Modified: websites/production/commons/content/proper/commons-compress/apidocs/src-html/org/apache/commons/compress/compressors/lz4/FramedLZ4CompressorInputStream.html ============================================================================== --- websites/production/commons/content/proper/commons-compress/apidocs/src-html/org/apache/commons/compress/compressors/lz4/FramedLZ4CompressorInputStream.html (original) +++ websites/production/commons/content/proper/commons-compress/apidocs/src-html/org/apache/commons/compress/compressors/lz4/FramedLZ4CompressorInputStream.html Sat Feb 4 16:19:37 2017 @@ -28,263 +28,262 @@ <span class="sourceLineNo">020</span><a name="line.20"></a> <span class="sourceLineNo">021</span>import java.io.IOException;<a name="line.21"></a> <span class="sourceLineNo">022</span>import java.io.InputStream;<a name="line.22"></a> -<span class="sourceLineNo">023</span>import java.io.PushbackInputStream;<a name="line.23"></a> -<span class="sourceLineNo">024</span>import java.util.Arrays;<a name="line.24"></a> -<span class="sourceLineNo">025</span><a name="line.25"></a> -<span class="sourceLineNo">026</span>import org.apache.commons.compress.compressors.CompressorInputStream;<a name="line.26"></a> -<span class="sourceLineNo">027</span>import org.apache.commons.compress.utils.BoundedInputStream;<a name="line.27"></a> -<span class="sourceLineNo">028</span>import org.apache.commons.compress.utils.ByteUtils;<a name="line.28"></a> -<span class="sourceLineNo">029</span>import org.apache.commons.compress.utils.IOUtils;<a name="line.29"></a> -<span class="sourceLineNo">030</span><a name="line.30"></a> -<span class="sourceLineNo">031</span>/**<a name="line.31"></a> -<span class="sourceLineNo">032</span> * CompressorInputStream for the LZ4 frame format.<a name="line.32"></a> -<span class="sourceLineNo">033</span> *<a name="line.33"></a> -<span class="sourceLineNo">034</span> * <p>Based on the "spec" in the version "1.5.1 (31/03/2015)"</p><a name="line.34"></a> -<span class="sourceLineNo">035</span> *<a name="line.35"></a> -<span class="sourceLineNo">036</span> * @see <a href="http://lz4.github.io/lz4/lz4_Frame_format.html">LZ4 Frame Format Description</a><a name="line.36"></a> -<span class="sourceLineNo">037</span> * @since 1.14<a name="line.37"></a> -<span class="sourceLineNo">038</span> * @NotThreadSafe<a name="line.38"></a> -<span class="sourceLineNo">039</span> */<a name="line.39"></a> -<span class="sourceLineNo">040</span>public class FramedLZ4CompressorInputStream extends CompressorInputStream {<a name="line.40"></a> -<span class="sourceLineNo">041</span> /*<a name="line.41"></a> -<span class="sourceLineNo">042</span> * TODO before releasing 1.14:<a name="line.42"></a> -<span class="sourceLineNo">043</span> *<a name="line.43"></a> -<span class="sourceLineNo">044</span> * + xxhash32 checksum validation<a name="line.44"></a> -<span class="sourceLineNo">045</span> * + skippable frames<a name="line.45"></a> -<span class="sourceLineNo">046</span> * + decompressConcatenated<a name="line.46"></a> -<span class="sourceLineNo">047</span> * + block dependence<a name="line.47"></a> -<span class="sourceLineNo">048</span> */<a name="line.48"></a> -<span class="sourceLineNo">049</span><a name="line.49"></a> -<span class="sourceLineNo">050</span> // used by FramedLZ4CompressorOutputStream as well<a name="line.50"></a> -<span class="sourceLineNo">051</span> static final byte[] LZ4_SIGNATURE = new byte[] { //NOSONAR<a name="line.51"></a> -<span class="sourceLineNo">052</span> 4, 0x22, 0x4d, 0x18<a name="line.52"></a> -<span class="sourceLineNo">053</span> };<a name="line.53"></a> -<span class="sourceLineNo">054</span><a name="line.54"></a> -<span class="sourceLineNo">055</span> static final int VERSION_MASK = 0xC0;<a name="line.55"></a> -<span