third_party/WebKit/Source/web/tests/PageSerializerTest.cpp

   1 /*
   2  * Copyright (c) 2013, Opera Software ASA. All rights reserved.
   3  *
   4  * Redistribution and use in source and binary forms, with or without
   5  * modification, are permitted provided that the following conditions
   6  * are met:
   7  *
   8  * 1. Redistributions of source code must retain the above copyright
   9  *    notice, this list of conditions and the following disclaimer.
  10  * 2. Redistributions in binary form must reproduce the above copyright
  11  *    notice, this list of conditions and the following disclaimer in the
  12  *    documentation and/or other materials provided with the distribution.
  13  * 3. Neither the name of Opera Software ASA nor the names of its
  14  *    contributors may be used to endorse or promote products derived
  15  *    from this software without specific prior written permission.
  16  *
  17  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  18  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  19  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  20  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
  21  * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
  22  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
  23  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
  24  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  25  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
  26  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  27  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
  28  * OF THE POSSIBILITY OF SUCH DAMAGE.
  29  */
  30
  31 #include "config.h"
  32 #include "core/page/PageSerializer.h"
  33
  34 #include "bindings/core/v8/V8Binding.h"
  35 #include "bindings/core/v8/V8BindingForTesting.h"
  36 #include "core/page/Page.h"
  37 #include "platform/SerializedResource.h"
  38 #include "platform/testing/URLTestHelpers.h"
  39 #include "public/platform/Platform.h"
  40 #include "public/platform/WebString.h"
  41 #include "public/platform/WebThread.h"
  42 #include "public/platform/WebURL.h"
  43 #include "public/platform/WebURLRequest.h"
  44 #include "public/platform/WebURLResponse.h"
  45 #include "public/platform/WebUnitTestSupport.h"
  46 #include "public/web/WebSettings.h"
  47 #include "web/WebLocalFrameImpl.h"
  48 #include "web/WebViewImpl.h"
  49 #include "web/tests/FrameTestHelpers.h"
  50 #include "wtf/Vector.h"
  51 #include <gtest/gtest.h>
  52
  53 using blink::URLTestHelpers::toKURL;
  54 using blink::URLTestHelpers::registerMockedURLLoad;
  55
  56 namespace blink {
  57
  58 class PageSerializerTest : public testing::Test {
  59 public:
  60     PageSerializerTest()
  61         : m_folder(WebString::fromUTF8("pageserializer/"))
  62         , m_baseUrl(toKURL("http://www.test.com"))
  63     {
  64     }
  65
  66 protected:
  67     void SetUp() override
  68     {
  69         // We want the images to load and JavaScript to be on.
  70         m_helper.initialize(true, 0, 0, &configureSettings);
  71     }
  72
  73     void TearDown() override
  74     {
  75         Platform::current()->unitTestSupport()->unregisterAllMockedURLs();
  76     }
  77
  78     void setBaseFolder(const char* folder)
  79     {
  80         m_folder = WebString::fromUTF8(folder);
  81     }
  82
  83     void setRewriteURLFolder(const char* folder)
  84     {
  85         m_rewriteFolder = folder;
  86     }
  87
  88     void registerURL(const char* url, const char* file, const char* mimeType)
  89     {
  90         registerMockedURLLoad(KURL(m_baseUrl, url), WebString::fromUTF8(file), m_folder, WebString::fromUTF8(mimeType));
  91     }
  92
  93     void registerURL(const char* file, const char* mimeType)
  94     {
  95         registerURL(file, file, mimeType);
  96     }
  97
  98     void registerErrorURL(const char* file, int statusCode)
  99     {
 100         WebURLError error;
 101         error.reason = 0xdead + statusCode;
