package org.argeo.cms.internal.kernel;
+import static javax.jcr.Property.JCR_DESCRIPTION;
+import static javax.jcr.Property.JCR_LAST_MODIFIED;
+import static javax.jcr.Property.JCR_TITLE;
+import static org.argeo.cms.CmsTypes.CMS_IMAGE;
+
import java.io.IOException;
+import java.io.PrintWriter;
+import java.security.PrivilegedExceptionAction;
import java.security.cert.X509Certificate;
+import java.util.Calendar;
+import java.util.Collection;
import java.util.Enumeration;
+import javax.jcr.Node;
+import javax.jcr.NodeIterator;
+import javax.jcr.Repository;
+import javax.jcr.RepositoryException;
+import javax.jcr.Session;
+import javax.security.auth.Subject;
import javax.servlet.FilterChain;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServlet;
import org.apache.commons.logging.LogFactory;
import org.argeo.cms.CmsException;
import org.argeo.jcr.ArgeoJcrConstants;
-import org.eclipse.equinox.http.servlet.ExtendedHttpService;
+import org.argeo.jcr.JcrUtils;
+import org.osgi.framework.BundleContext;
+import org.osgi.framework.ServiceReference;
+import org.osgi.service.http.HttpService;
/**
* Intercepts and enriches http access, mainly focusing on security and
// private final DoSFilter dosFilter;
// private final QoSFilter qosFilter;
- NodeHttp(ExtendedHttpService httpService) {
+ private BundleContext bc;
+
+ NodeHttp(HttpService httpService, BundleContext bc) {
+ this.bc = bc;
// rootFilter = new RootFilter();
// dosFilter = new CustomDosFilter();
// qosFilter = new QoSFilter();
try {
httpService.registerServlet("/!", new LinkServlet(), null, null);
- // httpService.registerFilter("/", rootFilter, null, null);
+ httpService.registerServlet("/robots.txt", new RobotServlet(), null, null);
} catch (Exception e) {
throw new CmsException("Cannot register filters", e);
}
private static final long serialVersionUID = 3749990143146845708L;
@Override
- protected void service(HttpServletRequest request,
- HttpServletResponse response) throws ServletException,
- IOException {
+ protected void service(HttpServletRequest request, HttpServletResponse response)
+ throws ServletException, IOException {
String path = request.getPathInfo();
String userAgent = request.getHeader("User-Agent").toLowerCase();
boolean isBot = false;
boolean isCompatibleBrowser = false;
- if (userAgent.contains("bot")) {
+ if (userAgent.contains("bot") || userAgent.contains("facebook") || userAgent.contains("twitter")) {
isBot = true;
- } else if (userAgent.contains("webkit")
- || userAgent.contains("gecko")
- || userAgent.contains("firefox")
- || userAgent.contains("msie")
- || userAgent.contains("chrome")
- || userAgent.contains("chromium")
- || userAgent.contains("opera")
- || userAgent.contains("browser")) {
+ } else if (userAgent.contains("webkit") || userAgent.contains("gecko") || userAgent.contains("firefox")
+ || userAgent.contains("msie") || userAgent.contains("chrome") || userAgent.contains("chromium")
+ || userAgent.contains("opera") || userAgent.contains("browser")) {
isCompatibleBrowser = true;
}
- if (isCompatibleBrowser) {// redirect
- response.setHeader("Location", "/#" + path);
- response.setStatus(HttpServletResponse.SC_FOUND);
- } else {
- if (isBot && log.isDebugEnabled())
- log.debug(request.getHeader("User-Agent") + " is a bot");
- // TODO pure html
- throw new UnsupportedOperationException();
+ if (isBot) {
+ log.warn("# BOT " + request.getHeader("User-Agent"));
+ canonicalAnswer(request, response, path);
+ return;
+ }
+
+ if (isCompatibleBrowser && log.isTraceEnabled())
+ log.trace("# BWS " + request.getHeader("User-Agent"));
+ redirectTo(response, "/#" + path);
+ }
+
+ private void redirectTo(HttpServletResponse response, String location) {
+ response.setHeader("Location", location);
+ response.setStatus(HttpServletResponse.SC_FOUND);
+ }
+
+ // private boolean canonicalAnswerNeededBy(HttpServletRequest request) {
+ // String userAgent = request.getHeader("User-Agent").toLowerCase();
+ // return userAgent.startsWith("facebookexternalhit/");
+ // }
+
+ /** For bots which don't understand RWT. */
+ private void canonicalAnswer(HttpServletRequest request, HttpServletResponse response, String path) {
+ Session session = null;
+ try {
+ PrintWriter writer = response.getWriter();
+ session = Subject.doAs(KernelUtils.anonymousLogin(), new PrivilegedExceptionAction<Session>() {
+
+ @Override
+ public Session run() throws Exception {
+ Collection<ServiceReference<Repository>> srs = bc.getServiceReferences(Repository.class, "("
+ + ArgeoJcrConstants.JCR_REPOSITORY_ALIAS + "=" + ArgeoJcrConstants.ALIAS_NODE + ")");
+ Repository repository = bc.getService(srs.iterator().next());
+ return repository.login();
+ }
+
+ });
+ Node node = session.getNode(path);
+ String title = node.hasProperty(JCR_TITLE) ? node.getProperty(JCR_TITLE).getString() : node.getName();
+ String desc = node.hasProperty(JCR_DESCRIPTION) ? node.getProperty(JCR_DESCRIPTION).getString() : null;
+ Calendar lastUpdate = node.hasProperty(JCR_LAST_MODIFIED)
