You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by gi...@apache.org on 2018/03/08 14:54:06 UTC

[12/25] hbase-site git commit: Published site at a03d09abd72789bbf9364d8a9b2c54d0e9351af9.

http://git-wip-us.apache.org/repos/asf/hbase-site/blob/e9a81b89/devapidocs/src-html/org/apache/hadoop/hbase/filter/RegexStringComparator.EngineType.html
----------------------------------------------------------------------
diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/filter/RegexStringComparator.EngineType.html b/devapidocs/src-html/org/apache/hadoop/hbase/filter/RegexStringComparator.EngineType.html
index ce6ff26..577439c 100644
--- a/devapidocs/src-html/org/apache/hadoop/hbase/filter/RegexStringComparator.EngineType.html
+++ b/devapidocs/src-html/org/apache/hadoop/hbase/filter/RegexStringComparator.EngineType.html
@@ -77,353 +77,354 @@
 <span class="sourceLineNo">069</span> * @see java.util.regex.Pattern<a name="line.69"></a>
 <span class="sourceLineNo">070</span> */<a name="line.70"></a>
 <span class="sourceLineNo">071</span>@InterfaceAudience.Public<a name="line.71"></a>
-<span class="sourceLineNo">072</span>public class RegexStringComparator extends ByteArrayComparable {<a name="line.72"></a>
-<span class="sourceLineNo">073</span><a name="line.73"></a>
-<span class="sourceLineNo">074</span>  private static final Logger LOG = LoggerFactory.getLogger(RegexStringComparator.class);<a name="line.74"></a>
-<span class="sourceLineNo">075</span><a name="line.75"></a>
-<span class="sourceLineNo">076</span>  private Engine engine;<a name="line.76"></a>
-<span class="sourceLineNo">077</span><a name="line.77"></a>
-<span class="sourceLineNo">078</span>  /** Engine implementation type (default=JAVA) */<a name="line.78"></a>
-<span class="sourceLineNo">079</span>  @InterfaceAudience.Public<a name="line.79"></a>
-<span class="sourceLineNo">080</span>  public enum EngineType {<a name="line.80"></a>
-<span class="sourceLineNo">081</span>    JAVA,<a name="line.81"></a>
-<span class="sourceLineNo">082</span>    JONI<a name="line.82"></a>
-<span class="sourceLineNo">083</span>  }<a name="line.83"></a>
-<span class="sourceLineNo">084</span><a name="line.84"></a>
-<span class="sourceLineNo">085</span>  /**<a name="line.85"></a>
-<span class="sourceLineNo">086</span>   * Constructor<a name="line.86"></a>
-<span class="sourceLineNo">087</span>   * Adds Pattern.DOTALL to the underlying Pattern<a name="line.87"></a>
-<span class="sourceLineNo">088</span>   * @param expr a valid regular expression<a name="line.88"></a>
-<span class="sourceLineNo">089</span>   */<a name="line.89"></a>
-<span class="sourceLineNo">090</span>  public RegexStringComparator(String expr) {<a name="line.90"></a>
-<span class="sourceLineNo">091</span>    this(expr, Pattern.DOTALL);<a name="line.91"></a>
-<span class="sourceLineNo">092</span>  }<a name="line.92"></a>
-<span class="sourceLineNo">093</span><a name="line.93"></a>
-<span class="sourceLineNo">094</span>  /**<a name="line.94"></a>
-<span class="sourceLineNo">095</span>   * Constructor<a name="line.95"></a>
-<span class="sourceLineNo">096</span>   * Adds Pattern.DOTALL to the underlying Pattern<a name="line.96"></a>
-<span class="sourceLineNo">097</span>   * @param expr a valid regular expression<a name="line.97"></a>
-<span class="sourceLineNo">098</span>   * @param engine engine implementation type<a name="line.98"></a>
-<span class="sourceLineNo">099</span>   */<a name="line.99"></a>
-<span class="sourceLineNo">100</span>  public RegexStringComparator(String expr, EngineType engine) {<a name="line.100"></a>
-<span class="sourceLineNo">101</span>    this(expr, Pattern.DOTALL, engine);<a name="line.101"></a>
-<span class="sourceLineNo">102</span>  }<a name="line.102"></a>
-<span class="sourceLineNo">103</span><a name="line.103"></a>
-<span class="sourceLineNo">104</span>  /**<a name="line.104"></a>
-<span class="sourceLineNo">105</span>   * Constructor<a name="line.105"></a>
-<span class="sourceLineNo">106</span>   * @param expr a valid regular expression<a name="line.106"></a>
-<span class="sourceLineNo">107</span>   * @param flags java.util.regex.Pattern flags<a name="line.107"></a>
-<span class="sourceLineNo">108</span>   */<a name="line.108"></a>
-<span class="sourceLineNo">109</span>  public RegexStringComparator(String expr, int flags) {<a name="line.109"></a>
-<span class="sourceLineNo">110</span>    this(expr, flags, EngineType.JAVA);<a name="line.110"></a>
-<span class="sourceLineNo">111</span>  }<a name="line.111"></a>
-<span class="sourceLineNo">112</span><a name="line.112"></a>
-<span class="sourceLineNo">113</span>  /**<a name="line.113"></a>
-<span class="sourceLineNo">114</span>   * Constructor<a name="line.114"></a>
-<span class="sourceLineNo">115</span>   * @param expr a valid regular expression<a name="line.115"></a>
-<span class="sourceLineNo">116</span>   * @param flags java.util.regex.Pattern flags<a name="line.116"></a>
-<span class="sourceLineNo">117</span>   * @param engine engine implementation type<a name="line.117"></a>
-<span class="sourceLineNo">118</span>   */<a name="line.118"></a>
-<span class="sourceLineNo">119</span>  public RegexStringComparator(String expr, int flags, EngineType engine) {<a name="line.119"></a>
-<span class="sourceLineNo">120</span>    super(Bytes.toBytes(expr));<a name="line.120"></a>
-<span class="sourceLineNo">121</span>    switch (engine) {<a name="line.121"></a>
-<span class="sourceLineNo">122</span>      case JAVA:<a name="line.122"></a>
-<span class="sourceLineNo">123</span>        this.engine = new JavaRegexEngine(expr, flags);<a name="line.123"></a>
-<span class="sourceLineNo">124</span>        break;<a name="line.124"></a>
-<span class="sourceLineNo">125</span>      case JONI:<a name="line.125"></a>
-<span class="sourceLineNo">126</span>        this.engine = new JoniRegexEngine(expr, flags);<a name="line.126"></a>
-<span class="sourceLineNo">127</span>        break;<a name="line.127"></a>
-<span class="sourceLineNo">128</span>    }<a name="line.128"></a>
-<span class="sourceLineNo">129</span>  }<a name="line.129"></a>
-<span class="sourceLineNo">130</span><a name="line.130"></a>
-<span class="sourceLineNo">131</span>  /**<a name="line.131"></a>
-<span class="sourceLineNo">132</span>   * Specifies the {@link Charset} to use to convert the row key to a String.<a name="line.132"></a>
-<span class="sourceLineNo">133</span>   * &lt;p&gt;<a name="line.133"></a>
-<span class="sourceLineNo">134</span>   * The row key needs to be converted to a String in order to be matched<a name="line.134"></a>
-<span class="sourceLineNo">135</span>   * against the regular expression.  This method controls which charset is<a name="line.135"></a>
-<span class="sourceLineNo">136</span>   * used to do this conversion.<a name="line.136"></a>
-<span class="sourceLineNo">137</span>   * &lt;p&gt;<a name="line.137"></a>
-<span class="sourceLineNo">138</span>   * If the row key is made of arbitrary bytes, the charset {@code ISO-8859-1}<a name="line.138"></a>
-<span class="sourceLineNo">139</span>   * is recommended.<a name="line.139"></a>
-<span class="sourceLineNo">140</span>   * @param charset The charset to use.<a name="line.140"></a>
-<span class="sourceLineNo">141</span>   */<a name="line.141"></a>
-<span class="sourceLineNo">142</span>  public void setCharset(final Charset charset) {<a name="line.142"></a>
-<span class="sourceLineNo">143</span>    engine.setCharset(charset.name());<a name="line.143"></a>
-<span class="sourceLineNo">144</span>  }<a name="line.144"></a>
-<span class="sourceLineNo">145</span><a name="line.145"></a>
-<span class="sourceLineNo">146</span>  @Override<a name="line.146"></a>
-<span class="sourceLineNo">147</span>  public int compareTo(byte[] value, int offset, int length) {<a name="line.147"></a>
-<span class="sourceLineNo">148</span>    return engine.compareTo(value, offset, length);<a name="line.148"></a>
-<span class="sourceLineNo">149</span>  }<a name="line.149"></a>
-<span class="sourceLineNo">150</span><a name="line.150"></a>