class="sourceLineNo">056</span> static final int SUPPORTED_VERSION = 0x40;<a name="line.56"></a> -<span class="sourceLineNo">057</span> static final int BLOCK_INDEPENDENCE_MASK = 0x20;<a name="line.57"></a> -<span class="sourceLineNo">058</span> static final int BLOCK_CHECKSUM_MASK = 0x10;<a name="line.58"></a> -<span class="sourceLineNo">059</span> static final int CONTENT_SIZE_MASK = 0x08;<a name="line.59"></a> -<span class="sourceLineNo">060</span> static final int CONTENT_CHECKSUM_MASK = 0x04;<a name="line.60"></a> -<span class="sourceLineNo">061</span> static final int BLOCK_MAX_SIZE_MASK = 0x70;<a name="line.61"></a> -<span class="sourceLineNo">062</span> static final int UNCOMPRESSED_FLAG_MASK = 0x80000000;<a name="line.62"></a> -<span class="sourceLineNo">063</span><a name="line.63"></a> -<span class="sourceLineNo">064</span> // used in no-arg read method<a name="line.64"></a> -<span class="sourceLineNo">065</span> private final byte[] oneByte = new byte[1];<a name="line.65"></a> -<span class="sourceLineNo">066</span><a name="line.66"></a> -<span class="sourceLineNo">067</span> private final ByteUtils.ByteSupplier supplier = new ByteUtils.ByteSupplier() {<a name="line.67"></a> -<span class="sourceLineNo">068</span> @Override<a name="line.68"></a> -<span class="sourceLineNo">069</span> public int getAsByte() throws IOException {<a name="line.69"></a> -<span class="sourceLineNo">070</span> return readOneByte();<a name="line.70"></a> -<span class="sourceLineNo">071</span> }<a name="line.71"></a> -<span class="sourceLineNo">072</span> };<a name="line.72"></a> -<span class="sourceLineNo">073</span><a name="line.73"></a> -<span class="sourceLineNo">074</span> private final InputStream in;<a name="line.74"></a> -<span class="sourceLineNo">075</span><a name="line.75"></a> -<span class="sourceLineNo">076</span> private boolean expectBlockChecksum;<a name="line.76"></a> -<span class="sourceLineNo">077</span> private boolean expectContentSize;<a name="line.77"></a> -<span class="sourceLineNo">078</span> private boolean expectContentChecksum;<a name="line.78"></a> -<span class="sourceLineNo">079</span><a name="line.79"></a> -<span class="sourceLineNo">080</span> private InputStream currentBlock;<a name="line.80"></a> -<span class="sourceLineNo">081</span> private boolean endReached, inUncompressed;<a name="line.81"></a> -<span class="sourceLineNo">082</span><a name="line.82"></a> -<span class="sourceLineNo">083</span> // used for frame header checksum and content checksum, if present<a name="line.83"></a> -<span class="sourceLineNo">084</span> private final XXHash32 contentHash = new XXHash32();<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> * Creates a new input stream that decompresses streams compressed<a name="line.87"></a> -<span class="sourceLineNo">088</span> * using the LZ4 frame format.<a name="line.88"></a> -<span class="sourceLineNo">089</span> * @param in the InputStream from which to read the compressed data<a name="line.89"></a> -<span class="sourceLineNo">090</span> * @throws IOException if reading fails<a name="line.90"></a> -<span class="sourceLineNo">091</span> */<a name="line.91"></a> -<span class="sourceLineNo">092</span> public FramedLZ4CompressorInputStream(InputStream in) throws IOException {<a name="line.92"></a> -<span class="sourceLineNo">093</span> this.in = in;<a name="line.93"></a> -<span class="sourceLineNo">094</span> readSignature();<a name="line.94"></a> -<span class="sourceLineNo">095</span> readFrameDescriptor();<a name="line.95"></a> -<span class="sourceLineNo">096</span> nextBlock();<a name="line.96"></a> -<span class="sourceLineNo">097</span> }<a name="line.97"></a> -<span class="sourceLineNo">098</span><a name="line.98"></a> -<span class="sourceLineNo">099</span> /** {@inheritDoc} */<a