Johannes/java-commons
Johannes
/
java-commons
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
java-commons/commons-http-server/src/main/java/io/gitlab/jfronny/commons/http/server/JLHTTPServer.java

1967 lines
86 KiB
Java
Raw Blame History

package io.gitlab.jfronny.commons.http.server;
import io.gitlab.jfronny.commons.http.server.api.ContextHandler;
import io.gitlab.jfronny.commons.http.server.io.ChunkedInputStream;
import io.gitlab.jfronny.commons.http.server.io.ChunkedOutputStream;
import io.gitlab.jfronny.commons.http.server.io.LimitedInputStream;
import io.gitlab.jfronny.commons.http.server.util.FileContextHandler;
import io.gitlab.jfronny.commons.http.server.util.Header;
import io.gitlab.jfronny.commons.http.server.util.Headers;
import io.gitlab.jfronny.commons.http.server.util.VirtualHost;
import javax.net.ServerSocketFactory;
import javax.net.ssl.SSLServerSocketFactory;
import javax.net.ssl.SSLSocket;
import java.io.*;
import java.net.*;
import java.nio.charset.StandardCharsets;
import java.text.ParseException;
import java.text.SimpleDateFormat;
import java.util.*;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.Executor;
import java.util.concurrent.Executors;
import java.util.zip.DeflaterOutputStream;
import java.util.zip.GZIPOutputStream;
/**
* The {@code HTTPServer} class implements a light-weight HTTP server.
* <p>
* This server implements all functionality required by RFC 2616 ("Hypertext
* Transfer Protocol -- HTTP/1.1"), as well as some of the optional
* functionality (this is termed "conditionally compliant" in the RFC).
* In fact, a couple of bugs in the RFC itself were discovered
* (and fixed) during the development of this server.
* <p>
* <b>Feature Overview</b>
* <ul>
* <li>RFC compliant - correctness is not sacrificed for the sake of size</li>
* <li>Virtual hosts - multiple domains and subdomains per server</li>
* <li>File serving - built-in handler to serve files and folders from disk</li>
* <li>Mime type mappings - configurable via API or a standard mime.types file</li>
* <li>Directory index generation - enables browsing folder contents</li>
* <li>Welcome files - configurable default filename (e.g. index.html)</li>
* <li>All HTTP methods supported - GET/HEAD/OPTIONS/TRACE/POST/PUT/DELETE/custom</li>
* <li>Conditional statuses - ETags and If-* header support</li>
* <li>Chunked transfer encoding - for serving dynamically-generated data streams</li>
* <li>Gzip/deflate compression - reduces bandwidth and download time</li>
* <li>HTTPS - secures all server communications</li>
* <li>Partial content - download continuation (a.k.a. byte range serving)</li>
* <li>File upload - multipart/form-data handling as stream or iterator</li>
* <li>Multiple context handlers - a different handler method per URL path</li>
* <li>@Context annotations - auto-detection of context handler methods</li>
* <li>Parameter parsing - from query string or x-www-form-urlencoded body</li>
* <li>A single source file - super-easy to integrate into any application</li>
* <li>Standalone - no dependencies other than the Java runtime</li>
* <li>Small footprint - standard jar is ~50K, stripped jar is ~35K</li>
* <li>Extensible design - easy to override, add or remove functionality</li>
* <li>Reusable utility methods to simplify your custom code</li>
* <li>Extensive documentation of API and implementation (&gt;40% of source lines)</li>
* </ul>
* <p>
* <b>Use Cases</b>
* <p>
* Being a lightweight, standalone, easily embeddable and tiny-footprint
* server, it is well-suited for
* <ul>
* <li>Resource-constrained environments such as embedded devices.
* For really extreme constraints, you can easily remove unneeded
* functionality to make it even smaller (and use the -Dstripped
* maven build option to strip away debug info, license, etc.)</li>
* <li>Unit and integration tests - fast setup/teardown times, small overhead
* and simple context handler setup make it a great web server for testing
* client components under various server response conditions.</li>
* <li>Embedding a web console into any headless application for
* administration, monitoring, or a full portable GUI.</li>
* <li>A full-fledged standalone web server serving static files,
* dynamically-generated content, REST APIs, pseudo-streaming, etc.</li>
* <li>A good reference for learning how HTTP works under the hood.</li>
* </ul>
* <p>
* <b>Implementation Notes</b>
* <p>
* The design and implementation of this server attempt to balance correctness,
* compliance, readability, size, features, extensibility and performance,
* and often prioritize them in this order, but some trade-offs must be made.
* <p>
* This server is multithreaded in its support for multiple concurrent HTTP
* connections, however most of its constituent classes are not thread-safe and
* require external synchronization if accessed by multiple threads concurrently.
* <p>
* <b>Source Structure and Documentation</b>
* <p>
* This server is intentionally written as a single source file, in order to make
* it as easy as possible to integrate into any existing project - by simply adding
* this single file to the project sources. It does, however, aim to maintain a
* structured and flexible design. There are no external package dependencies.
* <p>
* This file contains extensive documentation of its classes and methods, as
* well as implementation details and references to specific RFC sections
* which clarify the logic behind the code. It is recommended that anyone
* attempting to modify the protocol-level functionality become acquainted with
* the RFC, in order to make sure that protocol compliance is not broken.
* <p>
* <b>Getting Started</b>
* <p>
* For an example and a good starting point for learning how to use the API,
* see the {@link #main main} method at the bottom of the file, and follow
* the code into the API from there. Alternatively, you can just browse through
* the classes and utility methods and read their documentation and code.
*
* @author Amichai Rothman
* @since 2008-07-24
*/
public class JLHTTPServer {
private static final System.Logger LOGGER = System.getLogger("commons-http-server");
/**
* The SimpleDateFormat-compatible formats of dates which must be supported.
* Note that all generated date fields must be in the RFC 1123 format only,
* while the others are supported by recipients for backwards-compatibility.
*/
public static final String[] DATE_PATTERNS = {
"EEE, dd MMM yyyy HH:mm:ss z", // RFC 822, updated by RFC 1123
"EEEE, dd-MMM-yy HH:mm:ss z", // RFC 850, obsoleted by RFC 1036
"EEE MMM d HH:mm:ss yyyy" // ANSI C's asctime() format
};
/** A GMT (UTC) timezone instance. */
protected static final TimeZone GMT = TimeZone.getTimeZone("GMT");
/** Date format strings. */
protected static final char[]
DAYS = "Sun Mon Tue Wed Thu Fri Sat".toCharArray(),
MONTHS = "Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec".toCharArray();
/** A convenience array containing the carriage-return and line feed chars. */
public static final byte[] CRLF = { 0x0d, 0x0a };
/** The HTTP status description strings. */
protected static final String[] statuses = new String[600];
static {
// initialize status descriptions lookup table
Arrays.fill(statuses, "Unknown Status");
statuses[100] = "Continue";
statuses[200] = "OK";
statuses[204] = "No Content";
statuses[206] = "Partial Content";
statuses[301] = "Moved Permanently";
statuses[302] = "Found";
statuses[304] = "Not Modified";
statuses[307] = "Temporary Redirect";
statuses[400] = "Bad Request";
statuses[401] = "Unauthorized";
statuses[403] = "Forbidden";
statuses[404] = "Not Found";
statuses[405] = "Method Not Allowed";
statuses[408] = "Request Timeout";
statuses[412] = "Precondition Failed";
statuses[413] = "Request Entity Too Large";
statuses[414] = "Request-URI Too Large";
statuses[416] = "Requested Range Not Satisfiable";
statuses[417] = "Expectation Failed";
statuses[500] = "Internal Server Error";
statuses[501] = "Not Implemented";
statuses[502] = "Bad Gateway";
statuses[503] = "Service Unavailable";
statuses[504] = "Gateway Time-out";
}
/**
* A mapping of path suffixes (e.g. file extensions) to their
* corresponding MIME types.