+ ? node.getProperty(JCR_LAST_MODIFIED).getDate() : null;
+ String url = KernelUtils.getCanonicalUrl(node, request);
+ String imgUrl = null;
+ loop: for (NodeIterator it = node.getNodes(); it.hasNext();) {
+ // Takes the first found cms:image
+ Node child = it.nextNode();
+ if (child.isNodeType(CMS_IMAGE)) {
+ imgUrl = KernelUtils.getDataUrl(child, request);
+ break loop;
+ }
+ }
+ StringBuilder buf = new StringBuilder();
+ buf.append("<html>");
+ buf.append("<head>");
+ writeMeta(buf, "og:title", escapeHTML(title));
+ writeMeta(buf, "og:type", "website");
+ buf.append("<meta name='twitter:card' content='summary' />");
+ buf.append("<meta name='twitter:site' content='@argeo_org' />");
+ writeMeta(buf, "og:url", url);
+ if (desc != null)
+ writeMeta(buf, "og:description", escapeHTML(desc));
+ if (imgUrl != null)
+ writeMeta(buf, "og:image", imgUrl);
+ if (lastUpdate != null)
+ writeMeta(buf, "og:updated_time", Long.toString(lastUpdate.getTime().getTime()));
+ buf.append("</head>");
+ buf.append("<body>");
+ buf.append(
+ "<p><b>!! This page is meant for indexing robots, not for real people," + " visit <a href='/#")
+ .append(path).append("'>").append(escapeHTML(title)).append("</a> instead.</b></p>");
+ writeCanonical(buf, node);
+ buf.append("</body>");
+ buf.append("</html>");
+ writer.print(buf.toString());
+
+ response.setHeader("Content-Type", "text/html");
+ writer.flush();
+ } catch (Exception e) {
+ throw new CmsException("Cannot write canonical answer", e);
+ } finally {
+ JcrUtils.logoutQuietly(session);
}
}
+
+ /**
+ * From
+ * http://stackoverflow.com/questions/1265282/recommended-method-for-
+ * escaping-html-in-java (+ escaping '). TODO Use
+ * org.apache.commons.lang.StringEscapeUtils
+ */
+ private String escapeHTML(String s) {
+ StringBuilder out = new StringBuilder(Math.max(16, s.length()));
+ for (int i = 0; i < s.length(); i++) {
+ char c = s.charAt(i);
+ if (c > 127 || c == '\'' || c == '"' || c == '<' || c == '>' || c == '&') {
+ out.append("&#");
+ out.append((int) c);
+ out.append(';');
+ } else {
+ out.append(c);
+ }
+ }
+ return out.toString();
+ }
+
+ private void writeMeta(StringBuilder buf, String tag, String value) {
+ buf.append("<meta property='").append(tag).append("' content='").append(value).append("'/>");
+ }
+
+ private void writeCanonical(StringBuilder buf, Node node) throws RepositoryException {
+ buf.append("<div>");
+ if (node.hasProperty(JCR_TITLE))
+ buf.append("<p>").append(node.getProperty(JCR_TITLE).getString()).append("</p>");
+ if (node.hasProperty(JCR_DESCRIPTION))
+ buf.append("<p>").append(node.getProperty(JCR_DESCRIPTION).getString()).append("</p>");
+ NodeIterator children = node.getNodes();
+ while (children.hasNext()) {
+ writeCanonical(buf, children.nextNode());
+ }
+ buf.append("</div>");
+ }
+ }
+
+ class RobotServlet extends HttpServlet {
+ private static final long serialVersionUID = 7935661175336419089L;
+
+ @Override
+ protected void service(HttpServletRequest request, HttpServletResponse response)
+ throws ServletException, IOException {
+ PrintWriter writer = response.getWriter();
+ writer.append("User-agent: *\n");
+ writer.append("Disallow:\n");
+ response.setHeader("Content-Type", "text/plain");
+ writer.flush();
+ }
+
}
/** Intercepts all requests. Authenticates. */
class RootFilter extends HttpFilter {
@Override
- public void doFilter(HttpSession httpSession,
- HttpServletRequest request, HttpServletResponse response,
+ public void doFilter(HttpSession httpSession, HttpServletRequest request, HttpServletResponse response,
FilterChain filterChain) throws IOException, ServletException {
if (log.isTraceEnabled()) {
- log.trace(request.getRequestURL().append(
- request.getQueryString() != null ? "?"
- + request.getQueryString() : ""));
+ log.trace(request.getRequestURL()
+ .append(request.getQueryString() != null ? "?" + request.getQueryString() : ""));
logRequest(request);
}
}
// redirect long RWT paths to anchor
- String path = request.getRequestURI().substring(
- servletPath.length());
+ String path = request.getRequestURI().substring(servletPath.length());
int pathLength = path.length();
- if (pathLength != 0 && (path.charAt(0) == '/')
- && !servletPath.endsWith("rwt-resources")
- && !path.startsWith(KernelConstants.PATH_WORKBENCH)
- && path.lastIndexOf('/') != 0) {
+ if (pathLength != 0 && (path.charAt(0) == '/') && !servletPath.endsWith("rwt-resources")
+ && !path.startsWith(KernelConstants.PATH_WORKBENCH) && path.lastIndexOf('/') != 0) {
String newLocation = request.getServletPath() + "#" + path;
response.setHeader("Location", newLocation);
response.setStatus(HttpServletResponse.SC_FOUND);
}
private X509Certificate extractCertificate(HttpServletRequest req) {
- X509Certificate[] certs = (X509Certificate[]) req
- .getAttribute("javax.servlet.request.X509Certificate");
+ X509Certificate[] certs = (X509Certificate[]) req.getAttribute("javax.servlet.request.X509Certificate");
if (null != certs && certs.length > 0) {
return certs[0];
}
projects / lgpl / argeo-commons.git / blobdiff