-<span class="sourceLineNo">151</span>  /**<a name="line.151"></a>
-<span class="sourceLineNo">152</span>   * @return The comparator serialized using pb<a name="line.152"></a>
-<span class="sourceLineNo">153</span>   */<a name="line.153"></a>
-<span class="sourceLineNo">154</span>  @Override<a name="line.154"></a>
-<span class="sourceLineNo">155</span>  public byte [] toByteArray() {<a name="line.155"></a>
-<span class="sourceLineNo">156</span>    return engine.toByteArray();<a name="line.156"></a>
-<span class="sourceLineNo">157</span>  }<a name="line.157"></a>
-<span class="sourceLineNo">158</span><a name="line.158"></a>
-<span class="sourceLineNo">159</span>  /**<a name="line.159"></a>
-<span class="sourceLineNo">160</span>   * @param pbBytes A pb serialized {@link RegexStringComparator} instance<a name="line.160"></a>
-<span class="sourceLineNo">161</span>   * @return An instance of {@link RegexStringComparator} made from &lt;code&gt;bytes&lt;/code&gt;<a name="line.161"></a>
-<span class="sourceLineNo">162</span>   * @throws DeserializationException<a name="line.162"></a>
-<span class="sourceLineNo">163</span>   * @see #toByteArray<a name="line.163"></a>
-<span class="sourceLineNo">164</span>   */<a name="line.164"></a>
-<span class="sourceLineNo">165</span>  public static RegexStringComparator parseFrom(final byte [] pbBytes)<a name="line.165"></a>
-<span class="sourceLineNo">166</span>  throws DeserializationException {<a name="line.166"></a>
-<span class="sourceLineNo">167</span>    ComparatorProtos.RegexStringComparator proto;<a name="line.167"></a>
-<span class="sourceLineNo">168</span>    try {<a name="line.168"></a>
-<span class="sourceLineNo">169</span>      proto = ComparatorProtos.RegexStringComparator.parseFrom(pbBytes);<a name="line.169"></a>
-<span class="sourceLineNo">170</span>    } catch (InvalidProtocolBufferException e) {<a name="line.170"></a>
-<span class="sourceLineNo">171</span>      throw new DeserializationException(e);<a name="line.171"></a>
-<span class="sourceLineNo">172</span>    }<a name="line.172"></a>
-<span class="sourceLineNo">173</span>    RegexStringComparator comparator;<a name="line.173"></a>
-<span class="sourceLineNo">174</span>    if (proto.hasEngine()) {<a name="line.174"></a>
-<span class="sourceLineNo">175</span>      EngineType engine = EngineType.valueOf(proto.getEngine());<a name="line.175"></a>
-<span class="sourceLineNo">176</span>      comparator = new RegexStringComparator(proto.getPattern(), proto.getPatternFlags(),<a name="line.176"></a>
-<span class="sourceLineNo">177</span>        engine);<a name="line.177"></a>
-<span class="sourceLineNo">178</span>    } else {<a name="line.178"></a>
-<span class="sourceLineNo">179</span>      comparator = new RegexStringComparator(proto.getPattern(), proto.getPatternFlags());<a name="line.179"></a>
-<span class="sourceLineNo">180</span>    }<a name="line.180"></a>
-<span class="sourceLineNo">181</span>    String charset = proto.getCharset();<a name="line.181"></a>
-<span class="sourceLineNo">182</span>    if (charset.length() &gt; 0) {<a name="line.182"></a>
-<span class="sourceLineNo">183</span>      try {<a name="line.183"></a>
-<span class="sourceLineNo">184</span>        comparator.getEngine().setCharset(charset);<a name="line.184"></a>
-<span class="sourceLineNo">185</span>      } catch (IllegalCharsetNameException e) {<a name="line.185"></a>
-<span class="sourceLineNo">186</span>        LOG.error("invalid charset", e);<a name="line.186"></a>
-<span class="sourceLineNo">187</span>      }<a name="line.187"></a>
-<span class="sourceLineNo">188</span>    }<a name="line.188"></a>
-<span class="sourceLineNo">189</span>    return comparator;<a name="line.189"></a>
-<span class="sourceLineNo">190</span>  }<a name="line.190"></a>
-<span class="sourceLineNo">191</span><a name="line.191"></a>
-<span class="sourceLineNo">192</span>  /**<a name="line.192"></a>
-<span class="sourceLineNo">193</span>   * @param other<a name="line.193"></a>
-<span class="sourceLineNo">194</span>   * @return true if and only if the fields of the comparator that are serialized<a name="line.194"></a>
-<span class="sourceLineNo">195</span>   * are equal to the corresponding fields in other.  Used for testing.<a name="line.195"></a>
-<span class="sourceLineNo">196</span>   */<a name="line.196"></a>
-<span class="sourceLineNo">197</span>  @Override<a name="line.197"></a>
-<span class="sourceLineNo">198</span>  boolean areSerializedFieldsEqual(ByteArrayComparable other) {<a name="line.198"></a>
-<span class="sourceLineNo">199</span>    if (other == this) return true;<a name="line.199"></a>
-<span class="sourceLineNo">200</span>    if (!(other instanceof RegexStringComparator)) return false;<a name="line.200"></a>
-<span class="sourceLineNo">201</span>    RegexStringComparator comparator = (RegexStringComparator)other;<a name="line.201"></a>
-<span class="sourceLineNo">202</span>    return super.areSerializedFieldsEqual(comparator)<a name="line.202"></a>
-<span class="sourceLineNo">203</span>      &amp;&amp; engine.getClass().isInstance(comparator.getEngine())<a name="line.203"></a>
-<span class="sourceLineNo">204</span>      &amp;&amp; engine.getPattern().equals(comparator.getEngine().getPattern())<a name="line.204"></a>
-<span class="sourceLineNo">205</span>      &amp;&amp; engine.getFlags() == comparator.getEngine().getFlags()<a name="line.205"></a>
-<span class="sourceLineNo">206</span>      &amp;&amp; engine.getCharset().equals(comparator.getEngine().getCharset());<a name="line.206"></a>
-<span class="sourceLineNo">207</span>  }<a name="line.207"></a>
-<span class="sourceLineNo">208</span><a name="line.208"></a>
-<span class="sourceLineNo">209</span>  Engine getEngine() {<a name="line.209"></a>
-<span class="sourceLineNo">210</span>    return engine;<a name="line.210"></a>
-<span class="sourceLineNo">211</span>  }<a name="line.211"></a>
-<span class="sourceLineNo">212</span><a name="line.212"></a>
-<span class="sourceLineNo">213</span>  /**<a name="line.213"></a>
-<span class="sourceLineNo">214</span>   * This is an internal interface for abstracting access to different regular<a name="line.214"></a>
-<span class="sourceLineNo">215</span>   * expression matching engines.<a name="line.215"></a>
-<span class="sourceLineNo">216</span>   */<a name="line.216"></a>
-<span class="sourceLineNo">217</span>  static interface Engine {<a name="line.217"></a>
-<span class="sourceLineNo">218</span>    /**<a name="line.218"></a>
-<span class="sourceLineNo">219</span>     * Returns the string representation of the configured regular expression<a name="line.219"></a>
-<span class="sourceLineNo">220</span>     * for matching<a name="line.220"></a>
-<span class="sourceLineNo">221</span>     */<a name="line.221"></a>
-<span class="sourceLineNo">222</span>    String getPattern();<a name="line.222"></a>
-<span class="sourceLineNo">223</span><a name="line.223"></a>
-<span class="sourceLineNo">224</span>    /**<a name="line.224"></a>
-<span class="sourceLineNo">225</span>     * Returns the set of configured match flags, a bit mask that may include<a name="line.225"></a>
-<span class="sourceLineNo">226</span>     * {@link Pattern} flags<a name="line.226"></a>
-<span class="sourceLineNo">227</span>     */<a name="line.227"></a>
-<span class="sourceLineNo">228</span>    int getFlags();<a name="line.228"></a>
-<span class="sourceLineNo">229</span><a name="line.229"></a>
-<span class="sourceLineNo">230</span>    /**<a name="line.230"></a>
-<span class="sourceLineNo">231</span>     * Returns the name of the configured charset<a name="line.231"></a>
-<span class="sourceLineNo">232</span>     */<a name="line.232"></a>
-<span class="sourceLineNo">233</span>    String getCharset();<a name="line.233"></a>
-<span class="sourceLineNo">234</span><a name="line.234"></a>
-<span class="sourceLineNo">235</span>    /**<a name="line.235"></a>
-<span class="sourceLineNo">236</span>     * Set the charset used when matching<a name="line.236"></a>
-<span class="sourceLineNo">237</span>     * @param charset the name of the desired charset for matching<a name="line.237"></a>