*/
protected static final Map<String, String> contentTypes =
new ConcurrentHashMap<>();
static {
// add some default common content types
// see http://www.iana.org/assignments/media-types/ for full list
addContentType("application/font-woff", "woff");
addContentType("application/font-woff2", "woff2");
addContentType("application/java-archive", "jar");
addContentType("application/javascript", "js");
addContentType("application/json", "json");
addContentType("application/octet-stream", "exe");
addContentType("application/pdf", "pdf");
addContentType("application/x-7z-compressed", "7z");
addContentType("application/x-compressed", "tgz");
addContentType("application/x-gzip", "gz");
addContentType("application/x-tar", "tar");
addContentType("application/xhtml+xml", "xhtml");
addContentType("application/zip", "zip");
addContentType("audio/mpeg", "mp3");
addContentType("image/gif", "gif");
addContentType("image/jpeg", "jpg", "jpeg");
addContentType("image/png", "png");
addContentType("image/svg+xml", "svg");
addContentType("image/x-icon", "ico");
addContentType("text/css", "css");
addContentType("text/csv", "csv");
addContentType("text/html; charset=utf-8", "htm", "html");
addContentType("text/plain", "txt", "text", "log");
addContentType("text/xml", "xml");
}
/** The MIME types that can be compressed (prefix/suffix wildcards allowed). */
protected static String[] compressibleContentTypes =
{ "text/*", "*/javascript", "*icon", "*+xml", "*/json" };
/**
* The {@code Request} class encapsulates a single HTTP request.
*/
public class Request {
protected String method;
protected URI uri;
protected URL baseURL; // cached value
protected String version;
protected Headers headers;
protected InputStream body;
protected Map<String, String> params; // cached value
protected VirtualHost host; // cached value
protected VirtualHost.ContextInfo context; // cached value
/**
* Constructs a Request from the data in the given input stream.
*
* @param in the input stream from which the request is read
* @throws IOException if an error occurs
*/
public Request(InputStream in) throws IOException {
readRequestLine(in);
headers = readHeaders(in);
// RFC2616#3.6 - if "chunked" is used, it must be the last one
// RFC2616#4.4 - if non-identity Transfer-Encoding is present,
// it must either include "chunked" or close the connection after
// the body, and in any case ignore Content-Length.
// if there is no such Transfer-Encoding, use Content-Length
// if neither header exists, there is no body
String header = headers.get("Transfer-Encoding");
if (header != null && !header.toLowerCase(Locale.US).equals("identity")) {
if (Arrays.asList(splitElements(header, true)).contains("chunked"))
body = new ChunkedInputStream(in, headers);
else
body = in; // body ends when connection closes
} else {
header = headers.get("Content-Length");
long len = header == null ? 0 : parseULong(header, 10);
body = new LimitedInputStream(in, len, false);
}
}
/**
* Returns the request method.
*
* @return the request method
*/
public String getMethod() { return method; }
/**
* Returns the request URI.
*
* @return the request URI
*/
public URI getURI() { return uri; }
/**
* Returns the request version string.
*
* @return the request version string
*/
public String getVersion() { return version; }
/**
* Returns the request headers.
*
* @return the request headers
*/
public Headers getHeaders() { return headers; }
/**
* Returns the input stream containing the request body.
*
* @return the input stream containing the request body
*/
public InputStream getBody() { return body; }
/**
* Returns the path component of the request URI, after
* URL decoding has been applied (using the UTF-8 charset).
*
* @return the decoded path component of the request URI
*/
public String getPath() {
return uri.getPath();
}
/**
* Sets the path component of the request URI. This can be useful
* in URL rewriting, etc.
*
* @param path the path to set
* @throws IllegalArgumentException if the given path is malformed
*/
public void setPath(String path) {
try {
uri = new URI(uri.getScheme(), uri.getUserInfo(), uri.getHost(), uri.getPort(),
trimDuplicates(path, '/'), uri.getQuery(), uri.getFragment());
context = null; // clear cached context so it will be recalculated
} catch (URISyntaxException use) {
throw new IllegalArgumentException("error setting path", use);
}
}
/**
* Returns the base URL (scheme, host and port) of the request resource.
* The host name is taken from the request URI or the Host header or a
* default host (see RFC2616#5.2).
*
* @return the base URL of the requested resource, or null if it
* is malformed
*/
public URL getBaseURL() {
if (baseURL != null)
return baseURL;
// normalize host header
String host = uri.getHost();
if (host == null) {
host = headers.get("Host");
if (host == null) // missing in HTTP/1.0
host = detectLocalHostName();
}
int pos = host.indexOf(':');
host = pos < 0 ? host : host.substring(0, pos);
try {
return baseURL = new URL(secure ? "https" : "http", host, port, "");
} catch (MalformedURLException mue) {
return null;
}
}
/**
* Returns the request parameters, which are parsed both from the query
* part of the request URI, and from the request body if its content
* type is "application/x-www-form-urlencoded" (i.e. a submitted form).
* UTF-8 encoding is assumed in both cases.
* <p>
* The parameters are returned as a list of string arrays, each containing
* the parameter name as the first element and its corresponding value
* as the second element (or an empty string if there is no value).
* <p>
* The list retains the original order of the parameters.
*
* @return the request parameters name-value pairs,
* or an empty list if there are none
* @throws IOException if an error occurs
* @see JLHTTPServer#parseParamsList(String)
*/
public List<String[]> getParamsList() throws IOException {
List<String[]> queryParams = parseParamsList(uri.getRawQuery());
List<String[]> bodyParams = Collections.emptyList();
String ct = headers.get("Content-Type");
if (ct != null && ct.toLowerCase(Locale.US).startsWith("application/x-www-form-urlencoded"))
bodyParams = parseParamsList(readToken(body, -1, "UTF-8", 2097152)); // 2MB limit
if (bodyParams.isEmpty())
return queryParams;
if (queryParams.isEmpty())
return bodyParams;
queryParams.addAll(bodyParams);
return queryParams;
}
/**
* Returns the request parameters, which are parsed both from the query
* part of the request URI, and from the request body if its content
* type is "application/x-www-form-urlencoded" (i.e. a submitted form).
* UTF-8 encoding is assumed in both cases.
* <p>
* For multivalued parameters (i.e. multiple parameters with the same
* name), only the first one is considered. For access to all values,
* use {@link #getParamsList()} instead.
* <p>
* The map iteration retains the original order of the parameters.
*
* @return the request parameters name-value pairs,
* or an empty map if there are none
* @throws IOException if an error occurs
* @see #getParamsList()
*/
public Map<String, String> getParams() throws IOException {
if (params == null)
params = toMap(getParamsList());
return params;
}
/**
* Returns the absolute (zero-based) content range value read
* from the Range header. If multiple ranges are requested, a single
* range containing all of them is returned.
*
* @param length the full length of the requested resource
* @return the requested range, or null if the Range header
* is missing or invalid
*/
public long[] getRange(long length) {
String header = headers.get("Range");
return header == null || !header.startsWith("bytes=")
? null : parseRange(header.substring(6), length);
}
/**
* Reads the request line, parsing the method, URI and version string.
*
* @param in the input stream from which the request line is read
* @throws IOException if an error occurs or the request line is invalid
*/
protected void readRequestLine(InputStream in) throws IOException {
// RFC2616#4.1: should accept empty lines before request line
// RFC2616#19.3: tolerate additional whitespace between tokens
String line;
try {
do { line = readLine(in); } while (line.length() == 0);
} catch (IOException ioe) { // if EOF, timeout etc.
throw new IOException("missing request line"); // signal that the request did not begin
}
String[] tokens = split(line, " ", -1);
if (tokens.length != 3)
throw new IOException("invalid request line: \"" + line + "\"");
try {
method = tokens[0];
// must remove '//' prefix which constructor parses as host name
uri = new URI(trimDuplicates(tokens[1], '/'));
version = tokens[2]; // RFC2616#2.1: allow implied LWS; RFC7230#3.1.1: disallow it
} catch (URISyntaxException use) {
throw new IOException("invalid URI: " + use.getMessage());
}
}
/**
* Returns the virtual host corresponding to the requested host name,
* or the default host if none exists.
*
* @return the virtual host corresponding to the requested host name,
* or the default virtual host
*/
public VirtualHost getVirtualHost() {
if (host != null) return host;
URL baseUrl = getBaseURL();
if (baseUrl != null && (host = JLHTTPServer.this.getVirtualHost(baseUrl.getHost())) != null) return host;
return host = JLHTTPServer.this.getVirtualHost(null);
}
/**
* Returns the info of the context handling this request.
*
* @return the info of the context handling this request, or an empty context
*/
public VirtualHost.ContextInfo getContext() {
return context != null ? context : (context = getVirtualHost().getContext(getPath()));
}
}
/**
* The {@code Response} class encapsulates a single HTTP response.