 102         error.domain = "PageSerializerTest";
 103
 104         WebURLResponse response;
 105         response.initialize();
 106         response.setMIMEType("text/html");
 107         response.setHTTPStatusCode(statusCode);
 108
 109         Platform::current()->unitTestSupport()->registerMockedErrorURL(KURL(m_baseUrl, file), response, error);
 110     }
 111
 112     void registerRewriteURL(const char* fromURL, const char* toURL)
 113     {
 114         m_rewriteURLs.add(fromURL, toURL);
 115     }
 116
 117     void serialize(const char* url)
 118     {
 119         FrameTestHelpers::loadFrame(m_helper.webView()->mainFrame(), KURL(m_baseUrl, url).string().utf8().data());
 120         PageSerializer serializer(&m_resources, nullptr);
 121
 122         serializer.setRewriteURLFolder(m_rewriteFolder);
 123         for (const auto& rewriteURL: m_rewriteURLs)
 124             serializer.registerRewriteURL(rewriteURL.key, rewriteURL.value);
 125
 126         serializer.serialize(m_helper.webViewImpl()->mainFrameImpl()->frame()->page());
 127     }
 128
 129     Vector<SerializedResource>& getResources()
 130     {
 131         return m_resources;
 132     }
 133
 134
 135     const SerializedResource* getResource(const char* url, const char* mimeType)
 136     {
 137         KURL kURL = KURL(m_baseUrl, url);
 138         String mime(mimeType);
 139         for (size_t i = 0; i < m_resources.size(); ++i) {
 140             const SerializedResource& resource = m_resources[i];
 141             if (resource.url == kURL && !resource.data->isEmpty()
 142                 && (mime.isNull() || equalIgnoringCase(resource.mimeType, mime)))
 143                 return &resource;
 144         }
 145         return nullptr;
 146     }
 147
 148     bool isSerialized(const char* url, const char* mimeType = 0)
 149     {
 150         return getResource(url, mimeType);
 151     }
 152
 153     String getSerializedData(const char* url, const char* mimeType = 0)
 154     {
 155         const SerializedResource* resource = getResource(url, mimeType);
 156         if (resource)
 157             return String(resource->data->data(), resource->data->size());
 158         return String();
 159     }
 160
 161 private:
 162     static void configureSettings(WebSettings* settings)
 163     {
 164         settings->setImagesEnabled(true);
 165         settings->setLoadsImagesAutomatically(true);
 166         settings->setJavaScriptEnabled(true);
 167     }
 168
 169     FrameTestHelpers::WebViewHelper m_helper;
 170     WebString m_folder;
 171     KURL m_baseUrl;
 172     Vector<SerializedResource> m_resources;
 173     HashMap<String, String> m_rewriteURLs;
 174     String m_rewriteFolder;
 175 };
 176
 177 TEST_F(PageSerializerTest, HTMLElements)
 178 {
 179     setBaseFolder("pageserializer/elements/");
 180
 181     registerURL("elements.html", "text/html");
 182     registerURL("style.css", "style.css", "text/css");
 183     registerURL("copyright.html", "empty.txt", "text/html");
 184     registerURL("script.js", "empty.txt", "text/javascript");
 185
 186     registerURL("bodyBackground.png", "image.png", "image/png");
 187
 188     registerURL("imageSrc.png", "image.png", "image/png");
 189
 190     registerURL("inputImage.png", "image.png", "image/png");
 191
 192     registerURL("tableBackground.png", "image.png", "image/png");
 193     registerURL("trBackground.png", "image.png", "image/png");
 194     registerURL("tdBackground.png", "image.png", "image/png");
 195
 196     registerURL("blockquoteCite.html", "empty.txt", "text/html");
 197     registerURL("qCite.html", "empty.txt", "text/html");
 198     registerURL("delCite.html", "empty.txt", "text/html");
 199     registerURL("insCite.html", "empty.txt", "text/html");
 200
 201     registerErrorURL("nonExisting.png", 404);
 202
 203     serialize("elements.html");
 204
 205     EXPECT_EQ(8U, getResources().size());
 206
 207     EXPECT_TRUE(isSerialized("elements.html", "text/html"));
