/*
 * Copyright 2010 Google Inc.
 *
 * Licensed under the Apache License, Version 2.0 (the "License"); you may not
 * use this file except in compliance with the License. You may obtain a copy of
 * the License at
 *
 * http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
 * License for the specific language governing permissions and limitations under
 * the License.
 */
package com.google.gwt.safehtml.shared;

import com.google.gwt.core.shared.GWT;
import com.google.gwt.http.client.URL;
import com.google.gwt.regexp.shared.RegExp;
import com.google.gwt.safehtml.shared.annotations.IsSafeUri;
import com.google.gwt.safehtml.shared.annotations.SuppressIsSafeUriCastCheck;

import java.io.UnsupportedEncodingException;
import java.util.Locale;

/**
 * Utility class containing static methods for validating and sanitizing URIs.
 */
public final class UriUtils {

  /**
   * Characters that don't need %-escaping (minus letters and digits), according
   * to ECMAScript 5th edition for the {@code encodeURI} function.
   */
  static final String DONT_NEED_ENCODING = ";/?:@&=+$," // uriReserved
      + "-_.!~*'()" // uriMark
      + "#"
      + "[]"; // could be used in IPv6 addresses

  // used in conditional code in encode()
  private static final RegExp ESCAPED_LBRACKET_RE =
    GWT.isScript() ? RegExp.compile("%5B", "g") : null;
  private static final RegExp ESCAPED_RBRACKET_RE =
    GWT.isScript() ? RegExp.compile("%5D", "g") : null;

  /**
   * Encodes the URL.
   * <p>
   * In client code, this method delegates to {@link URL#encode(String)} and
   * then unescapes brackets, as they might be used for IPv6 addresses.
   *
   * @param uri the URL to encode
   * @return the %-escaped URL
   */
  public static String encode(String uri) {
    if (GWT.isScript()) {
      uri = URL.encode(uri);
      // Follow the same approach as SafeHtmlUtils.htmlEscape
      if (uri.indexOf("%5B") != -1) {
        uri = ESCAPED_LBRACKET_RE.replace(uri, "[");
      }
      if (uri.indexOf("%5D") != -1) {
        uri = ESCAPED_RBRACKET_RE.replace(uri, "]");
      }
      return uri;
    } else {
      StringBuilder sb = new StringBuilder();
      byte[] utf8bytes;
      try {
        utf8bytes = uri.getBytes("UTF-8");
      } catch (UnsupportedEncodingException e) {
        // UTF-8 is guaranteed to be implemented, this code won't ever run.
        return null;
      }
      for (byte b : utf8bytes) {
        int c = b & 0xFF;
        // This works because characters that don't need encoding are all
        // expressed as a single UTF-8 byte
        if (('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z') || ('0' <= c && c <= '9')
            || DONT_NEED_ENCODING.indexOf(c) != -1) {
          sb.append((char) c);
        } else {
          String hexByte = Integer.toHexString(c).toUpperCase(Locale.ROOT);
          if (hexByte.length() == 1) {
            hexByte = "0" + hexByte;
          }
          sb.append('%').append(hexByte);
        }
      }
      return sb.toString();
    }
  }

  /**
   * Encodes the URL, preserving existing %-escapes.
   *
   * @param uri the URL to encode
   * @return the %-escaped URL
   */
  public static String encodeAllowEscapes(String uri) {
    StringBuilder escaped = new StringBuilder();

    boolean firstSegment = true;
    for (String segment : uri.split("%", -1)) {
      if (firstSegment) {
        /*
         * The first segment is never part of a percent-escape, so we always
         * escape it. Note that if the input starts with a percent, we will get
         * an empty segment before that.
         */
        firstSegment = false;
        escaped.append(encode(segment));
        continue;
      }

      if (segment.length() >= 2 && segment.substring(0, 2).matches("[0-9a-fA-F]{2}")) {
        // Append the escape without encoding.
        escaped.append("%").append(segment.substring(0, 2));

        // Append the rest of the segment, escaped.
        escaped.append(encode(segment.substring(2)));
      } else {
        // The segment did not start with an escape, so encode the whole
        // segment.
        escaped.append("%25").append(encode(segment));
      }
    }
    return escaped.toString();
  }

  /**
   * Extracts the scheme of a URI.
   *
   * @param uri the URI to extract the scheme from
   * @return the URI's scheme, or {@code null} if the URI does not have one
   */
  public static String extractScheme(String uri) {
    int colonPos = uri.indexOf(':');
    if (colonPos < 0) {
      return null;
    }
    String scheme = uri.substring(0, colonPos);
    if (scheme.indexOf('/') >= 0 || scheme.indexOf('#') >= 0) {
      /*
       *  The URI's prefix up to the first ':' contains other URI special
       *  chars, and won't be interpreted as a scheme.
       *
       *  TODO(xtof): Consider basing this on URL#isValidProtocol or similar;
       *  however I'm worried that being too strict here will effectively
       *   allow dangerous schemes accepted in loosely parsing browsers.
       */
      return null;
    }
    return scheme;
  }