*/
public class Response implements Closeable {
protected OutputStream out; // the underlying output stream
protected OutputStream[] encoders = new OutputStream[4]; // chained encoder streams
protected Headers headers;
protected boolean discardBody;
protected int state; // nothing sent, headers sent, or closed
protected Request req; // request used in determining client capabilities
/**
* Constructs a Response whose output is written to the given stream.
*
* @param out the stream to which the response is written
*/
public Response(OutputStream out) {
this.out = out;
this.headers = new Headers();
}
/**
* Sets whether this response's body is discarded or sent.
*
* @param discardBody specifies whether the body is discarded or not
*/
public void setDiscardBody(boolean discardBody) {
this.discardBody = discardBody;
}
/**
* Sets the request which is used in determining the capabilities
* supported by the client (e.g. compression, encoding, etc.)
*
* @param req the request
*/
public void setClientCapabilities(Request req) { this.req = req; }
/**
* Returns the request headers collection.
*
* @return the request headers collection
*/
public Headers getHeaders() { return headers; }
/**
* Returns the underlying output stream to which the response is written.
* Except for special cases, you should use {@link #getBody()} instead.
*
* @return the underlying output stream to which the response is written
*/
public OutputStream getOutputStream() { return out; }
/**
* Returns whether the response headers were already sent.
*
* @return whether the response headers were already sent
*/
public boolean headersSent() { return state == 1; }
/**
* Returns an output stream into which the response body can be written.
* The stream applies encodings (e.g. compression) according to the sent headers.
* This method must be called after response headers have been sent
* that indicate there is a body. Normally, the content should be
* prepared (not sent) even before the headers are sent, so that any
* errors during processing can be caught and a proper error response returned -
* after the headers are sent, it's too late to change the status into an error.
*
* @return an output stream into which the response body can be written,
* or null if the body should not be written (e.g. it is discarded)
* @throws IOException if an error occurs
*/
public OutputStream getBody() throws IOException {
if (encoders[0] != null || discardBody)
return encoders[0]; // return the existing stream (or null)
// set up chain of encoding streams according to headers
List<String> te = Arrays.asList(splitElements(headers.get("Transfer-Encoding"), true));
List<String> ce = Arrays.asList(splitElements(headers.get("Content-Encoding"), true));
int i = encoders.length - 1;
encoders[i] = new FilterOutputStream(out) {
@Override
public void close() {} // keep underlying connection stream open for now
@Override // override the very inefficient default implementation
public void write(byte[] b, int off, int len) throws IOException { out.write(b, off, len); }
};
if (te.contains("chunked"))
encoders[--i] = new ChunkedOutputStream(encoders[i + 1]);
if (ce.contains("gzip") || te.contains("gzip"))
encoders[--i] = new GZIPOutputStream(encoders[i + 1], 4096);
else if (ce.contains("deflate") || te.contains("deflate"))
encoders[--i] = new DeflaterOutputStream(encoders[i + 1]);
encoders[0] = encoders[i];
encoders[i] = null; // prevent duplicate reference
return encoders[0]; // returned stream is always first
}
/**
* Closes this response and flushes all output.
*
* @throws IOException if an error occurs
*/
public void close() throws IOException {
state = -1; // closed
if (encoders[0] != null)
encoders[0].close(); // close all chained streams (except the underlying one)
out.flush(); // always flush underlying stream (even if getBody was never called)
}
/**
* Sends the response headers with the given response status.
* A Date header is added if it does not already exist.
* If the response has a body, the Content-Length/Transfer-Encoding
* and Content-Type headers must be set before sending the headers.
*
* @param status the response status
* @throws IOException if an error occurs or headers were already sent
* @see #sendHeaders(int, long, long, String, String, long[])
*/
public void sendHeaders(int status) throws IOException {
if (headersSent())
throw new IOException("headers were already sent");
if (!headers.contains("Date"))
headers.add("Date", formatDate(System.currentTimeMillis()));
headers.add("Server", "JLHTTP/2.5");
out.write(getBytes("HTTP/1.1 ", Integer.toString(status), " ", statuses[status]));
out.write(CRLF);
headers.writeTo(out);
state = 1; // headers sent
}
/**
* Sends the response headers, including the given response status
* and description, and all response headers. If they do not already
* exist, the following headers are added as necessary:
* Content-Range, Content-Type, Transfer-Encoding, Content-Encoding,
* Content-Length, Last-Modified, ETag, Connection and Date. Ranges are
* properly calculated as well, with a 200 status changed to a 206 status.
*
* @param status the response status
* @param length the response body length, or zero if there is no body,
* or negative if there is a body but its length is not yet known
* @param lastModified the last modified date of the response resource,
* or non-positive if unknown. A time in the future will be
* replaced with the current system time.
* @param etag the ETag of the response resource, or null if unknown
* (see RFC2616#3.11)
* @param contentType the content type of the response resource, or null
* if unknown (in which case "application/octet-stream" will be sent)
* @param range the content range that will be sent, or null if the
* entire resource will be sent
* @throws IOException if an error occurs
*/
public void sendHeaders(int status, long length, long lastModified,
String etag, String contentType, long[] range) throws IOException {
if (range != null) {
headers.add("Content-Range", "bytes " + range[0] + "-" +
range[1] + "/" + (length >= 0 ? length : "*"));
length = range[1] - range[0] + 1;
if (status == 200)
status = 206;
}
String ct = headers.get("Content-Type");
if (ct == null) {
ct = contentType != null ? contentType : "application/octet-stream";
headers.add("Content-Type", ct);
}
if (!headers.contains("Content-Length") && !headers.contains("Transfer-Encoding")) {
// RFC2616#3.6: transfer encodings are case-insensitive and must not be sent to an HTTP/1.0 client
boolean modern = req != null && req.getVersion().endsWith("1.1");
String accepted = req == null ? null : req.getHeaders().get("Accept-Encoding");
List<String> encodings = Arrays.asList(splitElements(accepted, true));
String compression = encodings.contains("gzip") ? "gzip" :
encodings.contains("deflate") ? "deflate" : null;
if (compression != null && (length < 0 || length > 300) && isCompressible(ct) && modern) {
headers.add("Transfer-Encoding", "chunked"); // compressed data is always unknown length
headers.add("Content-Encoding", compression);
} else if (length < 0 && modern) {
headers.add("Transfer-Encoding", "chunked"); // unknown length
} else if (length >= 0) {
headers.add("Content-Length", Long.toString(length)); // known length
}
}
if (!headers.contains("Vary")) // RFC7231#7.1.4: Vary field should include headers
headers.add("Vary", "Accept-Encoding"); // that are used in selecting representation
if (lastModified > 0 && !headers.contains("Last-Modified")) // RFC2616#14.29
headers.add("Last-Modified", formatDate(Math.min(lastModified, System.currentTimeMillis())));
if (etag != null && !headers.contains("ETag"))
headers.add("ETag", etag);
if (req != null && "close".equalsIgnoreCase(req.getHeaders().get("Connection"))
&& !headers.contains("Connection"))
headers.add("Connection", "close"); // #RFC7230#6.6: should reply to close with close
sendHeaders(status);
}
/**
* Sends the full response with the given status, and the given string
* as the body. The text is sent in the UTF-8 charset. If a
* Content-Type header was not explicitly set, it will be set to
* text/html, and so the text must contain valid (and properly
* {@link JLHTTPServer#escapeHTML escaped}) HTML.
*
* @param status the response status
* @param text the text body (sent as text/html)
* @throws IOException if an error occurs
*/
public void send(int status, String text) throws IOException {
byte[] content = text.getBytes(StandardCharsets.UTF_8);
sendHeaders(status, content.length, -1,
"W/\"" + Integer.toHexString(text.hashCode()) + "\"",
"text/html; charset=utf-8", null);
OutputStream out = getBody();
if (out != null)
out.write(content);
}
/**
* Sends an error response with the given status and detailed message.
* An HTML body is created containing the status and its description,
* as well as the message, which is escaped using the
* {@link JLHTTPServer#escapeHTML escape} method.
*
* @param status the response status
* @param text the text body (sent as text/html)
* @throws IOException if an error occurs
*/
public void sendError(int status, String text) throws IOException {
send(status, String.format(
"<!DOCTYPE html>%n<html>%n<head><title>%d %s</title></head>%n" +
"<body><h1>%d %s</h1>%n<p>%s</p>%n</body></html>",
status, statuses[status], status, statuses[status], escapeHTML(text)));
}
/**
* Sends an error response with the given status and default body.