 208     EXPECT_TRUE(isSerialized("style.css", "text/css"));
 209     EXPECT_TRUE(isSerialized("bodyBackground.png", "image/png"));
 210     EXPECT_TRUE(isSerialized("imageSrc.png", "image/png"));
 211     EXPECT_TRUE(isSerialized("inputImage.png", "image/png"));
 212     EXPECT_TRUE(isSerialized("tableBackground.png", "image/png"));
 213     EXPECT_TRUE(isSerialized("trBackground.png", "image/png"));
 214     EXPECT_TRUE(isSerialized("tdBackground.png", "image/png"));
 215     EXPECT_FALSE(isSerialized("nonExisting.png", "image/png"));
 216 }
 217
 218 TEST_F(PageSerializerTest, Frames)
 219 {
 220     setBaseFolder("pageserializer/frames/");
 221
 222     registerURL("simple_frames.html", "text/html");
 223     registerURL("simple_frames_top.html", "text/html");
 224     registerURL("simple_frames_1.html", "text/html");
 225     registerURL("simple_frames_3.html", "text/html");
 226
 227     registerURL("frame_1.png", "image.png", "image/png");
 228     registerURL("frame_2.png", "image.png", "image/png");
 229     registerURL("frame_3.png", "image.png", "image/png");
 230     registerURL("frame_4.png", "image.png", "image/png");
 231
 232     serialize("simple_frames.html");
 233
 234     EXPECT_EQ(8U, getResources().size());
 235
 236     EXPECT_TRUE(isSerialized("simple_frames.html", "text/html"));
 237     EXPECT_TRUE(isSerialized("simple_frames_top.html", "text/html"));
 238     EXPECT_TRUE(isSerialized("simple_frames_1.html", "text/html"));
 239     EXPECT_TRUE(isSerialized("simple_frames_3.html", "text/html"));
 240
 241     EXPECT_TRUE(isSerialized("frame_1.png", "image/png"));
 242     EXPECT_TRUE(isSerialized("frame_2.png", "image/png"));
 243     EXPECT_TRUE(isSerialized("frame_3.png", "image/png"));
 244     EXPECT_TRUE(isSerialized("frame_4.png", "image/png"));
 245 }
 246
 247 TEST_F(PageSerializerTest, IFrames)
 248 {
 249     setBaseFolder("pageserializer/frames/");
 250
 251     registerURL("top_frame.html", "text/html");
 252     registerURL("simple_iframe.html", "text/html");
 253     registerURL("object_iframe.html", "text/html");
 254     registerURL("embed_iframe.html", "text/html");
 255
 256     registerURL("top.png", "image.png", "image/png");
 257     registerURL("simple.png", "image.png", "image/png");
 258     registerURL("object.png", "image.png", "image/png");
 259     registerURL("embed.png", "image.png", "image/png");
 260
 261     serialize("top_frame.html");
 262
 263     EXPECT_EQ(8U, getResources().size());
 264
 265     EXPECT_TRUE(isSerialized("top_frame.html", "text/html"));
 266     EXPECT_TRUE(isSerialized("simple_iframe.html", "text/html"));
 267     EXPECT_TRUE(isSerialized("object_iframe.html", "text/html"));
 268     EXPECT_TRUE(isSerialized("embed_iframe.html", "text/html"));
 269
 270     EXPECT_TRUE(isSerialized("top.png", "image/png"));
 271     EXPECT_TRUE(isSerialized("simple.png", "image/png"));
 272     EXPECT_TRUE(isSerialized("object.png", "image/png"));
 273     EXPECT_TRUE(isSerialized("embed.png", "image/png"));
 274 }
 275
 276 // Tests that when serializing a page with blank frames these are reported with their resources.
 277 TEST_F(PageSerializerTest, BlankFrames)
 278 {
 279     setBaseFolder("pageserializer/frames/");
 280
 281     registerURL("blank_frames.html", "text/html");
 282     registerURL("red_background.png", "image.png", "image/png");
 283     registerURL("orange_background.png", "image.png", "image/png");
 284     registerURL("blue_background.png", "image.png", "image/png");
 285
 286     serialize("blank_frames.html");
 287
 288     EXPECT_EQ(7U, getResources().size());
 289
 290     EXPECT_TRUE(isSerialized("http://www.test.com/red_background.png", "image/png"));
 291     EXPECT_TRUE(isSerialized("http://www.test.com/orange_background.png", "image/png"));
 292     EXPECT_TRUE(isSerialized("http://www.test.com/blue_background.png", "image/png"));
 293     // The blank frames should have got a magic URL.