name="line.99"></a> -<span class="sourceLineNo">100</span> @Override<a name="line.100"></a> -<span class="sourceLineNo">101</span> public int read() throws IOException {<a name="line.101"></a> -<span class="sourceLineNo">102</span> return read(oneByte, 0, 1) == -1 ? -1 : oneByte[0] & 0xFF;<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> /** {@inheritDoc} */<a name="line.105"></a> -<span class="sourceLineNo">106</span> @Override<a name="line.106"></a> -<span class="sourceLineNo">107</span> public void close() throws IOException {<a name="line.107"></a> -<span class="sourceLineNo">108</span> if (currentBlock != null) {<a name="line.108"></a> -<span class="sourceLineNo">109</span> currentBlock.close();<a name="line.109"></a> -<span class="sourceLineNo">110</span> currentBlock = null;<a name="line.110"></a> -<span class="sourceLineNo">111</span> }<a name="line.111"></a> -<span class="sourceLineNo">112</span> in.close();<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> /** {@inheritDoc} */<a name="line.115"></a> -<span class="sourceLineNo">116</span> @Override<a name="line.116"></a> -<span class="sourceLineNo">117</span> public int read(final byte[] b, final int off, final int len) throws IOException {<a name="line.117"></a> -<span class="sourceLineNo">118</span> if (endReached) {<a name="line.118"></a> -<span class="sourceLineNo">119</span> return -1;<a name="line.119"></a> -<span class="sourceLineNo">120</span> }<a name="line.120"></a> -<span class="sourceLineNo">121</span> int r = readOnce(b, off, len);<a name="line.121"></a> -<span class="sourceLineNo">122</span> if (r == -1) {<a name="line.122"></a> -<span class="sourceLineNo">123</span> nextBlock();<a name="line.123"></a> -<span class="sourceLineNo">124</span> if (!endReached) {<a name="line.124"></a> -<span class="sourceLineNo">125</span> r = readOnce(b, off, len);<a name="line.125"></a> -<span class="sourceLineNo">126</span> }<a name="line.126"></a> -<span class="sourceLineNo">127</span> }<a name="line.127"></a> -<span class="sourceLineNo">128</span> if (expectContentChecksum && r != -1) {<a name="line.128"></a> -<span class="sourceLineNo">129</span> contentHash.update(b, off, r);<a name="line.129"></a> -<span class="sourceLineNo">130</span> }<a name="line.130"></a> -<span class="sourceLineNo">131</span> return r;<a name="line.131"></a> -<span class="sourceLineNo">132</span> }<a name="line.132"></a> -<span class="sourceLineNo">133</span><a name="line.133"></a> -<span class="sourceLineNo">134</span> private void readSignature() throws IOException {<a name="line.134"></a> -<span class="sourceLineNo">135</span> final byte[] b = new byte[4];<a name="line.135"></a> -<span class="sourceLineNo">136</span> final int read = IOUtils.readFully(in, b);<a name="line.136"></a> -<span class="sourceLineNo">137</span> count(read);<a name="line.137"></a> -<span class="sourceLineNo">138</span> if (4 != read || !matches(b, 4)) {<a name="line.138"></a> -<span class="sourceLineNo">139</span> throw new IOException("Not a LZ4 frame stream");<a name="line.139"></a> -<span class="sourceLineNo">140</span> }<a name="line.140"></a> -<span class="sourceLineNo">141</span> }<a name="line.141"></a> -<span class="sourceLineNo">142</span><a name="line.142"></a> -<span class="sourceLineNo">143</span> private void readFrameDescriptor() throws IOException {<a name="line.143"></a> -<span class="sourceLineNo">144</span> int flags = readOneByte();<a name="line.144"></a> -<span class="sourceLineNo">145</span> if (flags == -1) {<a name="line.145"></a> -<span class="sourceLineNo">146</span> throw new IOException("Premature end of stream while reading frame flags");<a name="line.146"></a> -<span class="sourceLineNo">147</span> }<a name="line.147"></a> -<span class="sourceLineNo">148</span> contentHash.update(flags);<a