*
* @param status the response status
* @throws IOException if an error occurs
*/
public void sendError(int status) throws IOException {
String text = status < 400 ? ":)" : "sorry it didn't work out :(";
sendError(status, text);
}
/**
* Sends the response body. This method must be called only after the
* response headers have been sent (and indicate that there is a body).
*
* @param body a stream containing the response body
* @param length the full length of the response body, or -1 for the whole stream
* @param range the sub-range within the response body that should be
* sent, or null if the entire body should be sent
* @throws IOException if an error occurs
*/
public void sendBody(InputStream body, long length, long[] range) throws IOException {
OutputStream out = getBody();
if (out != null) {
if (range != null) {
long offset = range[0];
length = range[1] - range[0] + 1;
while (offset > 0) {
long skip = body.skip(offset);
if (skip == 0)
throw new IOException("can't skip to " + range[0]);
offset -= skip;
}
}
transfer(body, out, length);
}
}
/**
* Sends a 301 or 302 response, redirecting the client to the given URL.
*
* @param url the absolute URL to which the client is redirected
* @param permanent specifies whether a permanent (301) or
* temporary (302) redirect status is sent
* @throws IOException if an IO error occurs or url is malformed
*/
public void redirect(String url, boolean permanent) throws IOException {
try {
url = new URI(url).toASCIIString();
} catch (URISyntaxException e) {
throw new IOException("malformed URL: " + url);
}
headers.add("Location", url);
// some user-agents expect a body, so we send it
if (permanent)
sendError(301, "Permanently moved to " + url);
else
sendError(302, "Temporarily moved to " + url);
}
}
/**
* The {@code SocketHandlerThread} handles accepted sockets.
*/
protected class SocketHandlerThread extends Thread {
@Override
public void run() {
setName(getClass().getSimpleName() + "-" + port);
try {
ServerSocket serv = JLHTTPServer.this.serv; // keep local to avoid NPE when stopped
while (serv != null && !serv.isClosed()) {
final Socket sock = serv.accept();
executor.execute(() -> {
try {
try {
sock.setSoTimeout(socketTimeout);
sock.setTcpNoDelay(true); // we buffer anyway, so improve latency
handleConnection(sock.getInputStream(), sock.getOutputStream());
} finally {
try {
// RFC7230#6.6 - close socket gracefully
// (except SSL socket which doesn't support half-closing)
if (!(sock instanceof SSLSocket)) {
sock.shutdownOutput(); // half-close socket (only output)
transfer(sock.getInputStream(), null, -1); // consume input
}
} finally {
sock.close(); // and finally close socket fully
}
}
} catch (IOException ignore) {}
});
}
} catch (IOException ignore) {}
}
}
protected volatile int port;
protected volatile int socketTimeout = 10000;
protected volatile ServerSocketFactory serverSocketFactory;
protected volatile boolean secure;
protected volatile Executor executor;
protected volatile ServerSocket serv;
protected final Map<String, VirtualHost> hosts = new ConcurrentHashMap<>();
/**
* Constructs an HTTPServer which can accept connections on the given port.
* Note: the {@link #start()} method must be called to start accepting
* connections.
*
* @param port the port on which this server will accept connections
*/
public JLHTTPServer(int port) {
setPort(port);
addVirtualHost(new VirtualHost(null)); // add default virtual host
}
/**
* Constructs an HTTPServer which can accept connections on the default HTTP port 80.
* Note: the {@link #start()} method must be called to start accepting connections.
*/
public JLHTTPServer() {
this(80);
}
/**
* Sets the port on which this server will accept connections.
*
* @param port the port on which this server will accept connections
*/
public void setPort(int port) {
this.port = port;
}
/**
* Sets the factory used to create the server socket.
* If null or not set, the default {@link ServerSocketFactory#getDefault()} is used.
* For secure sockets (HTTPS), use an SSLServerSocketFactory instance.
* The port should usually also be changed for HTTPS, e.g. port 443 instead of 80.
* <p>
* If using the default SSLServerSocketFactory returned by
* {@link SSLServerSocketFactory#getDefault()}, the appropriate system properties
* must be set to configure the default JSSE provider, such as
* {@code javax.net.ssl.keyStore} and {@code javax.net.ssl.keyStorePassword}.
*
* @param factory the server socket factory to use
*/
public void setServerSocketFactory(ServerSocketFactory factory) {
this.serverSocketFactory = factory;
this.secure = factory instanceof SSLServerSocketFactory;
}
/**
* Sets the socket timeout for established connections.
*
* @param timeout the socket timeout in milliseconds
*/
public void setSocketTimeout(int timeout) { this.socketTimeout = timeout; }
/**
* Sets the executor used in servicing HTTP connections.
* If null, a default executor is used. The caller is responsible
* for shutting down the provided executor when necessary.
*
* @param executor the executor to use
*/
public void setExecutor(Executor executor) {
this.executor = executor;
}
/**
* Returns the virtual host with the given name.
*
* @param name the name of the virtual host to return,
* or null for the default virtual host
* @return the virtual host with the given name, or null if it doesn't exist
*/
public VirtualHost getVirtualHost(String name) {
return hosts.get(name == null ? "" : name);
}
/**
* Returns all virtual hosts.
*
* @return all virtual hosts (as an unmodifiable set)
*/
public Set<VirtualHost> getVirtualHosts() {
return Set.copyOf(hosts.values());
}
/**
* Adds the given virtual host to the server.
* If the host's name or aliases already exist, they are overwritten.
*
* @param host the virtual host to add
*/
public void addVirtualHost(VirtualHost host) {
String name = host.getName();
hosts.put(name == null ? "" : name, host);
}
/**
* Creates the server socket used to accept connections, using the configured
* {@link #setServerSocketFactory ServerSocketFactory} and {@link #setPort port}.
* <p>
* Cryptic errors seen here often mean the factory configuration details are wrong.
*
* @return the created server socket
* @throws IOException if the socket cannot be created
*/
protected ServerSocket createServerSocket() throws IOException {
ServerSocket serv = serverSocketFactory.createServerSocket();
serv.setReuseAddress(true);
serv.bind(new InetSocketAddress(port));
return serv;
}
/**
* Starts this server. If it is already started, does nothing.
* Note: Once the server is started, configuration-altering methods
* of the server and its virtual hosts must not be used. To modify the
* configuration, the server must first be stopped.
*
* @throws IOException if the server cannot begin accepting connections
*/
public synchronized void start() throws IOException {
if (serv != null)
return;
if (serverSocketFactory == null) // assign default server socket factory if needed
serverSocketFactory = ServerSocketFactory.getDefault(); // plain sockets
serv = createServerSocket();
if (executor == null) // assign default executor if needed
executor = Executors.newCachedThreadPool(); // consumes no resources when idle
// register all host aliases (which may have been modified)
for (VirtualHost host : getVirtualHosts())
for (String alias : host.getAliases())
hosts.put(alias, host);
// start handling incoming connections
new SocketHandlerThread().start();
}
/**
* Stops this server. If it is already stopped, does nothing.
* Note that if an {@link #setExecutor Executor} was set, it must be closed separately.
*/
public synchronized void stop() {
try {
if (serv != null)
serv.close();
} catch (IOException ignore) {}
serv = null;
}
/**
* Handles communications for a single connection over the given streams.
* Multiple subsequent transactions are handled on the connection,
* until the streams are closed, an error occurs, or the request
* contains a "Connection: close" header which explicitly requests
* the connection be closed after the transaction ends.
*
* @param in the stream from which the incoming requests are read
* @param out the stream into which the outgoing responses are written
* @throws IOException if an error occurs
*/
protected void handleConnection(InputStream in, OutputStream out) throws IOException {
in = new BufferedInputStream(in, 4096);
out = new BufferedOutputStream(out, 4096);
Request req;
Response resp;
do {
// create request and response and handle transaction
req = null;
resp = new Response(out);
try {
req = new Request(in);
handleTransaction(req, resp);
} catch (Throwable t) { // unhandled errors (not normal error responses like 404)
if (req == null) { // error reading request
if (t instanceof IOException && t.getMessage().contains("missing request line"))
break; // we're not in the middle of a transaction - so just disconnect
resp.getHeaders().add("Connection", "close"); // about to close connection
if (t instanceof InterruptedIOException) // e.g. SocketTimeoutException
resp.sendError(408, "Timeout waiting for client request");
else
resp.sendError(400, "Invalid request: " + t.getMessage());
} else if (!resp.headersSent()) { // if headers were not already sent, we can send an error response
resp = new Response(out); // ignore whatever headers may have already been set
resp.getHeaders().add("Connection", "close"); // about to close connection
resp.sendError(500, "Error processing request: " + t.getMessage());
LOGGER.log(System.Logger.Level.WARNING, "Error processing request", t);
} // otherwise just abort the connection since we can't recover
break; // proceed to close connection
} finally {
resp.close(); // close response and flush output
}
// consume any leftover body data so next request can be processed
transfer(req.getBody(), null, -1);
// RFC7230#6.6: persist connection unless client or server close explicitly (or legacy client)
} while (!"close".equalsIgnoreCase(req.getHeaders().get("Connection"))
&& !"close".equalsIgnoreCase(resp.getHeaders().get("Connection")) && req.getVersion().endsWith("1.1"));
}
/**
* Handles a single transaction on a connection.