 294     EXPECT_TRUE(isSerialized("wyciwyg://frame/0", "text/html"));
 295     EXPECT_TRUE(isSerialized("wyciwyg://frame/1", "text/html"));
 296     EXPECT_TRUE(isSerialized("wyciwyg://frame/2", "text/html"));
 297 }
 298
 299 TEST_F(PageSerializerTest, CSS)
 300 {
 301     setBaseFolder("pageserializer/css/");
 302
 303     registerURL("css_test_page.html", "text/html");
 304     registerURL("link_styles.css", "text/css");
 305     registerURL("import_style_from_link.css", "text/css");
 306     registerURL("import_styles.css", "text/css");
 307     registerURL("do_not_serialize.png", "image.png", "image/png");
 308     registerURL("red_background.png", "image.png", "image/png");
 309     registerURL("orange_background.png", "image.png", "image/png");
 310     registerURL("yellow_background.png", "image.png", "image/png");
 311     registerURL("green_background.png", "image.png", "image/png");
 312     registerURL("blue_background.png", "image.png", "image/png");
 313     registerURL("purple_background.png", "image.png", "image/png");
 314     registerURL("pink_background.png", "image.png", "image/png");
 315     registerURL("brown_background.png", "image.png", "image/png");
 316     registerURL("ul-dot.png", "image.png", "image/png");
 317     registerURL("ol-dot.png", "image.png", "image/png");
 318
 319     serialize("css_test_page.html");
 320
 321     EXPECT_EQ(14U, getResources().size());
 322
 323     EXPECT_FALSE(isSerialized("do_not_serialize.png", "image/png"));
 324
 325     EXPECT_TRUE(isSerialized("css_test_page.html", "text/html"));
 326     EXPECT_TRUE(isSerialized("link_styles.css", "text/css"));
 327     EXPECT_TRUE(isSerialized("import_styles.css", "text/css"));
 328     EXPECT_TRUE(isSerialized("import_style_from_link.css", "text/css"));
 329     EXPECT_TRUE(isSerialized("red_background.png", "image/png"));
 330     EXPECT_TRUE(isSerialized("orange_background.png", "image/png"));
 331     EXPECT_TRUE(isSerialized("yellow_background.png", "image/png"));
 332     EXPECT_TRUE(isSerialized("green_background.png", "image/png"));
 333     EXPECT_TRUE(isSerialized("blue_background.png", "image/png"));
 334     EXPECT_TRUE(isSerialized("purple_background.png", "image/png"));
 335     EXPECT_TRUE(isSerialized("pink_background.png", "image/png"));
 336     EXPECT_TRUE(isSerialized("brown_background.png", "image/png"));
 337     EXPECT_TRUE(isSerialized("ul-dot.png", "image/png"));
 338     EXPECT_TRUE(isSerialized("ol-dot.png", "image/png"));
 339 }
 340
 341 TEST_F(PageSerializerTest, CSSImport)
 342 {
 343     setBaseFolder("pageserializer/css/");
 344
 345     registerURL("import.html", "text/html");
 346     registerURL("import/base.css", "text/css");
 347     registerURL("import/relative/red-background.css", "text/css");
 348     registerURL("import/absolute/green-header.css", "text/css");
 349
 350     serialize("import.html");
 351
 352     EXPECT_TRUE(isSerialized("import.html", "text/html"));
 353     EXPECT_TRUE(isSerialized("import/base.css", "text/css"));
 354     EXPECT_TRUE(isSerialized("import/relative/red-background.css", "text/css"));
 355     EXPECT_TRUE(isSerialized("import/absolute/green-header.css", "text/css"));
 356 }
 357
 358 TEST_F(PageSerializerTest, XMLDeclaration)
 359 {
 360     V8TestingScope scope(v8::Isolate::GetCurrent());
 361     setBaseFolder("pageserializer/xml/");
 362
 363     registerURL("xmldecl.xml", "text/xml");
 364     serialize("xmldecl.xml");
 365
 366     String expectedStart("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
 367     EXPECT_TRUE(getSerializedData("xmldecl.xml").startsWith(expectedStart));
 368 }
 369
 370 TEST_F(PageSerializerTest, DTD)
 371 {
 372     setBaseFolder("pageserializer/dtd/");
 373
 374     registerURL("html5.html", "text/html");
 375     serialize("html5.html");
 376
 377     String expectedStart("<!DOCTYPE html>");
 378     EXPECT_TRUE(getSerializedData("html5.html").startsWith(expectedStart));
 379 }
 380
 381 TEST_F(PageSerializerTest, Font)
 382 {
 383     setBaseFolder("pageserializer/font/");
 384
 385     registerURL("font.html", "text/html");
 386     registerURL("font.ttf", "application/octet-stream");
 387
 388     serialize("font.html");
 389
 390     EXPECT_TRUE(isSerialized("font.ttf", "application/octet-stream"));
 391 }
 392