name="line.148"></a> -<span class="sourceLineNo">149</span> if ((flags & VERSION_MASK) != SUPPORTED_VERSION) {<a name="line.149"></a> -<span class="sourceLineNo">150</span> throw new IOException("Unsupported version " + (flags >> 6));<a name="line.150"></a> -<span class="sourceLineNo">151</span> }<a name="line.151"></a> -<span class="sourceLineNo">152</span> if ((flags & BLOCK_INDEPENDENCE_MASK) == 0) {<a name="line.152"></a> -<span class="sourceLineNo">153</span> throw new IOException("Block dependence is not supported");<a name="line.153"></a> -<span class="sourceLineNo">154</span> }<a name="line.154"></a> -<span class="sourceLineNo">155</span> expectBlockChecksum = (flags & BLOCK_CHECKSUM_MASK) != 0;<a name="line.155"></a> -<span class="sourceLineNo">156</span> expectContentSize = (flags & CONTENT_SIZE_MASK) != 0;<a name="line.156"></a> -<span class="sourceLineNo">157</span> expectContentChecksum = (flags & CONTENT_CHECKSUM_MASK) != 0;<a name="line.157"></a> -<span class="sourceLineNo">158</span> int bdByte = readOneByte();<a name="line.158"></a> -<span class="sourceLineNo">159</span> if (bdByte == -1) { // max size is irrelevant for this implementation<a name="line.159"></a> -<span class="sourceLineNo">160</span> throw new IOException("Premature end of stream while reading frame BD byte");<a name="line.160"></a> -<span class="sourceLineNo">161</span> }<a name="line.161"></a> -<span class="sourceLineNo">162</span> contentHash.update(bdByte);<a name="line.162"></a> -<span class="sourceLineNo">163</span> if (expectContentSize) { // for now we don't care, contains the uncompressed size<a name="line.163"></a> -<span class="sourceLineNo">164</span> byte[] contentSize = new byte[8];<a name="line.164"></a> -<span class="sourceLineNo">165</span> int skipped = (int) IOUtils.readFully(in, contentSize);<a name="line.165"></a> -<span class="sourceLineNo">166</span> count(skipped);<a name="line.166"></a> -<span class="sourceLineNo">167</span> if (8 != skipped) {<a name="line.167"></a> -<span class="sourceLineNo">168</span> throw new IOException("Premature end of stream while reading content size");<a name="line.168"></a> -<span class="sourceLineNo">169</span> }<a name="line.169"></a> -<span class="sourceLineNo">170</span> contentHash.update(contentSize, 0, contentSize.length);<a name="line.170"></a> -<span class="sourceLineNo">171</span> }<a name="line.171"></a> -<span class="sourceLineNo">172</span> int headerHash = readOneByte();<a name="line.172"></a> -<span class="sourceLineNo">173</span> if (headerHash == -1) { // partial hash of header.<a name="line.173"></a> -<span class="sourceLineNo">174</span> throw new IOException("Premature end of stream while reading frame header checksum");<a name="line.174"></a> -<span class="sourceLineNo">175</span> }<a name="line.175"></a> -<span class="sourceLineNo">176</span> int expectedHash = (int) ((contentHash.getValue() >> 8) & 0xff);<a name="line.176"></a> -<span class="sourceLineNo">177</span> contentHash.reset();<a name="line.177"></a> -<span class="sourceLineNo">178</span> if (headerHash != expectedHash) {<a name="line.178"></a> -<span class="sourceLineNo">179</span> throw new IOException("frame header checksum mismatch.");<a name="line.179"></a> -<span class="sourceLineNo">180</span> }<a name="line.180"></a> -<span class="sourceLineNo">181</span> }<a name="line.181"></a> -<span class="sourceLineNo">182</span><a name="line.182"></a> -<span class="sourceLineNo">183</span> private void nextBlock() throws IOException {<a name="line.183"></a> -<span class="sourceLineNo">184</span> maybeFinishCurrentBlock();<a name="line.184"></a> -<span class="sourceLineNo">185</span> long len = ByteUtils.fromLittleEndian(supplier, 4);<a name="line.185"></a> -<span class="sourceLineNo">186</span> boolean