* <p>
* Subclasses can override this method to perform filtering on the
* request or response, apply wrappers to them, or further customize
* the transaction processing in some other way.
*
* @param req the transaction request
* @param resp the transaction response (into which the response is written)
* @throws IOException if and error occurs
*/
protected void handleTransaction(Request req, Response resp) throws IOException {
resp.setClientCapabilities(req);
if (preprocessTransaction(req, resp))
handleMethod(req, resp);
}
/**
* Preprocesses a transaction, performing various validation checks
* and required special header handling, possibly returning an
* appropriate response.
*
* @param req the request
* @param resp the response
* @return whether further processing should be performed on the transaction
* @throws IOException if an error occurs
*/
protected boolean preprocessTransaction(Request req, Response resp) throws IOException {
Headers reqHeaders = req.getHeaders();
// validate request
String version = req.getVersion();
if (version.equals("HTTP/1.1")) {
if (!reqHeaders.contains("Host")) {
// RFC2616#14.23: missing Host header gets 400
resp.sendError(400, "Missing required Host header");
return false;
}
// return a continue response before reading body
String expect = reqHeaders.get("Expect");
if (expect != null) {
if (expect.equalsIgnoreCase("100-continue")) {
Response tempResp = new Response(resp.getOutputStream());
tempResp.sendHeaders(100);
resp.getOutputStream().flush();
} else {
// RFC2616#14.20: if unknown expect, send 417
resp.sendError(417);
return false;
}
}
} else if (version.equals("HTTP/1.0") || version.equals("HTTP/0.9")) {
// RFC2616#14.10 - remove connection headers from older versions
for (String token : splitElements(reqHeaders.get("Connection"), false))
reqHeaders.remove(token);
} else {
resp.sendError(400, "Unknown version: " + version);
return false;
}
return true;
}
/**
* Handles a transaction according to the request method.
*
* @param req the transaction request
* @param resp the transaction response (into which the response is written)
* @throws IOException if and error occurs
*/
protected void handleMethod(Request req, Response resp) throws IOException {
String method = req.getMethod();
Map<String, ContextHandler> handlers = req.getContext().getHandlers();
// RFC 2616#5.1.1 - GET and HEAD must be supported
if (method.equals("GET") || handlers.containsKey(method)) {
serve(req, resp); // method is handled by context handler (or 404)
} else if (method.equals("HEAD")) { // default HEAD handler
req.method = "GET"; // identical to a GET
resp.setDiscardBody(true); // process normally but discard body
serve(req, resp);
} else if (method.equals("TRACE")) { // default TRACE handler
handleTrace(req, resp);
} else {
Set<String> methods = new LinkedHashSet<>(Arrays.asList("GET", "HEAD", "TRACE", "OPTIONS")); // built-in methods
// "*" is a special server-wide (no-context) request supported by OPTIONS
boolean isServerOptions = req.getPath().equals("*") && method.equals("OPTIONS");
methods.addAll(isServerOptions ? req.getVirtualHost().getMethods() : handlers.keySet());
resp.getHeaders().add("Allow", join(", ", methods));
if (method.equals("OPTIONS")) { // default OPTIONS handler
resp.getHeaders().add("Content-Length", "0"); // RFC2616#9.2
resp.sendHeaders(200);
} else if (req.getVirtualHost().getMethods().contains(method)) {
resp.sendHeaders(405); // supported by server, but not this context (nor built-in)
} else {
resp.sendError(501); // unsupported method
}
}
}
/**
* Handles a TRACE method request.
*
* @param req the request
* @param resp the response into which the content is written
* @throws IOException if an error occurs
*/
public void handleTrace(Request req, Response resp) throws IOException {
resp.sendHeaders(200, -1, -1, null, "message/http", null);
OutputStream out = resp.getBody();
out.write(getBytes("TRACE ", req.getURI().toString(), " ", req.getVersion()));
out.write(CRLF);
req.getHeaders().writeTo(out);
transfer(req.getBody(), out, -1);
}
/**
* Serves the content for a request by invoking the context
* handler for the requested context (path) and HTTP method.
*
* @param req the request
* @param resp the response into which the content is written
* @throws IOException if an error occurs
*/
protected void serve(Request req, Response resp) throws IOException {
// get context handler to handle request
ContextHandler handler = req.getContext().getHandlers().get(req.getMethod());
if (handler == null) {
resp.sendError(404);
return;
}
// serve request
int status = 404;
// add directory index if necessary
String path = req.getPath();
if (path.endsWith("/")) {
String index = req.getVirtualHost().getDirectoryIndex();
if (index != null) {
req.setPath(path + index);
status = handler.serve(req, resp);
req.setPath(path);
}
}
if (status == 404)
status = handler.serve(req, resp);
if (status > 0)
resp.sendError(status);
}
/**
* Adds a Content-Type mapping for the given path suffixes.
* If any of the path suffixes had a previous Content-Type associated
* with it, it is replaced with the given one. Path suffixes are
* considered case-insensitive, and contentType is converted to lowercase.
*
* @param contentType the content type (MIME type) to be associated with
* the given path suffixes
* @param suffixes the path suffixes which will be associated with
* the contentType, e.g. the file extensions of served files
* (excluding the '.' character)
*/
public static void addContentType(String contentType, String... suffixes) {
for (String suffix : suffixes)
contentTypes.put(suffix.toLowerCase(Locale.US), contentType.toLowerCase(Locale.US));
}
/**
* Adds Content-Type mappings from a standard mime.types file.
*
* @param in a stream containing a mime.types file
* @throws IOException if an error occurs
* @throws FileNotFoundException if the file is not found or cannot be read
*/
public static void addContentTypes(InputStream in) throws IOException {
try (in) {
while (true) {
String line = readLine(in).trim(); // throws EOFException when done
if (line.length() > 0 && line.charAt(0) != '#') {
String[] tokens = split(line, " \t", -1);
for (int i = 1; i < tokens.length; i++)
addContentType(tokens[0], tokens[i]);
}
}
} catch (EOFException ignore) { // the end of file was reached - it's ok
}
}
/**
* Returns the content type for the given path, according to its suffix,
* or the given default content type if none can be determined.
*
* @param path the path whose content type is requested
* @param def a default content type which is returned if none can be
* determined
* @return the content type for the given path, or the given default
*/
public static String getContentType(String path, String def) {
int dot = path.lastIndexOf('.');
String type = dot < 0 ? def : contentTypes.get(path.substring(dot + 1).toLowerCase(Locale.US));
return type != null ? type : def;
}
/**
* Checks whether data of the given content type (MIME type) is compressible.
*
* @param contentType the content type
* @return true if the data is compressible, false if not
*/
public static boolean isCompressible(String contentType) {
int pos = contentType.indexOf(';'); // exclude params
String ct = pos < 0 ? contentType : contentType.substring(0, pos);
for (String s : compressibleContentTypes)
if (s.equals(ct) || s.charAt(0) == '*' && ct.endsWith(s.substring(1))
|| s.charAt(s.length() - 1) == '*' && ct.startsWith(s.substring(0, s.length() - 1)))
return true;
return false;
}
/**
* Returns the local host's auto-detected name.
*
* @return the local host name
*/
public static String detectLocalHostName() {
try {
return InetAddress.getLocalHost().getCanonicalHostName();
} catch (UnknownHostException uhe) {
return "localhost";
}
}
/**
* Parses name-value pair parameters from the given "x-www-form-urlencoded"
* MIME-type string. This is the encoding used both for parameters passed
* as the query of an HTTP GET method, and as the content of HTML forms
* submitted using the HTTP POST method (as long as they use the default
* "application/x-www-form-urlencoded" encoding in their ENCTYPE attribute).