  /**
   * Returns a {@link SafeUri} constructed from a value that is fully under
   * the control of the program, e.g., a constant.
   *
   * <p>
   * The string is not sanitized and no checks are performed.  The assumption
   * that the resulting value adheres to the {@link SafeUri} type contract
   * is entirely based on the argument being fully under program control and
   * not being derived from a program input.
   *
   * <p>
   * <strong>Convention of use:</strong> This method must only be invoked on
   * values that are fully under the program's control, such as string literals.
   *
   * @param s the input String
   * @return a SafeUri instance
   */
  @SuppressIsSafeUriCastCheck
  public static SafeUri fromSafeConstant(String s) {
    SafeUriHostedModeUtils.maybeCheckValidUri(s);
    return new SafeUriString(s);
  }

  /**
   * Returns a {@link SafeUri} obtained by sanitizing the provided string.
   *
   * <p>
   * The input string is sanitized using {@link #sanitizeUri(String)}.
   *
   * @param s the input String
   * @return a SafeUri instance
   */
  public static SafeUri fromString(String s) {
    return new SafeUriString(sanitizeUri(s));
  }

  /**
   * Returns a {@link SafeUri} constructed from a trusted string, i.e., without
   * sanitizing the string. No checks are performed. The calling code should be
   * carefully reviewed to ensure the argument meets the SafeUri contract.
   *
   * @param s the input String
   * @return a SafeUri instance
   */
  @SuppressIsSafeUriCastCheck
  public static SafeUri fromTrustedString(String s) {
    SafeUriHostedModeUtils.maybeCheckValidUri(s);
    return new SafeUriString(s);
  }

  /**
   * Determines if a {@link String} is safe to use as the value of a URI-valued
   * HTML attribute such as {@code src} or {@code href}.
   *
   * <p>
   * In this context, a URI is safe if it can be established that using it as
   * the value of a URI-valued HTML attribute such as {@code src} or {@code
   * href} cannot result in script execution. Specifically, this method deems a
   * URI safe if it either does not have a scheme, or its scheme is one of
   * {@code http, https, ftp, mailto}.
   *
   * @param uri the URI to validate
   * @return {@code true} if {@code uri} is safe in the above sense; {@code
   *         false} otherwise
   */
  public static boolean isSafeUri(String uri) {
    String scheme = extractScheme(uri);
    if (scheme == null) {
      return true;
    }
    /*
     * Special care is be taken with case-insensitive 'i' in the Turkish locale.
     * i -> to upper in Turkish locale -> İ
     * I -> to lower in Turkish locale -> ı
     * For this reason there are two checks for mailto: "mailto" and "MAILTO"
     * For details, see: http://www.i18nguy.com/unicode/turkish-i18n.html
     */
    String schemeLc = scheme.toLowerCase(Locale.ROOT);
    return ("http".equals(schemeLc)
        || "https".equals(schemeLc)
        || "ftp".equals(schemeLc)
        || "mailto".equals(schemeLc)
        || "MAILTO".equals(scheme.toUpperCase(Locale.ROOT)));
  }

  /**
   * Sanitizes a URI.
   *
   * <p>
   * This method returns the URI provided if it is safe to use as the value
   * of a URI-valued HTML attribute according to {@link #isSafeUri}, or the URI
   * "{@code #}" otherwise.
   *
   * @param uri the URI to sanitize
   * @return a sanitized String
   */
  @IsSafeUri
  @SuppressIsSafeUriCastCheck
  public static String sanitizeUri(String uri) {
    if (isSafeUri(uri)) {
      return encodeAllowEscapes(uri);
    } else {
      return "#";
    }
  }

  /**
   * Returns a {@link SafeUri} constructed from an untrusted string but without
   * sanitizing it.
   *
   * <strong>Despite this method creating a SafeUri instance, no checks are
   * performed, so the returned SafeUri is absolutely NOT guaranteed to be
   * safe!</strong>
   *
   * @param s the input String
   * @return a SafeUri instance
   * @deprecated This method is intended only for use in APIs that use
   *             {@link SafeUri} to represent URIs, but for backwards
   *             compatibility have methods that accept URI parameters as plain
   *             strings.
   */
  @Deprecated
  @SuppressIsSafeUriCastCheck
  public static SafeUri unsafeCastFromUntrustedString(String s) {
    return new SafeUriString(s);
  }

  // prevent instantiation
  private UriUtils() {
  }
}