-<span class="sourceLineNo">238</span>     */<a name="line.238"></a>
-<span class="sourceLineNo">239</span>    void setCharset(final String charset);<a name="line.239"></a>
-<span class="sourceLineNo">240</span><a name="line.240"></a>
-<span class="sourceLineNo">241</span>    /**<a name="line.241"></a>
-<span class="sourceLineNo">242</span>     * Return the serialized form of the configured matcher<a name="line.242"></a>
-<span class="sourceLineNo">243</span>     */<a name="line.243"></a>
-<span class="sourceLineNo">244</span>    byte [] toByteArray();<a name="line.244"></a>
-<span class="sourceLineNo">245</span><a name="line.245"></a>
-<span class="sourceLineNo">246</span>    /**<a name="line.246"></a>
-<span class="sourceLineNo">247</span>     * Match the given input against the configured pattern<a name="line.247"></a>
-<span class="sourceLineNo">248</span>     * @param value the data to be matched<a name="line.248"></a>
-<span class="sourceLineNo">249</span>     * @param offset offset of the data to be matched<a name="line.249"></a>
-<span class="sourceLineNo">250</span>     * @param length length of the data to be matched<a name="line.250"></a>
-<span class="sourceLineNo">251</span>     * @return 0 if a match was made, 1 otherwise<a name="line.251"></a>
-<span class="sourceLineNo">252</span>     */<a name="line.252"></a>
-<span class="sourceLineNo">253</span>    int compareTo(byte[] value, int offset, int length);<a name="line.253"></a>
-<span class="sourceLineNo">254</span>  }<a name="line.254"></a>
-<span class="sourceLineNo">255</span><a name="line.255"></a>
-<span class="sourceLineNo">256</span>  /**<a name="line.256"></a>
-<span class="sourceLineNo">257</span>   * Implementation of the Engine interface using Java's Pattern.<a name="line.257"></a>
-<span class="sourceLineNo">258</span>   * &lt;p&gt;<a name="line.258"></a>
-<span class="sourceLineNo">259</span>   * This is the default engine.<a name="line.259"></a>
-<span class="sourceLineNo">260</span>   */<a name="line.260"></a>
-<span class="sourceLineNo">261</span>  static class JavaRegexEngine implements Engine {<a name="line.261"></a>
-<span class="sourceLineNo">262</span>    private Charset charset = Charset.forName("UTF-8");<a name="line.262"></a>
-<span class="sourceLineNo">263</span>    private Pattern pattern;<a name="line.263"></a>
-<span class="sourceLineNo">264</span><a name="line.264"></a>
-<span class="sourceLineNo">265</span>    public JavaRegexEngine(String regex, int flags) {<a name="line.265"></a>
-<span class="sourceLineNo">266</span>      this.pattern = Pattern.compile(regex, flags);<a name="line.266"></a>
-<span class="sourceLineNo">267</span>    }<a name="line.267"></a>
-<span class="sourceLineNo">268</span><a name="line.268"></a>
-<span class="sourceLineNo">269</span>    @Override<a name="line.269"></a>
-<span class="sourceLineNo">270</span>    public String getPattern() {<a name="line.270"></a>
-<span class="sourceLineNo">271</span>      return pattern.toString();<a name="line.271"></a>
-<span class="sourceLineNo">272</span>    }<a name="line.272"></a>
-<span class="sourceLineNo">273</span><a name="line.273"></a>
-<span class="sourceLineNo">274</span>    @Override<a name="line.274"></a>
-<span class="sourceLineNo">275</span>    public int getFlags() {<a name="line.275"></a>
-<span class="sourceLineNo">276</span>      return pattern.flags();<a name="line.276"></a>
-<span class="sourceLineNo">277</span>    }<a name="line.277"></a>
-<span class="sourceLineNo">278</span><a name="line.278"></a>
-<span class="sourceLineNo">279</span>    @Override<a name="line.279"></a>
-<span class="sourceLineNo">280</span>    public String getCharset() {<a name="line.280"></a>
-<span class="sourceLineNo">281</span>      return charset.name();<a name="line.281"></a>
-<span class="sourceLineNo">282</span>    }<a name="line.282"></a>
-<span class="sourceLineNo">283</span><a name="line.283"></a>
-<span class="sourceLineNo">284</span>    @Override<a name="line.284"></a>
-<span class="sourceLineNo">285</span>    public void setCharset(String charset) {<a name="line.285"></a>
-<span class="sourceLineNo">286</span>      this.charset = Charset.forName(charset);<a name="line.286"></a>
-<span class="sourceLineNo">287</span>    }<a name="line.287"></a>
-<span class="sourceLineNo">288</span><a name="line.288"></a>
-<span class="sourceLineNo">289</span>    @Override<a name="line.289"></a>
-<span class="sourceLineNo">290</span>    public int compareTo(byte[] value, int offset, int length) {<a name="line.290"></a>
-<span class="sourceLineNo">291</span>      // Use find() for subsequence match instead of matches() (full sequence<a name="line.291"></a>
-<span class="sourceLineNo">292</span>      // match) to adhere to the principle of least surprise.<a name="line.292"></a>
-<span class="sourceLineNo">293</span>      String tmp;<a name="line.293"></a>
-<span class="sourceLineNo">294</span>      if (length &lt; value.length / 2) {<a name="line.294"></a>
-<span class="sourceLineNo">295</span>        // See HBASE-9428. Make a copy of the relevant part of the byte[],<a name="line.295"></a>
-<span class="sourceLineNo">296</span>        // or the JDK will copy the entire byte[] during String decode<a name="line.296"></a>
-<span class="sourceLineNo">297</span>        tmp = new String(Arrays.copyOfRange(value, offset, offset + length), charset);<a name="line.297"></a>
-<span class="sourceLineNo">298</span>      } else {<a name="line.298"></a>
-<span class="sourceLineNo">299</span>        tmp = new String(value, offset, length, charset);<a name="line.299"></a>
-<span class="sourceLineNo">300</span>      }<a name="line.300"></a>
-<span class="sourceLineNo">301</span>      return pattern.matcher(tmp).find() ? 0 : 1;<a name="line.301"></a>
-<span class="sourceLineNo">302</span>    }<a name="line.302"></a>
-<span class="sourceLineNo">303</span><a name="line.303"></a>
-<span class="sourceLineNo">304</span>    @Override<a name="line.304"></a>
-<span class="sourceLineNo">305</span>    public byte[] toByteArray() {<a name="line.305"></a>
-<span class="sourceLineNo">306</span>      ComparatorProtos.RegexStringComparator.Builder builder =<a name="line.306"></a>
-<span class="sourceLineNo">307</span>          ComparatorProtos.RegexStringComparator.newBuilder();<a name="line.307"></a>
-<span class="sourceLineNo">308</span>      builder.setPattern(pattern.pattern());<a name="line.308"></a>
-<span class="sourceLineNo">309</span>      builder.setPatternFlags(pattern.flags());<a name="line.309"></a>
-<span class="sourceLineNo">310</span>      builder.setCharset(charset.name());<a name="line.310"></a>
-<span class="sourceLineNo">311</span>      builder.setEngine(EngineType.JAVA.name());<a name="line.311"></a>
-<span class="sourceLineNo">312</span>      return builder.build().toByteArray();<a name="line.312"></a>
-<span class="sourceLineNo">313</span>    }<a name="line.313"></a>
-<span class="sourceLineNo">314</span>  }<a name="line.314"></a>
-<span class="sourceLineNo">315</span><a name="line.315"></a>
-<span class="sourceLineNo">316</span>  /**<a name="line.316"></a>
-<span class="sourceLineNo">317</span>   * Implementation of the Engine interface using Jruby's joni regex engine.<a name="line.317"></a>
-<span class="sourceLineNo">318</span>   * &lt;p&gt;<a name="line.318"></a>
-<span class="sourceLineNo">319</span>   * This engine operates on byte arrays directly so is expected to be more GC<a name="line.319"></a>
-<span class="sourceLineNo">320</span>   * friendly, and reportedly is twice as fast as Java's Pattern engine.<a name="line.320"></a>
-<span class="sourceLineNo">321</span>   * &lt;p&gt;<a name="line.321"></a>
-<span class="sourceLineNo">322</span>   * NOTE: Only the {@link Pattern} flags CASE_INSENSITIVE, DOTALL, and<a name="line.322"></a>
-<span class="sourceLineNo">323</span>   * MULTILINE are supported.<a name="line.323"></a>
-<span class="sourceLineNo">324</span>   */<a name="line.324"></a>
-<span class="sourceLineNo">325</span>  static class JoniRegexEngine implements Engine {<a name="line.325"></a>
