HTTP/HTTPS and Java

Question

Why is the difference between the following classes? Why do different versions of these classes exist the way they do? (This question is based off of Sun's Java 1.3.1 release).
java.net.HttpURLConnection
<==>
sun.net.www.protocol.http.HttpURLConnection
sun.plugin.protocol.https.BrowserHttpsURLConnection
<==>
com.sun.net.ssl.internal.www.protocol.https.HttpsURLConnection
I'm writing code that needs to make HTTPS connections (I am using JSSE for this). I can make HTTP connections just fine, but the 'connect' member variable in URLConnection always remains false when I attempt an HTTPS connection. I get no status codes or error messages, just a dead connection. In an effort to debug this, I figured I'd take a peek out the source to get a hint as to what was going wrong. Anyway, I found myself more confused looking at the source, because there are different versions of things in different places. This is what led me to the question above.
Anyway help or insight would be much appreciated. Thanks.
Jake

Answer 1

the command-line looks like:
/opt/third-party/pkg/java-1.3.1_SUN/bin/java -classpath /opt/quality-ass
urance/lib/jars.current/HTTPClient.jar:/opt/quality-assurance/lib/jars.current/antlr.jar:/opt/quality-assurance/lib/jars.current/frank.jar:/opt/quality-assurance/lib/jars.current/org.jar:/opt/third-party/pkg/junit-3.2/jars/junit.jar:/opt/third-party/pkg/openxml-1.2/lib/openxml-1.2.jar:/opt/app/oracle/product/8.1.6/jdbc/lib/classes111.zip:/opt/third-party/jars/jsse.jar:/opt/third-party/jars/jnet.jar:/opt/third-party/jars/jcert.jar -Djava.library.path=/opt/quality-assurance/lib/jars.current qa.TestServer.GenericServer
the starting point where we are concerned is the run method of the class HttpTask. also i don't include the implementations, but methods like createFullyQualifiedURL() and originalURL(), etc return strings like 'http://www.company.com/one/two/three.html' or 'https://www.company.com/one/two/three.html' and also 'http://www.company.com:80/one/two/three.html' and
'https://www.company.com:443/one/two/three.html' depending on the startup params...
i know the details are sparse, but hopefully you understand the context. also i can assume the ca side of things is working, because i can use my browser to access the site via https. the code i'm writing which is somewhat functioning as a web client cannot access the site via https however.
jake
public class HttpTask extends SubTask {
// snip ..
public ResultIf run() {
// If we have already fetched, return NoResult
if (this.statusCode() != -1) {
return new NoResult(this.parent());
// Initialize URL
this.createFullyQualifiedURL();
if (this.originalURL().length() == 0) {
return this.defaultErrorResult(new Error("Empty URL"));
try {
// setup connection
HttpURLConnection connection = null;
connection = setupConnection();
connection.connect();
Log.log().debug("HTTP", "Fetching " + this.originalURL());
// process page
processPage(connection);
if (statusCode >= 300 && statusCode < 400) {
handleRedirect(connection);
} catch (Exception e) {
System.err.println(e);
return new HttpResult(this, true, "");
public HttpURLConnection setupConnection() throws Exception {
java.net.HttpURLConnection.setFollowRedirects(false);
java.security.Security.addProvider(new com.sun.net.ssl.internal.ssl.Pro\
vider());
java.lang.System.setProperty("java.protocol.handler.pkgs","com.sun.net.\
ssl.internal.www.protocol");
java.lang.System.setProperty("javax.net.debug", "ssl,handshake,data,tru\
stmanager");
HttpURLConnection connection = null;
URL tempURL = new URL(this.originalURL());
URLConnection uc = tempURL.openConnection();
connection = (HttpURLConnection) uc;
connection.setRequestProperty("User-Agent", "Frank/3.0");
// Deal w/ cookies:
this.setCookieForFetch(connection);
// FIX what should the behavior for this be?
// Send POST headers if any
if (this.method().startsWith("POST")) {
ResultIf result = this.doPost(connection);
// returns non-null only if error occurs.
if (result != null) {
return null;
return connection;
public void handleRedirect(HttpURLConnection connection) throws Exception {
while (statusCode >= 300 && statusCode < 400) {
// shouldn't need to handle this for POST case,
// some things may redirect on POST even though it's
// against the HTTP spec
if (method().equals("GETS") || method().equals("POSTS")) {
method("GETS");
} else {
method ("GET");
redirectURL = connection.getHeaderField("Location");
if (! redirectURL.startsWith("http")) {
if (method().equals("GETS") || method().equals("POSTS")) {
redirectURL = "https://" + this.baseURL() + ":" + myTask().variable("HTTPS_PORT") + redirectURL;
} else {
redirectURL = "http://" + this.baseURL() + ":" + myTask().variable("HTTP_PORT") + redirectURL;
this.originalURL(redirectURL);
connection = setupConnection();
connection.connect();
statusCode = connection.getResponseCode();
processPage(connection);
public void processPage(HttpURLConnection connection) throws Exception {
this.statusCode(connection.getResponseCode());
this.pageSize(connection.getContentLength());
if (this.getDynamicCookies()) {
String dCookies = getCookiesFromReply(connection);
Hashtable cookies = new Hashtable();
this.addCookiesFromString(cookies, dCookies);
// HACK: this is a quick tweak because I don't think
// it is a good thing to have for every fetch.
cookies.remove("path");
Hashtable dynamicCookies = Tool.cloneIt(this.dynamicCookies());
Tool.mergeHashes(dynamicCookies, cookies);
this.dynamicCookies(dynamicCookies);
// Get HTML if applicable
if (! this.method().equals("HEAD")) {
this.html(Tool.slurp(connection.getInputStream()));
// WARN: check to see if length is different from above pageSize
this.pageSize(html.length());
// snip...
}