uncompressed = (len & UNCOMPRESSED_FLAG_MASK) != 0;<a name="line.186"></a> -<span class="sourceLineNo">187</span> int realLen = (int) (len & (~UNCOMPRESSED_FLAG_MASK));<a name="line.187"></a> -<span class="sourceLineNo">188</span> if (realLen == 0) {<a name="line.188"></a> -<span class="sourceLineNo">189</span> endReached = true;<a name="line.189"></a> -<span class="sourceLineNo">190</span> verifyContentChecksum();<a name="line.190"></a> -<span class="sourceLineNo">191</span> return;<a name="line.191"></a> -<span class="sourceLineNo">192</span> }<a name="line.192"></a> -<span class="sourceLineNo">193</span> InputStream capped = new BoundedInputStream(in, realLen);<a name="line.193"></a> -<span class="sourceLineNo">194</span> if (uncompressed) {<a name="line.194"></a> -<span class="sourceLineNo">195</span> inUncompressed = true;<a name="line.195"></a> -<span class="sourceLineNo">196</span> currentBlock = capped;<a name="line.196"></a> -<span class="sourceLineNo">197</span> } else {<a name="line.197"></a> -<span class="sourceLineNo">198</span> inUncompressed = false;<a name="line.198"></a> -<span class="sourceLineNo">199</span> currentBlock = new BlockLZ4CompressorInputStream(capped);<a name="line.199"></a> -<span class="sourceLineNo">200</span> }<a name="line.200"></a> -<span class="sourceLineNo">201</span> }<a name="line.201"></a> -<span class="sourceLineNo">202</span><a name="line.202"></a> -<span class="sourceLineNo">203</span> private void maybeFinishCurrentBlock() throws IOException {<a name="line.203"></a> -<span class="sourceLineNo">204</span> if (currentBlock != null) {<a name="line.204"></a> -<span class="sourceLineNo">205</span> currentBlock.close();<a name="line.205"></a> -<span class="sourceLineNo">206</span> currentBlock = null;<a name="line.206"></a> -<span class="sourceLineNo">207</span> if (expectBlockChecksum) {<a name="line.207"></a> -<span class="sourceLineNo">208</span> int skipped = (int) IOUtils.skip(in, 4);<a name="line.208"></a> -<span class="sourceLineNo">209</span> count(skipped);<a name="line.209"></a> -<span class="sourceLineNo">210</span> if (4 != skipped) {<a name="line.210"></a> -<span class="sourceLineNo">211</span> throw new IOException("Premature end of stream while reading block checksum");<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> }<a name="line.215"></a> -<span class="sourceLineNo">216</span><a name="line.216"></a> -<span class="sourceLineNo">217</span> private void verifyContentChecksum() throws IOException {<a name="line.217"></a> -<span class="sourceLineNo">218</span> if (expectContentChecksum) {<a name="line.218"></a> -<span class="sourceLineNo">219</span> byte[] checksum = new byte[4];<a name="line.219"></a> -<span class="sourceLineNo">220</span> int read = IOUtils.readFully(in, checksum);<a name="line.220"></a> -<span class="sourceLineNo">221</span> count(read);<a name="line.221"></a> -<span class="sourceLineNo">222</span> if (4 != read) {<a name="line.222"></a> -<span class="sourceLineNo">223</span> throw new IOException("Premature end of stream while reading content checksum");<a name="line.223"></a> -<span class="sourceLineNo">224</span> }<a name="line.224"></a> -<span class="sourceLineNo">225</span> long expectedHash = contentHash.getValue();<a name="line.225"></a> -<span class="sourceLineNo">226</span> if (expectedHash != ByteUtils.fromLittleEndian(checksum)) {<a name="line.226"></a> -<span class="sourceLineNo">227</span> throw new IOException("content checksum mismatch.");<a name="line.227"></a> -<span class="sourceLineNo">228</span> }<a name="line.228"></a> -<span class="sourceLineNo">229</span> contentHash.reset();<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><a name="line.232"></a> -<span class="sourceLineNo">233</span> private int readOneByte() throws