-<span class="sourceLineNo">326</span>    private Encoding encoding = UTF8Encoding.INSTANCE;<a name="line.326"></a>
-<span class="sourceLineNo">327</span>    private String regex;<a name="line.327"></a>
-<span class="sourceLineNo">328</span>    private Regex pattern;<a name="line.328"></a>
-<span class="sourceLineNo">329</span><a name="line.329"></a>
-<span class="sourceLineNo">330</span>    public JoniRegexEngine(String regex, int flags) {<a name="line.330"></a>
-<span class="sourceLineNo">331</span>      this.regex = regex;<a name="line.331"></a>
-<span class="sourceLineNo">332</span>      byte[] b = Bytes.toBytes(regex);<a name="line.332"></a>
-<span class="sourceLineNo">333</span>      this.pattern = new Regex(b, 0, b.length, patternToJoniFlags(flags), encoding, Syntax.Java);<a name="line.333"></a>
-<span class="sourceLineNo">334</span>    }<a name="line.334"></a>
-<span class="sourceLineNo">335</span><a name="line.335"></a>
-<span class="sourceLineNo">336</span>    @Override<a name="line.336"></a>
-<span class="sourceLineNo">337</span>    public String getPattern() {<a name="line.337"></a>
-<span class="sourceLineNo">338</span>      return regex;<a name="line.338"></a>
-<span class="sourceLineNo">339</span>    }<a name="line.339"></a>
-<span class="sourceLineNo">340</span><a name="line.340"></a>
-<span class="sourceLineNo">341</span>    @Override<a name="line.341"></a>
-<span class="sourceLineNo">342</span>    public int getFlags() {<a name="line.342"></a>
-<span class="sourceLineNo">343</span>      return pattern.getOptions();<a name="line.343"></a>
-<span class="sourceLineNo">344</span>    }<a name="line.344"></a>
-<span class="sourceLineNo">345</span><a name="line.345"></a>
-<span class="sourceLineNo">346</span>    @Override<a name="line.346"></a>
-<span class="sourceLineNo">347</span>    public String getCharset() {<a name="line.347"></a>
-<span class="sourceLineNo">348</span>      return encoding.getCharsetName();<a name="line.348"></a>
-<span class="sourceLineNo">349</span>    }<a name="line.349"></a>
-<span class="sourceLineNo">350</span><a name="line.350"></a>
-<span class="sourceLineNo">351</span>    @Override<a name="line.351"></a>
-<span class="sourceLineNo">352</span>    public void setCharset(String name) {<a name="line.352"></a>
-<span class="sourceLineNo">353</span>      setEncoding(name);<a name="line.353"></a>
-<span class="sourceLineNo">354</span>    }<a name="line.354"></a>
-<span class="sourceLineNo">355</span><a name="line.355"></a>
-<span class="sourceLineNo">356</span>    @Override<a name="line.356"></a>
-<span class="sourceLineNo">357</span>    public int compareTo(byte[] value, int offset, int length) {<a name="line.357"></a>
-<span class="sourceLineNo">358</span>      // Use subsequence match instead of full sequence match to adhere to the<a name="line.358"></a>
-<span class="sourceLineNo">359</span>      // principle of least surprise.<a name="line.359"></a>
-<span class="sourceLineNo">360</span>      Matcher m = pattern.matcher(value);<a name="line.360"></a>
-<span class="sourceLineNo">361</span>      return m.search(offset, length, pattern.getOptions()) &lt; 0 ? 1 : 0;<a name="line.361"></a>
-<span class="sourceLineNo">362</span>    }<a name="line.362"></a>
-<span class="sourceLineNo">363</span><a name="line.363"></a>
-<span class="sourceLineNo">364</span>    @Override<a name="line.364"></a>
-<span class="sourceLineNo">365</span>    public byte[] toByteArray() {<a name="line.365"></a>
-<span class="sourceLineNo">366</span>      ComparatorProtos.RegexStringComparator.Builder builder =<a name="line.366"></a>
-<span class="sourceLineNo">367</span>          ComparatorProtos.RegexStringComparator.newBuilder();<a name="line.367"></a>
-<span class="sourceLineNo">368</span>        builder.setPattern(regex);<a name="line.368"></a>
-<span class="sourceLineNo">369</span>        builder.setPatternFlags(joniToPatternFlags(pattern.getOptions()));<a name="line.369"></a>
-<span class="sourceLineNo">370</span>        builder.setCharset(encoding.getCharsetName());<a name="line.370"></a>
-<span class="sourceLineNo">371</span>        builder.setEngine(EngineType.JONI.name());<a name="line.371"></a>
-<span class="sourceLineNo">372</span>        return builder.build().toByteArray();<a name="line.372"></a>
-<span class="sourceLineNo">373</span>    }<a name="line.373"></a>
-<span class="sourceLineNo">374</span><a name="line.374"></a>
-<span class="sourceLineNo">375</span>    private int patternToJoniFlags(int flags) {<a name="line.375"></a>
-<span class="sourceLineNo">376</span>      int newFlags = 0;<a name="line.376"></a>
-<span class="sourceLineNo">377</span>      if ((flags &amp; Pattern.CASE_INSENSITIVE) != 0) {<a name="line.377"></a>
-<span class="sourceLineNo">378</span>        newFlags |= Option.IGNORECASE;<a name="line.378"></a>
-<span class="sourceLineNo">379</span>      }<a name="line.379"></a>
-<span class="sourceLineNo">380</span>      if ((flags &amp; Pattern.DOTALL) != 0) {<a name="line.380"></a>
-<span class="sourceLineNo">381</span>        // This does NOT mean Pattern.MULTILINE<a name="line.381"></a>
-<span class="sourceLineNo">382</span>        newFlags |= Option.MULTILINE;<a name="line.382"></a>
-<span class="sourceLineNo">383</span>      }<a name="line.383"></a>
-<span class="sourceLineNo">384</span>      if ((flags &amp; Pattern.MULTILINE) != 0) {<a name="line.384"></a>
-<span class="sourceLineNo">385</span>        // This is what Java 8's Nashorn engine does when using joni and<a name="line.385"></a>
-<span class="sourceLineNo">386</span>        // translating Pattern's MULTILINE flag<a name="line.386"></a>
-<span class="sourceLineNo">387</span>        newFlags &amp;= ~Option.SINGLELINE;<a name="line.387"></a>
-<span class="sourceLineNo">388</span>        newFlags |= Option.NEGATE_SINGLELINE;<a name="line.388"></a>
-<span class="sourceLineNo">389</span>      }<a name="line.389"></a>
-<span class="sourceLineNo">390</span>      return newFlags;<a name="line.390"></a>
-<span class="sourceLineNo">391</span>    }<a name="line.391"></a>
-<span class="sourceLineNo">392</span><a name="line.392"></a>
-<span class="sourceLineNo">393</span>    private int joniToPatternFlags(int flags) {<a name="line.393"></a>
-<span class="sourceLineNo">394</span>      int newFlags = 0;<a name="line.394"></a>
-<span class="sourceLineNo">395</span>      if ((flags &amp; Option.IGNORECASE) != 0) {<a name="line.395"></a>
-<span class="sourceLineNo">396</span>        newFlags |= Pattern.CASE_INSENSITIVE;<a name="line.396"></a>
-<span class="sourceLineNo">397</span>      }<a name="line.397"></a>
-<span class="sourceLineNo">398</span>      // This does NOT mean Pattern.MULTILINE, this is equivalent to Pattern.DOTALL<a name="line.398"></a>
-<span class="sourceLineNo">399</span>      if ((flags &amp; Option.MULTILINE) != 0) {<a name="line.399"></a>
-<span class="sourceLineNo">400</span>        newFlags |= Pattern.DOTALL;<a name="line.400"></a>
-<span class="sourceLineNo">401</span>      }<a name="line.401"></a>
-<span class="sourceLineNo">402</span>      // This means Pattern.MULTILINE. Nice<a name="line.402"></a>
-<span class="sourceLineNo">403</span>      if ((flags &amp; Option.NEGATE_SINGLELINE) != 0) {<a name="line.403"></a>
-<span class="sourceLineNo">404</span>        newFlags |= Pattern.MULTILINE;<a name="line.404"></a>
-<span class="sourceLineNo">405</span>      }<a name="line.405"></a>
-<span class="sourceLineNo">406</span>      return newFlags;<a name="line.406"></a>
-<span class="sourceLineNo">407</span>    }<a name="line.407"></a>
-<span class="sourceLineNo">408</span><a name="line.408"></a>
-<span class="sourceLineNo">409</span>    private void setEncoding(String name) {<a name="line.409"></a>
-<span class="sourceLineNo">410</span>      EncodingDB.Entry e = EncodingDB.getEncodings().get(Bytes.toBytes(name));<a name="line.410"></a>
-<span class="sourceLineNo">411</span>      if (e != null) {<a name="line.411"></a>
-<span class="sourceLineNo">412</span>        encoding = e.getEncoding();<a name="line.412"></a>
