 // Read original contents for later comparison.
829 std::string original_contents
;
830 ASSERT_TRUE(base::ReadFileToString(xml_file_path
, &original_contents
));
832 GURL file_url
= net::FilePathToFileURL(page_file_path
);
833 GURL xml_file_url
= net::FilePathToFileURL(xml_file_path
);
834 ASSERT_TRUE(file_url
.SchemeIsFile());
835 // Load the test file.
836 NavigateToURL(shell(), file_url
);
838 PostTaskToInProcessRendererAndWait(
840 &DomSerializerTests::SerializeXMLDocWithBuiltInEntitiesOnRenderer
,
841 base::Unretained(this), xml_file_url
, original_contents
));
844 // When serializing DOM, we add MOTW declaration before html tag.
845 IN_PROC_BROWSER_TEST_F(DomSerializerTests
, SerializeHTMLDOMWithAddingMOTW
) {
846 base::FilePath page_file_path
=
847 GetTestFilePath("dom_serializer", "youtube_2.htm");
848 // Read original contents for later comparison .
849 std::string original_contents
;
850 ASSERT_TRUE(base::ReadFileToString(page_file_path
, &original_contents
));
852 GURL file_url
= net::FilePathToFileURL(page_file_path
);
853 ASSERT_TRUE(file_url
.SchemeIsFile());
855 // Load the test file.
856 NavigateToURL(shell(), file_url
);
858 PostTaskToInProcessRendererAndWait(
860 &DomSerializerTests::SerializeHTMLDOMWithAddingMOTWOnRenderer
,
861 base::Unretained(this), file_url
, original_contents
));
864 // When serializing DOM, we will add the META which have correct charset
865 // declaration as first child of HEAD element for resolving WebKit bug:
866 // http://bugs.webkit.org/show_bug.cgi?id=16621 even the original document
867 // does not have META charset declaration.
868 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
869 SerializeHTMLDOMWithNoMetaCharsetInOriginalDoc
) {
870 base::FilePath page_file_path
=
871 GetTestFilePath("dom_serializer", "youtube_1.htm");
873 GURL file_url
= net::FilePathToFileURL(page_file_path
);
874 ASSERT_TRUE(file_url
.SchemeIsFile());
875 // Load the test file.
876 NavigateToURL(shell(), file_url
);
878 PostTaskToInProcessRendererAndWait(
880 &DomSerializerTests::
881 SerializeHTMLDOMWithNoMetaCharsetInOriginalDocOnRenderer
,
882 base::Unretained(this), file_url
));
885 // When serializing DOM, if the original document has multiple META charset
886 // declaration, we will add the META which have correct charset declaration
887 // as first child of HEAD element and remove all original META charset
889 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
890 SerializeHTMLDOMWithMultipleMetaCharsetInOriginalDoc
) {
891 base::FilePath page_file_path
=
892 GetTestFilePath("dom_serializer", "youtube_2.htm");
894 GURL file_url
= net::FilePathToFileURL(page_file_path
);
895 ASSERT_TRUE(file_url
.SchemeIsFile());
896 // Load the test file.
897 NavigateToURL(shell(), file_url
);
899 PostTaskToInProcessRendererAndWait(
901 &DomSerializerTests::
902 SerializeHTMLDOMWithMultipleMetaCharsetInOriginalDocOnRenderer
,
903 base::Unretained(this), file_url
));
906 // Test situation of html entities in text when serializing HTML DOM.
907 IN_PROC_BROWSER_TEST_F(DomSerializerTests
, SerializeHTMLDOMWithEntitiesInText
) {
908 // Need to spin up the renderer and also navigate to a file url so that the
909 // renderer code doesn't attempt a fork when it sees a load to file scheme
910 // from non-file scheme.
911 NavigateToURL(shell(), GetTestUrl(".", "simple_page.html"));
913 PostTaskToInProcessRendererAndWait(
915 &DomSerializerTests::SerializeHTMLDOMWithEntitiesInTextOnRenderer
,
916 base::Unretained(this)));
919 // Test situation of html entities in attribute value when serializing
921 // This test started to fail at WebKit r65388. See http://crbug.com/52279.
923 // TODO(tiger@opera.com): Disabled in preparation of page serializer merge --
924 // Some attributes are handled differently in the merged serializer.
925 // Bug: http://crbug.com/328354
926 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
927 DISABLED_SerializeHTMLDOMWithEntitiesInAttributeValue
) {
928 // Need to spin up the renderer and also navigate to a file url so that the
929 // renderer code doesn't attempt a fork when it sees a load to file scheme
930 // from non-file scheme.