IOException {<a name="line.233"></a> -<span class="sourceLineNo">234</span> final int b = in.read();<a name="line.234"></a> -<span class="sourceLineNo">235</span> if (b != -1) {<a name="line.235"></a> -<span class="sourceLineNo">236</span> count(1);<a name="line.236"></a> -<span class="sourceLineNo">237</span> return b & 0xFF;<a name="line.237"></a> -<span class="sourceLineNo">238</span> }<a name="line.238"></a> -<span class="sourceLineNo">239</span> return -1;<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> private int readOnce(byte[] b, int off, int len) throws IOException {<a name="line.242"></a> -<span class="sourceLineNo">243</span> if (inUncompressed) {<a name="line.243"></a> -<span class="sourceLineNo">244</span> int cnt = currentBlock.read(b, off, len);<a name="line.244"></a> -<span class="sourceLineNo">245</span> count(cnt);<a name="line.245"></a> -<span class="sourceLineNo">246</span> return cnt;<a name="line.246"></a> -<span class="sourceLineNo">247</span> } else {<a name="line.247"></a> -<span class="sourceLineNo">248</span> BlockLZ4CompressorInputStream l = (BlockLZ4CompressorInputStream) currentBlock;<a name="line.248"></a> -<span class="sourceLineNo">249</span> long before = l.getBytesRead();<a name="line.249"></a> -<span class="sourceLineNo">250</span> int cnt = currentBlock.read(b, off, len);<a name="line.250"></a> -<span class="sourceLineNo">251</span> count(l.getBytesRead() - before);<a name="line.251"></a> -<span class="sourceLineNo">252</span> return cnt;<a name="line.252"></a> -<span class="sourceLineNo">253</span> }<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> * Checks if the signature matches what is expected for a .lz4 file.<a name="line.257"></a> -<span class="sourceLineNo">258</span> *<a name="line.258"></a> -<span class="sourceLineNo">259</span> * <p>.lz4 files start with a four byte signature.</p><a name="line.259"></a> -<span class="sourceLineNo">260</span> *<a name="line.260"></a> -<span class="sourceLineNo">261</span> * @param signature the bytes to check<a name="line.261"></a> -<span class="sourceLineNo">262</span> * @param length the number of bytes to check<a name="line.262"></a> -<span class="sourceLineNo">263</span> * @return true if this is a .sz stream, false otherwise<a name="line.263"></a> -<span class="sourceLineNo">264</span> */<a name="line.264"></a> -<span class="sourceLineNo">265</span> public static boolean matches(final byte[] signature, final int length) {<a name="line.265"></a> -<span class="sourceLineNo">266</span><a name="line.266"></a> -<span class="sourceLineNo">267</span> if (length < LZ4_SIGNATURE.length) {<a name="line.267"></a> -<span class="sourceLineNo">268</span> return false;<a name="line.268"></a> -<span class="sourceLineNo">269</span> }<a name="line.269"></a> -<span class="sourceLineNo">270</span><a name="line.270"></a> -<span class="sourceLineNo">271</span> byte[] shortenedSig = signature;<a name="line.271"></a> -<span class="sourceLineNo">272</span> if (signature.length > LZ4_SIGNATURE.length) {<a name="line.272"></a> -<span class="sourceLineNo">273</span> shortenedSig = new byte[LZ4_SIGNATURE.length];<a name="line.273"></a> -<span class="sourceLineNo">274</span> System.arraycopy(signature, 0, shortenedSig, 0, LZ4_SIGNATURE.length);<a name="line.274"></a> -<span class="sourceLineNo">275</span> }<a name="line.275"></a> -<span class="sourceLineNo">276</span><a name="line.276"></a> -<span class="sourceLineNo">277</span> return Arrays.equals(shortenedSig, LZ4_SIGNATURE);<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">023</span>import java.util.Arrays;<a name="line.23"></a> +<span class="sourceLineNo">024</span><a name="line.24"></a> +<span class="sourceLineNo">025</span>import org.apache.commons.compress.compressors.CompressorInputStream;<a