 393 TEST_F(PageSerializerTest, DataURI)
 394 {
 395     setBaseFolder("pageserializer/datauri/");
 396
 397     registerURL("page_with_data.html", "text/html");
 398
 399     serialize("page_with_data.html");
 400
 401     EXPECT_EQ(1U, getResources().size());
 402     EXPECT_TRUE(isSerialized("page_with_data.html", "text/html"));
 403 }
 404
 405 TEST_F(PageSerializerTest, DataURIMorphing)
 406 {
 407     setBaseFolder("pageserializer/datauri/");
 408
 409     registerURL("page_with_morphing_data.html", "text/html");
 410
 411     serialize("page_with_morphing_data.html");
 412
 413     EXPECT_EQ(2U, getResources().size());
 414     EXPECT_TRUE(isSerialized("page_with_morphing_data.html", "text/html"));
 415 }
 416
 417 TEST_F(PageSerializerTest, RewriteLinksSimple)
 418 {
 419     setBaseFolder("pageserializer/rewritelinks/");
 420     setRewriteURLFolder("folder");
 421
 422     registerURL("rewritelinks_simple.html", "text/html");
 423     registerURL("absolute.png", "image.png", "image/png");
 424     registerURL("relative.png", "image.png", "image/png");
 425     registerRewriteURL("http://www.test.com/absolute.png", "a.png");
 426     registerRewriteURL("http://www.test.com/relative.png", "b.png");
 427
 428     serialize("rewritelinks_simple.html");
 429
 430     EXPECT_EQ(3U, getResources().size());
 431     EXPECT_NE(getSerializedData("rewritelinks_simple.html", "text/html").find("\"folder/a.png\""), kNotFound);
 432     EXPECT_NE(getSerializedData("rewritelinks_simple.html", "text/html").find("\"folder/b.png\""), kNotFound);
 433 }
 434
 435 TEST_F(PageSerializerTest, RewriteLinksBase)
 436 {
 437     setBaseFolder("pageserializer/rewritelinks/");
 438     setRewriteURLFolder("folder");
 439
 440     registerURL("rewritelinks_base.html", "text/html");
 441     registerURL("images/here/image.png", "image.png", "image/png");
 442     registerURL("images/here/or/in/here/image.png", "image.png", "image/png");
 443     registerURL("or/absolute.png", "image.png", "image/png");
 444     registerRewriteURL("http://www.test.com/images/here/image.png", "a.png");
 445     registerRewriteURL("http://www.test.com/images/here/or/in/here/image.png", "b.png");
 446     registerRewriteURL("http://www.test.com/or/absolute.png", "c.png");
 447
 448     serialize("rewritelinks_base.html");
 449
 450     EXPECT_EQ(4U, getResources().size());
 451     EXPECT_NE(getSerializedData("rewritelinks_base.html", "text/html").find("\"folder/a.png\""), kNotFound);
 452     EXPECT_NE(getSerializedData("rewritelinks_base.html", "text/html").find("\"folder/b.png\""), kNotFound);
 453     EXPECT_NE(getSerializedData("rewritelinks_base.html", "text/html").find("\"folder/c.png\""), kNotFound);
 454 }
 455
 456 // Test that we don't regress https://bugs.webkit.org/show_bug.cgi?id=99105
 457 TEST_F(PageSerializerTest, SVGImageDontCrash)
 458 {
 459     setBaseFolder("pageserializer/svg/");
 460
 461     registerURL("page_with_svg_image.html", "text/html");
 462     registerURL("green_rectangle.svg", "image/svg+xml");
 463
 464     serialize("page_with_svg_image.html");
 465
 466     EXPECT_EQ(2U, getResources().size());
 467
 468     EXPECT_TRUE(isSerialized("green_rectangle.svg", "image/svg+xml"));
 469     EXPECT_GT(getSerializedData("green_rectangle.svg", "image/svg+xml").length(), 250U);
 470 }
 471
 472 TEST_F(PageSerializerTest, DontIncludeErrorImage)
 473 {
 474     setBaseFolder("pageserializer/image/");
 475
 476     registerURL("page_with_img_error.html", "text/html");
 477     registerURL("error_image.png", "image/png");
 478
 479     serialize("page_with_img_error.html");
 480
 481     EXPECT_EQ(1U, getResources().size());
 482     EXPECT_TRUE(isSerialized("page_with_img_error.html", "text/html"));
 483     EXPECT_FALSE(isSerialized("error_image.png", "image/png"));
 484 }
 485
 486 TEST_F(PageSerializerTest, NamespaceElementsDontCrash)
 487 {
 488     setBaseFolder("pageserializer/namespace/");
 489
 490     registerURL("namespace_element.html", "text/html");
 491
 492     serialize("namespace_element.html");
 493
 494     EXPECT_EQ(1U, getResources().size());
 495     EXPECT_TRUE(isSerialized("namespace_element.html", "text/html"));
 496     EXPECT_GT(getSerializedData("namespace_element.html", "text/html").length(), 0U);
 497 }
 498
 499 } // namespace blink