-<span class="sourceLineNo">413</span>      } else {<a name="line.413"></a>
-<span class="sourceLineNo">414</span>        throw new IllegalCharsetNameException(name);<a name="line.414"></a>
-<span class="sourceLineNo">415</span>      }<a name="line.415"></a>
-<span class="sourceLineNo">416</span>    }<a name="line.416"></a>
-<span class="sourceLineNo">417</span>  }<a name="line.417"></a>
-<span class="sourceLineNo">418</span>}<a name="line.418"></a>
+<span class="sourceLineNo">072</span>@SuppressWarnings("ComparableType") // Should this move to Comparator usage?<a name="line.72"></a>
+<span class="sourceLineNo">073</span>public class RegexStringComparator extends ByteArrayComparable {<a name="line.73"></a>
+<span class="sourceLineNo">074</span><a name="line.74"></a>
+<span class="sourceLineNo">075</span>  private static final Logger LOG = LoggerFactory.getLogger(RegexStringComparator.class);<a name="line.75"></a>
+<span class="sourceLineNo">076</span><a name="line.76"></a>
+<span class="sourceLineNo">077</span>  private Engine engine;<a name="line.77"></a>
+<span class="sourceLineNo">078</span><a name="line.78"></a>
+<span class="sourceLineNo">079</span>  /** Engine implementation type (default=JAVA) */<a name="line.79"></a>
+<span class="sourceLineNo">080</span>  @InterfaceAudience.Public<a name="line.80"></a>
+<span class="sourceLineNo">081</span>  public enum EngineType {<a name="line.81"></a>
+<span class="sourceLineNo">082</span>    JAVA,<a name="line.82"></a>
+<span class="sourceLineNo">083</span>    JONI<a name="line.83"></a>
+<span class="sourceLineNo">084</span>  }<a name="line.84"></a>
+<span class="sourceLineNo">085</span><a name="line.85"></a>
+<span class="sourceLineNo">086</span>  /**<a name="line.86"></a>
+<span class="sourceLineNo">087</span>   * Constructor<a name="line.87"></a>
+<span class="sourceLineNo">088</span>   * Adds Pattern.DOTALL to the underlying Pattern<a name="line.88"></a>
+<span class="sourceLineNo">089</span>   * @param expr a valid regular expression<a name="line.89"></a>
+<span class="sourceLineNo">090</span>   */<a name="line.90"></a>
+<span class="sourceLineNo">091</span>  public RegexStringComparator(String expr) {<a name="line.91"></a>
+<span class="sourceLineNo">092</span>    this(expr, Pattern.DOTALL);<a name="line.92"></a>
+<span class="sourceLineNo">093</span>  }<a name="line.93"></a>
+<span class="sourceLineNo">094</span><a name="line.94"></a>
+<span class="sourceLineNo">095</span>  /**<a name="line.95"></a>
+<span class="sourceLineNo">096</span>   * Constructor<a name="line.96"></a>
+<span class="sourceLineNo">097</span>   * Adds Pattern.DOTALL to the underlying Pattern<a name="line.97"></a>
+<span class="sourceLineNo">098</span>   * @param expr a valid regular expression<a name="line.98"></a>
+<span class="sourceLineNo">099</span>   * @param engine engine implementation type<a name="line.99"></a>
+<span class="sourceLineNo">100</span>   */<a name="line.100"></a>
+<span class="sourceLineNo">101</span>  public RegexStringComparator(String expr, EngineType engine) {<a name="line.101"></a>
+<span class="sourceLineNo">102</span>    this(expr, Pattern.DOTALL, engine);<a name="line.102"></a>
+<span class="sourceLineNo">103</span>  }<a name="line.103"></a>
+<span class="sourceLineNo">104</span><a name="line.104"></a>
+<span class="sourceLineNo">105</span>  /**<a name="line.105"></a>
+<span class="sourceLineNo">106</span>   * Constructor<a name="line.106"></a>
+<span class="sourceLineNo">107</span>   * @param expr a valid regular expression<a name="line.107"></a>
+<span class="sourceLineNo">108</span>   * @param flags java.util.regex.Pattern flags<a name="line.108"></a>
+<span class="sourceLineNo">109</span>   */<a name="line.109"></a>
+<span class="sourceLineNo">110</span>  public RegexStringComparator(String expr, int flags) {<a name="line.110"></a>
+<span class="sourceLineNo">111</span>    this(expr, flags, EngineType.JAVA);<a name="line.111"></a>
+<span class="sourceLineNo">112</span>  }<a name="line.112"></a>
+<span class="sourceLineNo">113</span><a name="line.113"></a>
+<span class="sourceLineNo">114</span>  /**<a name="line.114"></a>
+<span class="sourceLineNo">115</span>   * Constructor<a name="line.115"></a>
+<span class="sourceLineNo">116</span>   * @param expr a valid regular expression<a name="line.116"></a>
+<span class="sourceLineNo">117</span>   * @param flags java.util.regex.Pattern flags<a name="line.117"></a>
+<span class="sourceLineNo">118</span>   * @param engine engine implementation type<a name="line.118"></a>
+<span class="sourceLineNo">119</span>   */<a name="line.119"></a>
+<span class="sourceLineNo">120</span>  public RegexStringComparator(String expr, int flags, EngineType engine) {<a name="line.120"></a>
+<span class="sourceLineNo">121</span>    super(Bytes.toBytes(expr));<a name="line.121"></a>
+<span class="sourceLineNo">122</span>    switch (engine) {<a name="line.122"></a>
+<span class="sourceLineNo">123</span>      case JAVA:<a name="line.123"></a>
+<span class="sourceLineNo">124</span>        this.engine = new JavaRegexEngine(expr, flags);<a name="line.124"></a>
+<span class="sourceLineNo">125</span>        break;<a name="line.125"></a>
+<span class="sourceLineNo">126</span>      case JONI:<a name="line.126"></a>
+<span class="sourceLineNo">127</span>        this.engine = new JoniRegexEngine(expr, flags);<a name="line.127"></a>
+<span class="sourceLineNo">128</span>        break;<a name="line.128"></a>
+<span class="sourceLineNo">129</span>    }<a name="line.129"></a>
+<span class="sourceLineNo">130</span>  }<a name="line.130"></a>
+<span class="sourceLineNo">131</span><a name="line.131"></a>
+<span class="sourceLineNo">132</span>  /**<a name="line.132"></a>
+<span class="sourceLineNo">133</span>   * Specifies the {@link Charset} to use to convert the row key to a String.<a name="line.133"></a>
+<span class="sourceLineNo">134</span>   * &lt;p&gt;<a name="line.134"></a>
+<span class="sourceLineNo">135</span>   * The row key needs to be converted to a String in order to be matched<a name="line.135"></a>
+<span class="sourceLineNo">136</span>   * against the regular expression.  This method controls which charset is<a name="line.136"></a>
+<span class="sourceLineNo">137</span>   * used to do this conversion.<a name="line.137"></a>
+<span class="sourceLineNo">138</span>   * &lt;p&gt;<a name="line.138"></a>
+<span class="sourceLineNo">139</span>   * If the row key is made of arbitrary bytes, the charset {@code ISO-8859-1}<a name="line.139"></a>
+<span class="sourceLineNo">140</span>   * is recommended.<a name="line.140"></a>
+<span class="sourceLineNo">141</span>   * @param charset The charset to use.<a name="line.141"></a>
+<span class="sourceLineNo">142</span>   */<a name="line.142"></a>
+<span class="sourceLineNo">143</span>  public void setCharset(final Charset charset) {<a name="line.143"></a>
+<span class="sourceLineNo">144</span>    engine.setCharset(charset.name());<a name="line.144"></a>
+<span class="sourceLineNo">145</span>  }<a name="line.145"></a>
+<span class="sourceLineNo">146</span><a name="line.146"></a>
+<span class="sourceLineNo">147</span>  @Override<a name="line.147"></a>
+<span class="sourceLineNo">148</span>  public int compareTo(byte[] value, int offset, int length) {<a name="line.148"></a>
+<span class="sourceLineNo">149</span>    return engine.compareTo(value, offset, length);<a name="line.149"></a>
+<span class="sourceLineNo">150</span>  }<a name="line.150"></a>
+<span class="sourceLineNo">151</span><a name="line.151"></a>
+<span class="sourceLineNo">152</span>  /**<a name="line.152"></a>
+<span class="sourceLineNo">153</span>   * @return The comparator serialized using pb<a name="line.153"></a>
+<span class="sourceLineNo">154</span>   */<a name="line.154"></a>
+<span class="sourceLineNo">155</span>  @Override<a name="line.155"></a>
+<span class="sourceLineNo">156</span>  public byte [] toByteArray() {<a name="line.156"></a>
+<span class="sourceLineNo">157</span>    return engine.toByteArray();<a name="line.157"></a>