* UTF-8 encoding is assumed.
* <p>
* The parameters are returned as a list of string arrays, each containing
* the parameter name as the first element and its corresponding value
* as the second element (or an empty string if there is no value).
* <p>
* The list retains the original order of the parameters.
*
* @param s an "application/x-www-form-urlencoded" string
* @return the parameter name-value pairs parsed from the given string,
* or an empty list if there are none
*/
public static List<String[]> parseParamsList(String s) {
if (s == null || s.length() == 0)
return Collections.emptyList();
List<String[]> params = new ArrayList<>(8);
for (String pair : split(s, "&", -1)) {
int pos = pair.indexOf('=');
String name = pos < 0 ? pair : pair.substring(0, pos);
String val = pos < 0 ? "" : pair.substring(pos + 1);
name = URLDecoder.decode(name.trim(), StandardCharsets.UTF_8);
val = URLDecoder.decode(val.trim(), StandardCharsets.UTF_8);
if (name.length() > 0)
params.add(new String[] { name, val });
}
return params;
}
/**
* Converts a collection of pairs of objects (arrays of size two,
* each representing a key and corresponding value) into a Map.
* Duplicate keys are ignored (only the first occurrence of each key is considered).
* The map retains the original collection's iteration order.
*
* @param pairs a collection of arrays, each containing a key and corresponding value
* @param <K> the key type
* @param <V> the value type
* @return a map containing the paired keys and values, or an empty map
*/
@SuppressWarnings("unchecked")
public static <K, V> Map<K, V> toMap(Collection<? extends Object[]> pairs) {
if (pairs == null || pairs.isEmpty())
return Collections.emptyMap();
Map<K, V> map = new LinkedHashMap<>(pairs.size());
for (Object[] pair : pairs)
if (!map.containsKey(pair[0]))
map.put((K)pair[0], (V)pair[1]);
return map;
}
/**
* Returns the absolute (zero-based) content range value specified
* by the given range string. If multiple ranges are requested, a single
* range containing all of them is returned.
*
* @param range the string containing the range description
* @param length the full length of the requested resource
* @return the requested range, or null if the range value is invalid
*/
public static long[] parseRange(String range, long length) {
long min = Long.MAX_VALUE;
long max = Long.MIN_VALUE;
try {
for (String token : splitElements(range, false)) {
long start, end;
int dash = token.indexOf('-');
if (dash == 0) { // suffix range
start = length - parseULong(token.substring(1), 10);
end = length - 1;
} else if (dash == token.length() - 1) { // open range
start = parseULong(token.substring(0, dash), 10);
end = length - 1;
} else { // explicit range
start = parseULong(token.substring(0, dash), 10);
end = parseULong(token.substring(dash + 1), 10);
}
if (end < start)
throw new RuntimeException();
if (start < min)
min = start;
if (end > max)
max = end;
}
if (max < 0) // no tokens
throw new RuntimeException();
if (max >= length && min < length)
max = length - 1;
return new long[] { min, max }; // start might be >= length!
} catch (RuntimeException re) { // NFE, IOOBE or explicit RE
return null; // RFC2616#14.35.1 - ignore header if invalid
}
}
/**
* Parses an unsigned long value. This method behaves the same as calling
* {@link Long#parseLong(String, int)}, but considers the string invalid
* if it starts with an ASCII minus sign ('-') or plus sign ('+').
*
* @param s the String containing the long representation to be parsed
* @param radix the radix to be used while parsing s
* @return the long represented by s in the specified radix
* @throws NumberFormatException if the string does not contain a parsable
* long, or if it starts with an ASCII minus sign or plus sign
*/
public static long parseULong(String s, int radix) throws NumberFormatException {
long val = Long.parseLong(s, radix); // throws NumberFormatException
if (s.charAt(0) == '-' || s.charAt(0) == '+')
throw new NumberFormatException("invalid digit: " + s.charAt(0));
return val;
}
/**
* Parses a date string in one of the supported {@link #DATE_PATTERNS}.
* <p>
* Received date header values must be in one of the following formats:
* Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
* Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
* Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format
*
* @param time a string representation of a time value
* @return the parsed date value
* @throws IllegalArgumentException if the given string does not contain
* a valid date format in any of the supported formats
*/
public static Date parseDate(String time) {
for (String pattern : DATE_PATTERNS) {
try {
SimpleDateFormat df = new SimpleDateFormat(pattern, Locale.US);
df.setLenient(false);
df.setTimeZone(GMT);
return df.parse(time);
} catch (ParseException ignore) {}
}
throw new IllegalArgumentException("invalid date format: " + time);
}
/**
* Formats the given time value as a string in RFC 1123 format.
*
* @param time the time in milliseconds since January 1, 1970, 00:00:00 GMT
* @return the given time value as a string in RFC 1123 format
*/
public static String formatDate(long time) {
// this implementation performs far better than SimpleDateFormat instances, and even
// quite better than ThreadLocal SDFs - the server's CPU-bound benchmark gains over 20%!
if (time < -62167392000000L || time > 253402300799999L)
throw new IllegalArgumentException("year out of range (0001-9999): " + time);
char[] s = "DAY, 00 MON 0000 00:00:00 GMT".toCharArray(); // copy the format template
Calendar cal = new GregorianCalendar(GMT, Locale.US);
cal.setTimeInMillis(time);
System.arraycopy(DAYS, 4 * (cal.get(Calendar.DAY_OF_WEEK) - 1), s, 0, 3);
System.arraycopy(MONTHS, 4 * cal.get(Calendar.MONTH), s, 8, 3);
int n = cal.get(Calendar.DATE); s[5] += n / 10; s[6] += n % 10;
n = cal.get(Calendar.YEAR); s[12] += n / 1000; s[13] += n / 100 % 10;
s[14] += n / 10 % 10; s[15] += n % 10;
n = cal.get(Calendar.HOUR_OF_DAY); s[17] += n / 10; s[18] += n % 10;
n = cal.get(Calendar.MINUTE); s[20] += n / 10; s[21] += n % 10;
n = cal.get(Calendar.SECOND); s[23] += n / 10; s[24] += n % 10;
return new String(s);
}
/**
* Splits the given element list string (comma-separated header value)
* into its constituent non-empty trimmed elements.
* (RFC2616#2.1: element lists are delimited by a comma and optional LWS,
* and empty elements are ignored).
*
* @param list the element list string
* @param lower specifies whether the list elements should be lower-cased
* @return the non-empty elements in the list, or an empty array
*/
public static String[] splitElements(String list, boolean lower) {
return split(lower && list != null ? list.toLowerCase(Locale.US) : list, ",", -1);
}
/**
* Splits the given string into its constituent non-empty trimmed elements,
* which are delimited by any of the given delimiter characters.
* This is a more direct and efficient implementation than using a regex
* (e.g. String.split()), trimming the elements and removing empty ones.
*
* @param str the string to split
* @param delimiters the characters used as the delimiters between elements
* @param limit if positive, limits the returned array size (remaining of str in last element)
* @return the non-empty elements in the string, or an empty array
*/
public static String[] split(String str, String delimiters, int limit) {
if (str == null)
return new String[0];
Collection<String> elements = new ArrayList<>();
int len = str.length();
int start = 0;
int end;
while (start < len) {
for (end = --limit == 0 ? len : start;
end < len && delimiters.indexOf(str.charAt(end)) < 0; end++);
String element = str.substring(start, end).trim();
if (element.length() > 0)
elements.add(element);
start = end + 1;
}
return elements.toArray(String[]::new);
}
/**
* Returns a string constructed by joining the string representations of the
* iterated objects (in order), with the delimiter inserted between them.
*
* @param delim the delimiter that is inserted between the joined strings
* @param items the items whose string representations are joined
* @param <T> the item type
* @return the joined string
*/
public static <T> String join(String delim, Iterable<T> items) {
StringBuilder sb = new StringBuilder();
for (Iterator<T> it = items.iterator(); it.hasNext(); )
sb.append(it.next()).append(it.hasNext() ? delim : "");
return sb.toString();
}
/**
* Returns the parent of the given path.
*
* @param path the path whose parent is returned (must start with '/')
* @return the parent of the given path (excluding trailing slash),
* or null if given path is the root path
*/
public static String getParentPath(String path) {
path = trimRight(path, '/'); // remove trailing slash
int slash = path.lastIndexOf('/');
return slash < 0 ? null : path.substring(0, slash);
}
/**
* Returns the given string with all occurrences of the given character
* removed from its right side.