name="line.25"></a> +<span class="sourceLineNo">026</span>import org.apache.commons.compress.utils.BoundedInputStream;<a name="line.26"></a> +<span class="sourceLineNo">027</span>import org.apache.commons.compress.utils.ByteUtils;<a name="line.27"></a> +<span class="sourceLineNo">028</span>import org.apache.commons.compress.utils.IOUtils;<a name="line.28"></a> +<span class="sourceLineNo">029</span><a name="line.29"></a> +<span class="sourceLineNo">030</span>/**<a name="line.30"></a> +<span class="sourceLineNo">031</span> * CompressorInputStream for the LZ4 frame format.<a name="line.31"></a> +<span class="sourceLineNo">032</span> *<a name="line.32"></a> +<span class="sourceLineNo">033</span> * <p>Based on the "spec" in the version "1.5.1 (31/03/2015)"</p><a name="line.33"></a> +<span class="sourceLineNo">034</span> *<a name="line.34"></a> +<span class="sourceLineNo">035</span> * @see <a href="http://lz4.github.io/lz4/lz4_Frame_format.html">LZ4 Frame Format Description</a><a name="line.35"></a> +<span class="sourceLineNo">036</span> * @since 1.14<a name="line.36"></a> +<span class="sourceLineNo">037</span> * @NotThreadSafe<a name="line.37"></a> +<span class="sourceLineNo">038</span> */<a name="line.38"></a> +<span class="sourceLineNo">039</span>public class FramedLZ4CompressorInputStream extends CompressorInputStream {<a name="line.39"></a> +<span class="sourceLineNo">040</span> /*<a name="line.40"></a> +<span class="sourceLineNo">041</span> * TODO before releasing 1.14:<a name="line.41"></a> +<span class="sourceLineNo">042</span> *<a name="line.42"></a> +<span class="sourceLineNo">043</span> * + xxhash32 checksum validation<a name="line.43"></a> +<span class="sourceLineNo">044</span> * + skippable frames<a name="line.44"></a> +<span class="sourceLineNo">045</span> * + decompressConcatenated<a name="line.45"></a> +<span class="sourceLineNo">046</span> * + block dependence<a name="line.46"></a> +<span class="sourceLineNo">047</span> */<a name="line.47"></a> +<span class="sourceLineNo">048</span><a name="line.48"></a> +<span class="sourceLineNo">049</span> // used by FramedLZ4CompressorOutputStream as well<a name="line.49"></a> +<span class="sourceLineNo">050</span> static final byte[] LZ4_SIGNATURE = new byte[] { //NOSONAR<a name="line.50"></a> +<span class="sourceLineNo">051</span> 4, 0x22, 0x4d, 0x18<a name="line.51"></a> +<span class="sourceLineNo">052</span> };<a name="line.52"></a> +<span class="sourceLineNo">053</span><a name="line.53"></a> +<span class="sourceLineNo">054</span> static final int VERSION_MASK = 0xC0;<a name="line.54"></a> +<span class="sourceLineNo">055</span> static final int SUPPORTED_VERSION = 0x40;<a name="line.55"></a> +<span class="sourceLineNo">056</span> static final int BLOCK_INDEPENDENCE_MASK = 0x20;<a name="line.56"></a> +<span class="sourceLineNo">057</span> static final int BLOCK_CHECKSUM_MASK = 0x10;<a name="line.57"></a> +<span class="sourceLineNo">058</span> static final int CONTENT_SIZE_MASK = 0x08;<a name="line.58"></a> +<span class="sourceLineNo">059</span> static final int CONTENT_CHECKSUM_MASK = 0x04;<a name="line.59"></a> +<span class="sourceLineNo">060</span> static final int BLOCK_MAX_SIZE_MASK = 0x70;<a name="line.60"></a> +<span class="sourceLineNo">061</span> static final int UNCOMPRESSED_FLAG_MASK = 0x80000000;<a name="line.61"></a> +<span class="sourceLineNo">062</span><a name="line.62"></a> +<span class="sourceLineNo">063</span> // used in no-arg read method<a name="line.63"></a> +<span class="sourceLineNo">064</span> private final byte[] oneByte = new byte[1];<a name="line.64"></a> +<span class="sourceLineNo">065</span><a name="line.65"></a> +<span class="sourceLineNo">066</span> private final ByteUtils.ByteSupplier supplier = new ByteUtils.ByteSupplier() {<a name="line.66"></a> +<span