+<span class="sourceLineNo">158</span>  }<a name="line.158"></a>
+<span class="sourceLineNo">159</span><a name="line.159"></a>
+<span class="sourceLineNo">160</span>  /**<a name="line.160"></a>
+<span class="sourceLineNo">161</span>   * @param pbBytes A pb serialized {@link RegexStringComparator} instance<a name="line.161"></a>
+<span class="sourceLineNo">162</span>   * @return An instance of {@link RegexStringComparator} made from &lt;code&gt;bytes&lt;/code&gt;<a name="line.162"></a>
+<span class="sourceLineNo">163</span>   * @throws DeserializationException<a name="line.163"></a>
+<span class="sourceLineNo">164</span>   * @see #toByteArray<a name="line.164"></a>
+<span class="sourceLineNo">165</span>   */<a name="line.165"></a>
+<span class="sourceLineNo">166</span>  public static RegexStringComparator parseFrom(final byte [] pbBytes)<a name="line.166"></a>
+<span class="sourceLineNo">167</span>  throws DeserializationException {<a name="line.167"></a>
+<span class="sourceLineNo">168</span>    ComparatorProtos.RegexStringComparator proto;<a name="line.168"></a>
+<span class="sourceLineNo">169</span>    try {<a name="line.169"></a>
+<span class="sourceLineNo">170</span>      proto = ComparatorProtos.RegexStringComparator.parseFrom(pbBytes);<a name="line.170"></a>
+<span class="sourceLineNo">171</span>    } catch (InvalidProtocolBufferException e) {<a name="line.171"></a>
+<span class="sourceLineNo">172</span>      throw new DeserializationException(e);<a name="line.172"></a>
+<span class="sourceLineNo">173</span>    }<a name="line.173"></a>
+<span class="sourceLineNo">174</span>    RegexStringComparator comparator;<a name="line.174"></a>
+<span class="sourceLineNo">175</span>    if (proto.hasEngine()) {<a name="line.175"></a>
+<span class="sourceLineNo">176</span>      EngineType engine = EngineType.valueOf(proto.getEngine());<a name="line.176"></a>
+<span class="sourceLineNo">177</span>      comparator = new RegexStringComparator(proto.getPattern(), proto.getPatternFlags(),<a name="line.177"></a>
+<span class="sourceLineNo">178</span>        engine);<a name="line.178"></a>
+<span class="sourceLineNo">179</span>    } else {<a name="line.179"></a>
+<span class="sourceLineNo">180</span>      comparator = new RegexStringComparator(proto.getPattern(), proto.getPatternFlags());<a name="line.180"></a>
+<span class="sourceLineNo">181</span>    }<a name="line.181"></a>
+<span class="sourceLineNo">182</span>    String charset = proto.getCharset();<a name="line.182"></a>
+<span class="sourceLineNo">183</span>    if (charset.length() &gt; 0) {<a name="line.183"></a>
+<span class="sourceLineNo">184</span>      try {<a name="line.184"></a>
+<span class="sourceLineNo">185</span>        comparator.getEngine().setCharset(charset);<a name="line.185"></a>
+<span class="sourceLineNo">186</span>      } catch (IllegalCharsetNameException e) {<a name="line.186"></a>
+<span class="sourceLineNo">187</span>        LOG.error("invalid charset", e);<a name="line.187"></a>
+<span class="sourceLineNo">188</span>      }<a name="line.188"></a>
+<span class="sourceLineNo">189</span>    }<a name="line.189"></a>
+<span class="sourceLineNo">190</span>    return comparator;<a name="line.190"></a>
+<span class="sourceLineNo">191</span>  }<a name="line.191"></a>
+<span class="sourceLineNo">192</span><a name="line.192"></a>
+<span class="sourceLineNo">193</span>  /**<a name="line.193"></a>
+<span class="sourceLineNo">194</span>   * @param other<a name="line.194"></a>
+<span class="sourceLineNo">195</span>   * @return true if and only if the fields of the comparator that are serialized<a name="line.195"></a>
+<span class="sourceLineNo">196</span>   * are equal to the corresponding fields in other.  Used for testing.<a name="line.196"></a>
+<span class="sourceLineNo">197</span>   */<a name="line.197"></a>
+<span class="sourceLineNo">198</span>  @Override<a name="line.198"></a>
+<span class="sourceLineNo">199</span>  boolean areSerializedFieldsEqual(ByteArrayComparable other) {<a name="line.199"></a>
+<span class="sourceLineNo">200</span>    if (other == this) return true;<a name="line.200"></a>
+<span class="sourceLineNo">201</span>    if (!(other instanceof RegexStringComparator)) return false;<a name="line.201"></a>
+<span class="sourceLineNo">202</span>    RegexStringComparator comparator = (RegexStringComparator)other;<a name="line.202"></a>
+<span class="sourceLineNo">203</span>    return super.areSerializedFieldsEqual(comparator)<a name="line.203"></a>
+<span class="sourceLineNo">204</span>      &amp;&amp; engine.getClass().isInstance(comparator.getEngine())<a name="line.204"></a>
+<span class="sourceLineNo">205</span>      &amp;&amp; engine.getPattern().equals(comparator.getEngine().getPattern())<a name="line.205"></a>
+<span class="sourceLineNo">206</span>      &amp;&amp; engine.getFlags() == comparator.getEngine().getFlags()<a name="line.206"></a>
+<span class="sourceLineNo">207</span>      &amp;&amp; engine.getCharset().equals(comparator.getEngine().getCharset());<a name="line.207"></a>
+<span class="sourceLineNo">208</span>  }<a name="line.208"></a>
+<span class="sourceLineNo">209</span><a name="line.209"></a>
+<span class="sourceLineNo">210</span>  Engine getEngine() {<a name="line.210"></a>
+<span class="sourceLineNo">211</span>    return engine;<a name="line.211"></a>
+<span class="sourceLineNo">212</span>  }<a name="line.212"></a>
+<span class="sourceLineNo">213</span><a name="line.213"></a>
+<span class="sourceLineNo">214</span>  /**<a name="line.214"></a>
+<span class="sourceLineNo">215</span>   * This is an internal interface for abstracting access to different regular<a name="line.215"></a>
+<span class="sourceLineNo">216</span>   * expression matching engines.<a name="line.216"></a>
+<span class="sourceLineNo">217</span>   */<a name="line.217"></a>
+<span class="sourceLineNo">218</span>  static interface Engine {<a name="line.218"></a>
+<span class="sourceLineNo">219</span>    /**<a name="line.219"></a>
+<span class="sourceLineNo">220</span>     * Returns the string representation of the configured regular expression<a name="line.220"></a>
+<span class="sourceLineNo">221</span>     * for matching<a name="line.221"></a>
+<span class="sourceLineNo">222</span>     */<a name="line.222"></a>
+<span class="sourceLineNo">223</span>    String getPattern();<a name="line.223"></a>
+<span class="sourceLineNo">224</span><a name="line.224"></a>
+<span class="sourceLineNo">225</span>    /**<a name="line.225"></a>
+<span class="sourceLineNo">226</span>     * Returns the set of configured match flags, a bit mask that may include<a name="line.226"></a>
+<span class="sourceLineNo">227</span>     * {@link Pattern} flags<a name="line.227"></a>
+<span class="sourceLineNo">228</span>     */<a name="line.228"></a>
+<span class="sourceLineNo">229</span>    int getFlags();<a name="line.229"></a>
+<span class="sourceLineNo">230</span><a name="line.230"></a>
+<span class="sourceLineNo">231</span>    /**<a name="line.231"></a>
+<span class="sourceLineNo">232</span>     * Returns the name of the configured charset<a name="line.232"></a>
+<span class="sourceLineNo">233</span>     */<a name="line.233"></a>
+<span class="sourceLineNo">234</span>    String getCharset();<a name="line.234"></a>
+<span class="sourceLineNo">235</span><a name="line.235"></a>
+<span class="sourceLineNo">236</span>    /**<a name="line.236"></a>
+<span class="sourceLineNo">237</span>     * Set the charset used when matching<a name="line.237"></a>
+<span class="sourceLineNo">238</span>     * @param charset the name of the desired charset for matching<a name="line.238"></a>
+<span class="sourceLineNo">239</span>     */<a name="line.239"></a>
+<span class="sourceLineNo">240</span>    void setCharset(final String charset);<a name="line.240"></a>
+<span class="sourceLineNo">241</span><a name="line.241"></a>
+<span class="sourceLineNo">242</span>    /**<a name="line.242"></a>
+<span class="sourceLineNo">243</span>     * Return the serialized form of the configured matcher<a name="line.243"></a>