*
* @param s the string to trim
* @param c the character to remove
* @return the trimmed string
*/
public static String trimRight(String s, char c) {
int len = s.length() - 1;
int end;
for (end = len; end >= 0 && s.charAt(end) == c; end--);
return end == len ? s : s.substring(0, end + 1);
}
/**
* Returns the given string with all occurrences of the given character
* removed from its left side.
*
* @param s the string to trim
* @param c the character to remove
* @return the trimmed string
*/
public static String trimLeft(String s, char c) {
int len = s.length();
int start;
for (start = 0; start < len && s.charAt(start) == c; start++);
return start == 0 ? s : s.substring(start);
}
/**
* Trims duplicate consecutive occurrences of the given character within the
* given string, replacing them with a single instance of the character.
*
* @param s the string to trim
* @param c the character to trim
* @return the given string with duplicate consecutive occurrences of c
* replaced by a single instance of c
*/
public static String trimDuplicates(String s, char c) {
int start = 0;
while ((start = s.indexOf(c, start) + 1) > 0) {
int end;
for (end = start; end < s.length() && s.charAt(end) == c; end++);
if (end > start)
s = s.substring(0, start) + s.substring(end);
}
return s;
}
/**
* Returns a human-friendly string approximating the given data size,
* e.g. "316", "1.8K", "324M", etc.
*
* @param size the size to display
* @return a human-friendly string approximating the given data size
*/
public static String toSizeApproxString(long size) {
final char[] units = { ' ', 'K', 'M', 'G', 'T', 'P', 'E' };
int u;
double s;
for (u = 0, s = size; s >= 1000; u++, s /= 1024);
return String.format(s < 10 ? "%.1f%c" : "%.0f%c", s, units[u]);
}
/**
* Returns an HTML-escaped version of the given string for safe display
* within a web page. The characters '&amp;', '&gt;' and '&lt;' must always
* be escaped, and single and double quotes must be escaped within
* attribute values; this method escapes them always. This method can
* be used for generating both HTML and XHTML valid content.
*
* @param s the string to escape
* @return the escaped string
* @see <a href="http://www.w3.org/International/questions/qa-escapes">The W3C FAQ</a>
*/
public static String escapeHTML(String s) {
int len = s.length();
StringBuilder sb = new StringBuilder(len + 30);
int start = 0;
for (int i = 0; i < len; i++) {
String ref = switch (s.charAt(i)) {
case '&' -> "&amp;";
case '>' -> "&gt;";
case '<' -> "&lt;";
case '"' -> "&quot;";
case '\'' -> "&#39;";
default -> null;
};
if (ref != null) {
sb.append(s, start, i).append(ref);
start = i + 1;
}
}
return start == 0 ? s : sb.append(s.substring(start)).toString();
}
/**
* Converts strings to bytes by casting the chars to bytes.
* This is a fast way to encode a string as ISO-8859-1/US-ASCII bytes.
* If multiple strings are provided, their bytes are concatenated.
*
* @param strings the strings to convert (containing only ISO-8859-1 chars)
* @return the byte array
*/
public static byte[] getBytes(String... strings) {
int n = 0;
for (String s : strings)
n += s.length();
byte[] b = new byte[n];
n = 0;
for (String s : strings)
for (int i = 0, len = s.length(); i < len; i++)
b[n++] = (byte)s.charAt(i);
return b;
}
/**
* Transfers data from an input stream to an output stream.
*
* @param in the input stream to transfer from
* @param out the output stream to transfer to (or null to discard output)
* @param len the number of bytes to transfer. If negative, the entire
* contents of the input stream are transferred.
* @throws IOException if an IO error occurs or the input stream ends
* before the requested number of bytes have been read
*/
public static void transfer(InputStream in, OutputStream out, long len) throws IOException {
if (len == 0 || out == null && len < 0 && in.read() < 0)
return; // small optimization - avoid buffer creation
byte[] buf = new byte[4096];
while (len != 0) {
int count = len < 0 || buf.length < len ? buf.length : (int)len;
count = in.read(buf, 0, count);
if (count < 0) {
if (len > 0)
throw new IOException("unexpected end of stream");
break;
}
if (out != null)
out.write(buf, 0, count);
len -= len > 0 ? count : 0;
}
}
/**
* Reads the token starting at the current stream position and ending at
* the first occurrence of the given delimiter byte, in the given encoding.
* If LF is specified as the delimiter, a CRLF pair is also treated as one.
*
* @param in the stream from which the token is read
* @param delim the byte value which marks the end of the token,
* or -1 if the token ends at the end of the stream
* @param enc a character-encoding name
* @param maxLength the maximum length (in bytes) to read
* @return the read token, excluding the delimiter
* @throws UnsupportedEncodingException if the encoding is not supported
* @throws EOFException if the stream end is reached before a delimiter is found
* @throws IOException if an IO error occurs, or the maximum length
* is reached before the token end is reached
*/
public static String readToken(InputStream in, int delim,
String enc, int maxLength) throws IOException {
// note: we avoid using a ByteArrayOutputStream here because it
// suffers the overhead of synchronization for each byte written
int b;
int len = 0; // buffer length
int count = 0; // number of read bytes
byte[] buf = null; // optimization - lazy allocation only if necessary
while ((b = in.read()) != -1 && b != delim) {
if (count == len) { // expand buffer
if (count == maxLength)
throw new IOException("token too large (" + count + ")");
len = len > 0 ? 2 * len : 256; // start small, double each expansion
len = Math.min(maxLength, len);
byte[] expanded = new byte[len];
if (buf != null)
System.arraycopy(buf, 0, expanded, 0, count);
buf = expanded;
}
buf[count++] = (byte)b;
}
if (b < 0 && delim != -1)
throw new EOFException("unexpected end of stream");
if (delim == '\n' && count > 0 && buf[count - 1] == '\r')
count--;
return count > 0 ? new String(buf, 0, count, enc) : "";
}
/**
* Reads the ISO-8859-1 encoded string starting at the current stream
* position and ending at the first occurrence of the LF character.
*
* @param in the stream from which the line is read
* @return the read string, excluding the terminating LF character
* and (if exists) the CR character immediately preceding it
* @throws EOFException if the stream end is reached before an LF character is found
* @throws IOException if an IO error occurs, or the line is longer than 8192 bytes
* @see #readToken(InputStream, int, String, int)
*/
public static String readLine(InputStream in) throws IOException {
return readToken(in, '\n', "ISO8859_1", 8192);
}
/**
* Reads headers from the given stream. Headers are read according to the
* RFC, including folded headers, element lists, and multiple headers
* (which are concatenated into a single element list header).
* Leading and trailing whitespace is removed.
*
* @param in the stream from which the headers are read
* @return the read headers (possibly empty, if none exist)
* @throws IOException if an IO error occurs or the headers are malformed
* or there are more than 100 header lines
*/
public static Headers readHeaders(InputStream in) throws IOException {
Headers headers = new Headers();
String line;
String prevLine = "";
int count = 0;
while ((line = readLine(in)).length() > 0) {
int start; // start of line data (after whitespace)
for (start = 0; start < line.length() &&
Character.isWhitespace(line.charAt(start)); start++);
if (start > 0) // unfold header continuation line
line = prevLine + ' ' + line.substring(start);
int separator = line.indexOf(':');
if (separator < 0)
throw new IOException("invalid header: \"" + line + "\"");
String name = line.substring(0, separator);
String value = line.substring(separator + 1).trim(); // ignore LWS
Header replaced = headers.replace(name, value);
// concatenate repeated headers (distinguishing repeated from folded)
if (replaced != null && start == 0) {
value = replaced.getValue() + ", " + value;
line = name + ": " + value;
headers.replace(name, value);
}
prevLine = line;
if (++count > 100)
throw new IOException("too many header lines");
}
return headers;
}
/**
* Matches the given ETag value against the given ETags. A match is found
* if the given ETag is not null, and either the ETags contain a "*" value,
* or one of them is identical to the given ETag. If strong comparison is
* used, tags beginning with the weak ETag prefix "W/" never match.
* See RFC2616#3.11, RFC2616#13.3.3.