931 NavigateToURL(shell(), GetTestUrl(".", "simple_page.html"));
933 PostTaskToInProcessRendererAndWait(
935 &DomSerializerTests::
936 SerializeHTMLDOMWithEntitiesInAttributeValueOnRenderer
,
937 base::Unretained(this)));
940 // Test situation of non-standard HTML entities when serializing HTML DOM.
941 // This test started to fail at WebKit r65351. See http://crbug.com/52279.
942 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
943 SerializeHTMLDOMWithNonStandardEntities
) {
944 // Make a test file URL and load it.
945 base::FilePath page_file_path
= GetTestFilePath(
946 "dom_serializer", "nonstandard_htmlentities.htm");
947 GURL file_url
= net::FilePathToFileURL(page_file_path
);
948 NavigateToURL(shell(), file_url
);
950 PostTaskToInProcessRendererAndWait(
952 &DomSerializerTests::
953 SerializeHTMLDOMWithNonStandardEntitiesOnRenderer
,
954 base::Unretained(this), file_url
));
957 // Test situation of BASE tag in original document when serializing HTML DOM.
958 // When serializing, we should comment the BASE tag, append a new BASE tag.
959 // rewrite all the savable URLs to relative local path, and change other URLs
962 // TODO(tiger@opera.com): Disabled in preparation of page serializer merge --
963 // Base tags are handled a bit different in merged version.
964 // Bug: http://crbug.com/328354
965 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
966 DISABLED_SerializeHTMLDOMWithBaseTag
) {
967 base::FilePath page_file_path
= GetTestFilePath(
968 "dom_serializer", "html_doc_has_base_tag.htm");
970 // Get page dir URL which is base URL of this file.
971 base::FilePath dir_name
= page_file_path
.DirName();
972 dir_name
= dir_name
.Append(
973 base::FilePath::StringType(base::FilePath::kSeparators
[0], 1));
974 GURL path_dir_url
= net::FilePathToFileURL(dir_name
);
977 GURL file_url
= net::FilePathToFileURL(page_file_path
);
978 ASSERT_TRUE(file_url
.SchemeIsFile());
979 // Load the test file.
980 NavigateToURL(shell(), file_url
);
982 PostTaskToInProcessRendererAndWait(
984 &DomSerializerTests::SerializeHTMLDOMWithBaseTagOnRenderer
,
985 base::Unretained(this), file_url
, path_dir_url
));
988 // Serializing page which has an empty HEAD tag.
989 IN_PROC_BROWSER_TEST_F(DomSerializerTests
, SerializeHTMLDOMWithEmptyHead
) {
990 // Need to spin up the renderer and also navigate to a file url so that the
991 // renderer code doesn't attempt a fork when it sees a load to file scheme
992 // from non-file scheme.
993 NavigateToURL(shell(), GetTestUrl(".", "simple_page.html"));
995 PostTaskToInProcessRendererAndWait(
996 base::Bind(&DomSerializerTests::SerializeHTMLDOMWithEmptyHeadOnRenderer
,
997 base::Unretained(this)));
1000 // Test that we don't crash when the page contains an iframe that
1001 // was handled as a download (http://crbug.com/42212).
1002 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
1003 SerializeDocumentWithDownloadedIFrame
) {
1004 base::FilePath page_file_path
= GetTestFilePath(
1005 "dom_serializer", "iframe-src-is-exe.htm");
1006 GURL file_url
= net::FilePathToFileURL(page_file_path
);
1007 ASSERT_TRUE(file_url
.SchemeIsFile());
1008 // Load the test file.
1009 NavigateToURL(shell(), file_url
);
1011 PostTaskToInProcessRendererAndWait(
1013 &DomSerializerTests::
1014 SerializeDocumentWithDownloadedIFrameOnRenderer
,
1015 base::Unretained(this), file_url
));
1018 IN_PROC_BROWSER_TEST_F(DomSerializerTests
,
1019 SubResourceForElementsInNonHTMLNamespace
) {
1020 base::FilePath page_file_path
= GetTestFilePath(
1021 "dom_serializer", "non_html_namespace.htm");
1022 GURL file_url
= net::FilePathToFileURL(page_file_path
);
1023 NavigateToURL(shell(), file_url
);
1025 PostTaskToInProcessRendererAndWait(
1027 &DomSerializerTests::
1028 SubResourceForElementsInNonHTMLNamespaceOnRenderer
,
1029 base::Unretained(this), file_url
));
1032 } // namespace content