+<span class="sourceLineNo">244</span>     */<a name="line.244"></a>
+<span class="sourceLineNo">245</span>    byte [] toByteArray();<a name="line.245"></a>
+<span class="sourceLineNo">246</span><a name="line.246"></a>
+<span class="sourceLineNo">247</span>    /**<a name="line.247"></a>
+<span class="sourceLineNo">248</span>     * Match the given input against the configured pattern<a name="line.248"></a>
+<span class="sourceLineNo">249</span>     * @param value the data to be matched<a name="line.249"></a>
+<span class="sourceLineNo">250</span>     * @param offset offset of the data to be matched<a name="line.250"></a>
+<span class="sourceLineNo">251</span>     * @param length length of the data to be matched<a name="line.251"></a>
+<span class="sourceLineNo">252</span>     * @return 0 if a match was made, 1 otherwise<a name="line.252"></a>
+<span class="sourceLineNo">253</span>     */<a name="line.253"></a>
+<span class="sourceLineNo">254</span>    int compareTo(byte[] value, int offset, int length);<a name="line.254"></a>
+<span class="sourceLineNo">255</span>  }<a name="line.255"></a>
+<span class="sourceLineNo">256</span><a name="line.256"></a>
+<span class="sourceLineNo">257</span>  /**<a name="line.257"></a>
+<span class="sourceLineNo">258</span>   * Implementation of the Engine interface using Java's Pattern.<a name="line.258"></a>
+<span class="sourceLineNo">259</span>   * &lt;p&gt;<a name="line.259"></a>
+<span class="sourceLineNo">260</span>   * This is the default engine.<a name="line.260"></a>
+<span class="sourceLineNo">261</span>   */<a name="line.261"></a>
+<span class="sourceLineNo">262</span>  static class JavaRegexEngine implements Engine {<a name="line.262"></a>
+<span class="sourceLineNo">263</span>    private Charset charset = Charset.forName("UTF-8");<a name="line.263"></a>
+<span class="sourceLineNo">264</span>    private Pattern pattern;<a name="line.264"></a>
+<span class="sourceLineNo">265</span><a name="line.265"></a>
+<span class="sourceLineNo">266</span>    public JavaRegexEngine(String regex, int flags) {<a name="line.266"></a>
+<span class="sourceLineNo">267</span>      this.pattern = Pattern.compile(regex, flags);<a name="line.267"></a>
+<span class="sourceLineNo">268</span>    }<a name="line.268"></a>
+<span class="sourceLineNo">269</span><a name="line.269"></a>
+<span class="sourceLineNo">270</span>    @Override<a name="line.270"></a>
+<span class="sourceLineNo">271</span>    public String getPattern() {<a name="line.271"></a>
+<span class="sourceLineNo">272</span>      return pattern.toString();<a name="line.272"></a>
+<span class="sourceLineNo">273</span>    }<a name="line.273"></a>
+<span class="sourceLineNo">274</span><a name="line.274"></a>
+<span class="sourceLineNo">275</span>    @Override<a name="line.275"></a>
+<span class="sourceLineNo">276</span>    public int getFlags() {<a name="line.276"></a>
+<span class="sourceLineNo">277</span>      return pattern.flags();<a name="line.277"></a>
+<span class="sourceLineNo">278</span>    }<a name="line.278"></a>
+<span class="sourceLineNo">279</span><a name="line.279"></a>
+<span class="sourceLineNo">280</span>    @Override<a name="line.280"></a>
+<span class="sourceLineNo">281</span>    public String getCharset() {<a name="line.281"></a>
+<span class="sourceLineNo">282</span>      return charset.name();<a name="line.282"></a>
+<span class="sourceLineNo">283</span>    }<a name="line.283"></a>
+<span class="sourceLineNo">284</span><a name="line.284"></a>
+<span class="sourceLineNo">285</span>    @Override<a name="line.285"></a>
+<span class="sourceLineNo">286</span>    public void setCharset(String charset) {<a name="line.286"></a>
+<span class="sourceLineNo">287</span>      this.charset = Charset.forName(charset);<a name="line.287"></a>
+<span class="sourceLineNo">288</span>    }<a name="line.288"></a>
+<span class="sourceLineNo">289</span><a name="line.289"></a>
+<span class="sourceLineNo">290</span>    @Override<a name="line.290"></a>
+<span class="sourceLineNo">291</span>    public int compareTo(byte[] value, int offset, int length) {<a name="line.291"></a>
+<span class="sourceLineNo">292</span>      // Use find() for subsequence match instead of matches() (full sequence<a name="line.292"></a>
+<span class="sourceLineNo">293</span>      // match) to adhere to the principle of least surprise.<a name="line.293"></a>
+<span class="sourceLineNo">294</span>      String tmp;<a name="line.294"></a>
+<span class="sourceLineNo">295</span>      if (length &lt; value.length / 2) {<a name="line.295"></a>
+<span class="sourceLineNo">296</span>        // See HBASE-9428. Make a copy of the relevant part of the byte[],<a name="line.296"></a>
+<span class="sourceLineNo">297</span>        // or the JDK will copy the entire byte[] during String decode<a name="line.297"></a>
+<span class="sourceLineNo">298</span>        tmp = new String(Arrays.copyOfRange(value, offset, offset + length), charset);<a name="line.298"></a>
+<span class="sourceLineNo">299</span>      } else {<a name="line.299"></a>
+<span class="sourceLineNo">300</span>        tmp = new String(value, offset, length, charset);<a name="line.300"></a>
+<span class="sourceLineNo">301</span>      }<a name="line.301"></a>
+<span class="sourceLineNo">302</span>      return pattern.matcher(tmp).find() ? 0 : 1;<a name="line.302"></a>
+<span class="sourceLineNo">303</span>    }<a name="line.303"></a>
+<span class="sourceLineNo">304</span><a name="line.304"></a>
+<span class="sourceLineNo">305</span>    @Override<a name="line.305"></a>
+<span class="sourceLineNo">306</span>    public byte[] toByteArray() {<a name="line.306"></a>
+<span class="sourceLineNo">307</span>      ComparatorProtos.RegexStringComparator.Builder builder =<a name="line.307"></a>
+<span class="sourceLineNo">308</span>          ComparatorProtos.RegexStringComparator.newBuilder();<a name="line.308"></a>
+<span class="sourceLineNo">309</span>      builder.setPattern(pattern.pattern());<a name="line.309"></a>
+<span class="sourceLineNo">310</span>      builder.setPatternFlags(pattern.flags());<a name="line.310"></a>
+<span class="sourceLineNo">311</span>      builder.setCharset(charset.name());<a name="line.311"></a>
+<span class="sourceLineNo">312</span>      builder.setEngine(EngineType.JAVA.name());<a name="line.312"></a>
+<span class="sourceLineNo">313</span>      return builder.build().toByteArray();<a name="line.313"></a>
+<span class="sourceLineNo">314</span>    }<a name="line.314"></a>
+<span class="sourceLineNo">315</span>  }<a name="line.315"></a>
+<span class="sourceLineNo">316</span><a name="line.316"></a>
+<span class="sourceLineNo">317</span>  /**<a name="line.317"></a>
+<span class="sourceLineNo">318</span>   * Implementation of the Engine interface using Jruby's joni regex engine.<a name="line.318"></a>
+<span class="sourceLineNo">319</span>   * &lt;p&gt;<a name="line.319"></a>
+<span class="sourceLineNo">320</span>   * This engine operates on byte arrays directly so is expected to be more GC<a name="line.320"></a>
+<span class="sourceLineNo">321</span>   * friendly, and reportedly is twice as fast as Java's Pattern engine.<a name="line.321"></a>
+<span class="sourceLineNo">322</span>   * &lt;p&gt;<a name="line.322"></a>
+<span class="sourceLineNo">323</span>   * NOTE: Only the {@link Pattern} flags CASE_INSENSITIVE, DOTALL, and<a name="line.323"></a>
+<span class="sourceLineNo">324</span>   * MULTILINE are supported.<a name="line.324"></a>
+<span class="sourceLineNo">325</span>   */<a name="line.325"></a>
+<span class="sourceLineNo">326</span>  static class JoniRegexEngine implements Engine {<a name="line.326"></a>
+<span class="sourceLineNo">327</span>    private Encoding encoding = UTF8Encoding.INSTANCE;<a name="line.327"></a>
+<span class="sourceLineNo">328</span>    private String regex;<a name="line.328"></a>
+<span class="sourceLineNo">329</span>    private Regex pattern;<a name="line.329"></a>
+<span class="sourceLineNo">330</span><a name="line.330"></a>
+<span class="sourceLineNo">331</span>    public JoniRegexEngine(String regex, int flags) {<a name="line.331"></a>