*
* @param strong if true, strong comparison is used, otherwise weak
* comparison is used
* @param etags the ETags to match against
* @param etag the ETag to match
* @return true if the ETag is matched, false otherwise
*/
public static boolean match(boolean strong, String[] etags, String etag) {
if (etag == null || strong && etag.startsWith("W/"))
return false;
for (String e : etags)
if (e.equals("*") || (e.equals(etag) && !(strong && (e.startsWith("W/")))))
return true;
return false;
}
/**
* Calculates the appropriate response status for the given request and
* its resource's last-modified time and ETag, based on the conditional
* headers present in the request.
*
* @param req the request
* @param lastModified the resource's last modified time
* @param etag the resource's ETag
* @return the appropriate response status for the request
*/
public static int getConditionalStatus(Request req, long lastModified, String etag) {
Headers headers = req.getHeaders();
// If-Match
String header = headers.get("If-Match");
if (header != null && !match(true, splitElements(header, false), etag))
return 412;
// If-Unmodified-Since
Date date = headers.getDate("If-Unmodified-Since");
if (date != null && lastModified > date.getTime())
return 412;
// If-Modified-Since
int status = 200;
boolean force = false;
date = headers.getDate("If-Modified-Since");
if (date != null && date.getTime() <= System.currentTimeMillis()) {
if (lastModified > date.getTime())
force = true;
else
status = 304;
}
// If-None-Match
header = headers.get("If-None-Match");
if (header != null) {
if (match(false, splitElements(header, false), etag)) // RFC7232#3.2: use weak matching
status = req.getMethod().equals("GET")
|| req.getMethod().equals("HEAD") ? 304 : 412;
else
force = true;
}
return force ? 200 : status;
}
/**
* Serves a context's contents from a file based resource.
* <p>
* The file is located by stripping the given context prefix from
* the request's path, and appending the result to the given base directory.
* <p>
* Missing, forbidden and otherwise invalid files return the appropriate
* error response. Directories are served as an HTML index page if the
* virtual host allows one, or a forbidden error otherwise. Files are
* sent with their corresponding content types, and handle conditional
* and partial retrievals according to the RFC.
*
* @param base the base directory to which the context is mapped
* @param context the context which is mapped to the base directory
* @param req the request
* @param resp the response into which the content is written
* @return the HTTP status code to return, or 0 if a response was sent
* @throws IOException if an error occurs
*/
public static int serveFile(File base, String context,
Request req, Response resp) throws IOException {
String relativePath = req.getPath().substring(context.length());
File file = new File(base, relativePath).getCanonicalFile();
if (!file.exists() || file.isHidden() || file.getName().startsWith(".")) {
return 404;
} else if (!file.canRead() || !file.getPath().startsWith(base.getPath())) { // validate
return 403;
} else if (file.isDirectory()) {
if (relativePath.endsWith("/")) {
if (!req.getVirtualHost().isAllowGeneratedIndex())
return 403;
resp.send(200, createIndex(file, req.getPath()));
} else { // redirect to the normalized directory URL ending with '/'
resp.redirect(req.getBaseURL() + req.getPath() + "/", true);
}
} else if (relativePath.endsWith("/")) {
return 404; // non-directory ending with slash (File constructor removed it)
} else {
serveFileContent(file, req, resp);
}
return 0;
}
/**
* Serves the contents of a file, with its corresponding content type,
* last modification time, etc. conditional and partial retrievals are
* handled according to the RFC.
*
* @param file the existing and readable file whose contents are served
* @param req the request
* @param resp the response into which the content is written
* @throws IOException if an error occurs
*/
public static void serveFileContent(File file, Request req, Response resp) throws IOException {
long len = file.length();
long lastModified = file.lastModified();
String etag = "W/\"" + lastModified + "\""; // a weak tag based on date
int status = 200;
// handle range or conditional request
long[] range = req.getRange(len);
if (range == null || len == 0) {
status = getConditionalStatus(req, lastModified, etag);
} else {
String ifRange = req.getHeaders().get("If-Range");
if (ifRange == null) {
if (range[0] >= len)
status = 416; // unsatisfiable range
else
status = getConditionalStatus(req, lastModified, etag);
} else if (range[0] >= len) {
// RFC2616#14.16, 10.4.17: invalid If-Range gets everything
range = null;
} else { // send either range or everything
if (!ifRange.startsWith("\"") && !ifRange.startsWith("W/")) {
Date date = req.getHeaders().getDate("If-Range");
if (date != null && lastModified > date.getTime())
range = null; // modified - send everything
} else if (!ifRange.equals(etag)) {
range = null; // modified - send everything
}
}
}
// send the response
Headers respHeaders = resp.getHeaders();
switch (status) {
case 304 -> { // no other headers or body allowed
respHeaders.add("ETag", etag);
respHeaders.add("Vary", "Accept-Encoding");
respHeaders.add("Last-Modified", formatDate(lastModified));
resp.sendHeaders(304);
}
case 412 -> resp.sendHeaders(412);
case 416 -> {
respHeaders.add("Content-Range", "bytes */" + len);
resp.sendHeaders(416);
}
case 200 -> {
// send OK response
resp.sendHeaders(200, len, lastModified, etag,
getContentType(file.getName(), "application/octet-stream"), range);
// send body
try (InputStream in = new FileInputStream(file)) {
resp.sendBody(in, len, range);
}
}
default -> resp.sendHeaders(500); // should never happen
}
}
/**
* Serves the contents of a directory as an HTML file index.
*
* @param dir the existing and readable directory whose contents are served
* @param path the displayed base path corresponding to dir
* @return an HTML string containing the file index for the directory
*/
public static String createIndex(File dir, String path) {
if (!path.endsWith("/"))
path += "/";
// calculate name column width
int w = 21; // minimum width
for (String name : dir.list())
if (name.length() > w)
w = name.length();
w += 2; // with room for added slash and space
// note: we use apache's format, for consistent user experience
Formatter f = new Formatter(Locale.US);
f.format("<!DOCTYPE html>%n" +
"<html><head><title>Index of %s</title></head>%n" +
"<body><h1>Index of %s</h1>%n" +
"<pre> Name%" + (w - 5) + "s Last modified Size<hr>",
path, path, "");
if (path.length() > 1) // add parent link if not root path
f.format(" <a href=\"%s/\">Parent Directory</a>%"
+ (w + 5) + "s-%n", getParentPath(path), "");
for (File file : dir.listFiles()) {
try {
String name = file.getName() + (file.isDirectory() ? "/" : "");
String size = file.isDirectory() ? "- " : toSizeApproxString(file.length());
// properly url-encode the link
String link = new URI(null, path + name, null).toASCIIString();
if (!file.isHidden() && !name.startsWith("."))
f.format(" <a href=\"%s\">%s</a>%-" + (w - name.length()) +
"s&#8206;%td-%<tb-%<tY %<tR%6s%n",
link, name, "", file.lastModified(), size);
} catch (URISyntaxException ignore) {}
}
f.format("</pre></body></html>");
return f.toString();
}
/**
* Starts a stand-alone HTTP server, serving files from disk.
*
* @param args the command line arguments
*/
public static void main(String[] args) {
try {
if (args.length == 0) {
System.err.printf("Usage: java [-options] %s <directory> [port]%n" +
"To enable SSL: specify options -Djavax.net.ssl.keyStore, " +
"-Djavax.net.ssl.keyStorePassword, etc.%n", JLHTTPServer.class.getName());
return;
}
File dir = new File(args[0]);
if (!dir.canRead())
throw new FileNotFoundException(dir.getAbsolutePath());
int port = args.length < 2 ? 80 : (int)parseULong(args[1], 10);
// set up server
for (File f : Arrays.asList(new File("/etc/mime.types"), new File(dir, ".mime.types")))
if (f.exists())
addContentTypes(new FileInputStream(f));
JLHTTPServer server = new JLHTTPServer(port);
if (System.getProperty("javax.net.ssl.keyStore") != null) // enable SSL if configured
server.setServerSocketFactory(SSLServerSocketFactory.getDefault());
VirtualHost host = server.getVirtualHost(null); // default host
host.setAllowGeneratedIndex(true); // with directory index pages
host.addContext("/", new FileContextHandler(dir));
host.addContext("/api/time", (req, resp) -> {
long now = System.currentTimeMillis();
resp.getHeaders().add("Content-Type", "text/plain");
resp.send(200, String.format("%tF %<tT", now));
return 0;
});
server.start();
System.out.println("HTTPServer is listening on port " + port);
} catch (Exception e) {
System.err.println("error: " + e);
}
}
}
Reference in New Issue View Git Blame Copy Permalink