/**
* Tests if the {@link PageInstanceMapper} is used if {@link
* org.apache.wicket.settings.def.PageSettings#getRecreateMountedPagesAfterExpiry()} is disabled
*/
@Test
public void testLinkOnPageWithRecreationDisabled() {
tester.getApplication().getPageSettings().setRecreateMountedPagesAfterExpiry(false);
PageWithLink page =
tester.startPage(PageWithLink.class, new PageParameters().add("param", "value"));
Link<?> link = (Link<?>) page.get("link");
String url = link.getURL().toString();
assertEquals("./wicket/page?0-1.ILinkListener-link", url);
tester.executeUrl(url);
}
/** ... and this should throw a {@link PageExpiredException} if the page is expired */
@Test(expected = PageExpiredException.class)
public void testExpiredPageWithRecreationDisabled() {
tester.getApplication().getPageSettings().setRecreateMountedPagesAfterExpiry(false);
PageWithLink page =
tester.startPage(PageWithLink.class, new PageParameters().add("param", "value"));
Link<?> link = (Link<?>) page.get("link");
String url = link.getURL().toString();
assertEquals("./wicket/page?0-1.ILinkListener-link", url);
// simulate a page expiry
url = url.replace("page?0", "page?3");
tester.executeUrl(url);
}
public void run(String[] args) throws IOException {
System.out.println(
"Crawling for "
+ maxPages
+ " pages relevant to "
+ args[1]
+ " starting from "
+ startingUrl);
System.out.println("\n");
try {
URL url = new URL(startingUrl);
Link newLink = new Link(url, "");
knownUrls.put(url, newLink);
urlQueue.add(newLink);
allLinks.put(url, newLink);
} catch (MalformedURLException e) {
System.out.println("Invalid starting URL " + args[0]);
}
while ((!(urlQueue.size() == 0)) && (knownUrls.size() < maxPages)) {
Link newLink = urlQueue.poll();
URL newUrl = newLink.getURL();
if (debug) {
System.out.println(
"Downloading : " + newLink.getURL() + " with score = " + newLink.getScore());
}
if (!robotSafe(newUrl)) {
continue;
}
File dFile = downloadFile(newUrl);
if (dFile != null) {
if (debug) {
System.out.println("Received : " + newUrl);
}
// knownUrls.put(newUrl, newLink);
fetchAnchorLinks(dFile, newUrl);
}
}
}
public void fetchAnchorLinks(File file, URL parentUrl) throws IOException {
Document doc = Jsoup.parse(file, "UTF-8", parentUrl.toString());
Elements links = doc.select("a[href]");
for (Element link : links) {
String linkHref = link.attr("href");
String anchorText = link.text();
URL childUrl = new URL(parentUrl, linkHref);
Link newLink = new Link(childUrl, anchorText);
int score = computeScore(file, link, query, parentUrl);
if (knownUrls.size() == maxPages) {
break;
}
if (!knownUrls.containsKey(childUrl)) {
if (!urlQueue.contains(newLink)) {
newLink.setScore(score);
urlQueue.add(newLink);
// allLinks.put(childUrl, newLink);
if (debug) {
System.out.println(
"Adding to queue: " + newLink.getURL() + " with score = " + newLink.getScore());
}
}
} else {
Link retrievedLink = knownUrls.get(childUrl);
if (urlQueue.contains(retrievedLink)) {
int newScore = score + retrievedLink.getScore();
retrievedLink.setScore(newScore);
// knownUrls.put(childUrl,retrievedLink);
if (debug) {
System.out.println("Adding " + score + " to score of " + retrievedLink.getURL());
}
}
}
}
System.out.println("\n");
}
/**
* Tests if it is possible to re-instantiate the page if it is expired. The page should be
* instantiated with the same page parameters. The link will not be clicked however.
*/
@Test
public void testLinkOnExpiredPage() {
PageWithLink page =
tester.startPage(PageWithLink.class, new PageParameters().add("param", "value"));
assertEquals("value", page.getPageParameters().get("param").toString());
tester.assertContains("param=value");
Link<?> link = (Link<?>) page.get("link");
String url = link.getURL().toString();
// simulate a page expiry
url = url.replace("?0", "?3");
tester.executeUrl(url);
// request parameters to callback urls should be ignored for the re-created page
// (WICKET-4594)
tester.assertContainsNot("param=value");
}
/**
* Tests if the page parameters are part of the url of the link, and if the link actually works.
*/
@Test
public void testPageParametersInLink() {
PageWithLink page =
tester.startPage(PageWithLink.class, new PageParameters().add("param", "value"));
Link<?> link = (Link<?>) page.get("link");
String url = link.getURL().toString();
if (mount)
assertTrue(
"URL for link should contain 'mount/value/part2': " + url,
url.toString().contains("mount/value/part2"));
else
assertTrue(
"URL for link should contain 'param=value': " + url,
url.toString().contains("param=value"));
tester.executeUrl(url);
}