+<span class="sourceLineNo">332</span>      this.regex = regex;<a name="line.332"></a>
+<span class="sourceLineNo">333</span>      byte[] b = Bytes.toBytes(regex);<a name="line.333"></a>
+<span class="sourceLineNo">334</span>      this.pattern = new Regex(b, 0, b.length, patternToJoniFlags(flags), encoding, Syntax.Java);<a name="line.334"></a>
+<span class="sourceLineNo">335</span>    }<a name="line.335"></a>
+<span class="sourceLineNo">336</span><a name="line.336"></a>
+<span class="sourceLineNo">337</span>    @Override<a name="line.337"></a>
+<span class="sourceLineNo">338</span>    public String getPattern() {<a name="line.338"></a>
+<span class="sourceLineNo">339</span>      return regex;<a name="line.339"></a>
+<span class="sourceLineNo">340</span>    }<a name="line.340"></a>
+<span class="sourceLineNo">341</span><a name="line.341"></a>
+<span class="sourceLineNo">342</span>    @Override<a name="line.342"></a>
+<span class="sourceLineNo">343</span>    public int getFlags() {<a name="line.343"></a>
+<span class="sourceLineNo">344</span>      return pattern.getOptions();<a name="line.344"></a>
+<span class="sourceLineNo">345</span>    }<a name="line.345"></a>
+<span class="sourceLineNo">346</span><a name="line.346"></a>
+<span class="sourceLineNo">347</span>    @Override<a name="line.347"></a>
+<span class="sourceLineNo">348</span>    public String getCharset() {<a name="line.348"></a>
+<span class="sourceLineNo">349</span>      return encoding.getCharsetName();<a name="line.349"></a>
+<span class="sourceLineNo">350</span>    }<a name="line.350"></a>
+<span class="sourceLineNo">351</span><a name="line.351"></a>
+<span class="sourceLineNo">352</span>    @Override<a name="line.352"></a>
+<span class="sourceLineNo">353</span>    public void setCharset(String name) {<a name="line.353"></a>
+<span class="sourceLineNo">354</span>      setEncoding(name);<a name="line.354"></a>
+<span class="sourceLineNo">355</span>    }<a name="line.355"></a>
+<span class="sourceLineNo">356</span><a name="line.356"></a>
+<span class="sourceLineNo">357</span>    @Override<a name="line.357"></a>
+<span class="sourceLineNo">358</span>    public int compareTo(byte[] value, int offset, int length) {<a name="line.358"></a>
+<span class="sourceLineNo">359</span>      // Use subsequence match instead of full sequence match to adhere to the<a name="line.359"></a>
+<span class="sourceLineNo">360</span>      // principle of least surprise.<a name="line.360"></a>
+<span class="sourceLineNo">361</span>      Matcher m = pattern.matcher(value);<a name="line.361"></a>
+<span class="sourceLineNo">362</span>      return m.search(offset, length, pattern.getOptions()) &lt; 0 ? 1 : 0;<a name="line.362"></a>
+<span class="sourceLineNo">363</span>    }<a name="line.363"></a>
+<span class="sourceLineNo">364</span><a name="line.364"></a>
+<span class="sourceLineNo">365</span>    @Override<a name="line.365"></a>
+<span class="sourceLineNo">366</span>    public byte[] toByteArray() {<a name="line.366"></a>
+<span class="sourceLineNo">367</span>      ComparatorProtos.RegexStringComparator.Builder builder =<a name="line.367"></a>
+<span class="sourceLineNo">368</span>          ComparatorProtos.RegexStringComparator.newBuilder();<a name="line.368"></a>
+<span class="sourceLineNo">369</span>        builder.setPattern(regex);<a name="line.369"></a>
+<span class="sourceLineNo">370</span>        builder.setPatternFlags(joniToPatternFlags(pattern.getOptions()));<a name="line.370"></a>
+<span class="sourceLineNo">371</span>        builder.setCharset(encoding.getCharsetName());<a name="line.371"></a>
+<span class="sourceLineNo">372</span>        builder.setEngine(EngineType.JONI.name());<a name="line.372"></a>
+<span class="sourceLineNo">373</span>        return builder.build().toByteArray();<a name="line.373"></a>
+<span class="sourceLineNo">374</span>    }<a name="line.374"></a>
+<span class="sourceLineNo">375</span><a name="line.375"></a>
+<span class="sourceLineNo">376</span>    private int patternToJoniFlags(int flags) {<a name="line.376"></a>
+<span class="sourceLineNo">377</span>      int newFlags = 0;<a name="line.377"></a>
+<span class="sourceLineNo">378</span>      if ((flags &amp; Pattern.CASE_INSENSITIVE) != 0) {<a name="line.378"></a>
+<span class="sourceLineNo">379</span>        newFlags |= Option.IGNORECASE;<a name="line.379"></a>
+<span class="sourceLineNo">380</span>      }<a name="line.380"></a>
+<span class="sourceLineNo">381</span>      if ((flags &amp; Pattern.DOTALL) != 0) {<a name="line.381"></a>
+<span class="sourceLineNo">382</span>        // This does NOT mean Pattern.MULTILINE<a name="line.382"></a>
+<span class="sourceLineNo">383</span>        newFlags |= Option.MULTILINE;<a name="line.383"></a>
+<span class="sourceLineNo">384</span>      }<a name="line.384"></a>
+<span class="sourceLineNo">385</span>      if ((flags &amp; Pattern.MULTILINE) != 0) {<a name="line.385"></a>
+<span class="sourceLineNo">386</span>        // This is what Java 8's Nashorn engine does when using joni and<a name="line.386"></a>
+<span class="sourceLineNo">387</span>        // translating Pattern's MULTILINE flag<a name="line.387"></a>
+<span class="sourceLineNo">388</span>        newFlags &amp;= ~Option.SINGLELINE;<a name="line.388"></a>
+<span class="sourceLineNo">389</span>        newFlags |= Option.NEGATE_SINGLELINE;<a name="line.389"></a>
+<span class="sourceLineNo">390</span>      }<a name="line.390"></a>
+<span class="sourceLineNo">391</span>      return newFlags;<a name="line.391"></a>
+<span class="sourceLineNo">392</span>    }<a name="line.392"></a>
+<span class="sourceLineNo">393</span><a name="line.393"></a>
+<span class="sourceLineNo">394</span>    private int joniToPatternFlags(int flags) {<a name="line.394"></a>
+<span class="sourceLineNo">395</span>      int newFlags = 0;<a name="line.395"></a>
+<span class="sourceLineNo">396</span>      if ((flags &amp; Option.IGNORECASE) != 0) {<a name="line.396"></a>
+<span class="sourceLineNo">397</span>        newFlags |= Pattern.CASE_INSENSITIVE;<a name="line.397"></a>
+<span class="sourceLineNo">398</span>      }<a name="line.398"></a>
+<span class="sourceLineNo">399</span>      // This does NOT mean Pattern.MULTILINE, this is equivalent to Pattern.DOTALL<a name="line.399"></a>
+<span class="sourceLineNo">400</span>      if ((flags &amp; Option.MULTILINE) != 0) {<a name="line.400"></a>
+<span class="sourceLineNo">401</span>        newFlags |= Pattern.DOTALL;<a name="line.401"></a>
+<span class="sourceLineNo">402</span>      }<a name="line.402"></a>
+<span class="sourceLineNo">403</span>      // This means Pattern.MULTILINE. Nice<a name="line.403"></a>
+<span class="sourceLineNo">404</span>      if ((flags &amp; Option.NEGATE_SINGLELINE) != 0) {<a name="line.404"></a>
+<span class="sourceLineNo">405</span>        newFlags |= Pattern.MULTILINE;<a name="line.405"></a>
+<span class="sourceLineNo">406</span>      }<a name="line.406"></a>
+<span class="sourceLineNo">407</span>      return newFlags;<a name="line.407"></a>
+<span class="sourceLineNo">408</span>    }<a name="line.408"></a>
+<span class="sourceLineNo">409</span><a name="line.409"></a>
+<span class="sourceLineNo">410</span>    private void setEncoding(String name) {<a name="line.410"></a>
+<span class="sourceLineNo">411</span>      EncodingDB.Entry e = EncodingDB.getEncodings().get(Bytes.toBytes(name));<a name="line.411"></a>
+<span class="sourceLineNo">412</span>      if (e != null) {<a name="line.412"></a>
+<span class="sourceLineNo">413</span>        encoding = e.getEncoding();<a name="line.413"></a>
+<span class="sourceLineNo">414</span>      } else {<a name="line.414"></a>
+<span class="sourceLineNo">415</span>        throw new IllegalCharsetNameException(name);<a name="line.415"></a>
+<span class="sourceLineNo">416</span>      }<a name="line.416"></a>
+<span class="sourceLineNo">417</span>    }<a name="line.417"></a>
+<span class="sourceLineNo">418</span>  }<a name="line.418"></a>
+<span class="sourceLineNo">419</span>}<a name="line.419"></a>