class="sourceLineNo">067</span> @Override<a name="line.67"></a> +<span class="sourceLineNo">068</span> public int getAsByte() throws IOException {<a name="line.68"></a> +<span class="sourceLineNo">069</span> return readOneByte();<a name="line.69"></a> +<span class="sourceLineNo">070</span> }<a name="line.70"></a> +<span class="sourceLineNo">071</span> };<a name="line.71"></a> +<span class="sourceLineNo">072</span><a name="line.72"></a> +<span class="sourceLineNo">073</span> private final InputStream in;<a name="line.73"></a> +<span class="sourceLineNo">074</span><a name="line.74"></a> +<span class="sourceLineNo">075</span> private boolean expectBlockChecksum;<a name="line.75"></a> +<span class="sourceLineNo">076</span> private boolean expectContentSize;<a name="line.76"></a> +<span class="sourceLineNo">077</span> private boolean expectContentChecksum;<a name="line.77"></a> +<span class="sourceLineNo">078</span><a name="line.78"></a> +<span class="sourceLineNo">079</span> private InputStream currentBlock;<a name="line.79"></a> +<span class="sourceLineNo">080</span> private boolean endReached, inUncompressed;<a name="line.80"></a> +<span class="sourceLineNo">081</span><a name="line.81"></a> +<span class="sourceLineNo">082</span> // used for frame header checksum and content checksum, if present<a name="line.82"></a> +<span class="sourceLineNo">083</span> private final XXHash32 contentHash = new XXHash32();<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> * Creates a new input stream that decompresses streams compressed<a name="line.86"></a> +<span class="sourceLineNo">087</span> * using the LZ4 frame format.<a name="line.87"></a> +<span class="sourceLineNo">088</span> * @param in the InputStream from which to read the compressed data<a name="line.88"></a> +<span class="sourceLineNo">089</span> * @throws IOException if reading fails<a name="line.89"></a> +<span class="sourceLineNo">090</span> */<a name="line.90"></a> +<span class="sourceLineNo">091</span> public FramedLZ4CompressorInputStream(InputStream in) throws IOException {<a name="line.91"></a> +<span class="sourceLineNo">092</span> this.in = in;<a name="line.92"></a> +<span class="sourceLineNo">093</span> readSignature();<a name="line.93"></a> +<span class="sourceLineNo">094</span> readFrameDescriptor();<a name="line.94"></a> +<span class="sourceLineNo">095</span> nextBlock();<a name="line.95"></a> +<span class="sourceLineNo">096</span> }<a name="line.96"></a> +<span class="sourceLineNo">097</span><a name="line.97"></a> +<span class="sourceLineNo">098</span> /** {@inheritDoc} */<a name="line.98"></a> +<span class="sourceLineNo">099</span> @Override<a name="line.99"></a> +<span class="sourceLineNo">100</span> public int read() throws IOException {<a name="line.100"></a> +<span class="sourceLineNo">101</span> return read(oneByte, 0, 1) == -1 ? -1 : oneByte[0] & 0xFF;<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> /** {@inheritDoc} */<a name="line.104"></a> +<span class="sourceLineNo">105</span> @Override<a name="line.105"></a> +<span class="sourceLineNo">106</span> public void close() throws IOException {<a name="line.106"></a> +<span class="sourceLineNo">107</span> if (currentBlock != null) {<a name="line.107"></a> +<span class="sourceLineNo">108</span> currentBlock.close();<a name="line.108"></a> +<span class="sourceLineNo">109</span> currentBlock = null;<a name="line.109"></a> +<span class="sourceLineNo">110</span> }<a name="line.110"></a> +<span class="sourceLineNo">111</span> in.close();<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> /** {@inheritDoc} */<a name="line.114"></a> +<span class="sourceLineNo">115</span> @Override<a name="line.115"></a> +<span class="sourceLineNo">116</span> public int read(final byte[] b, final int off, final int len) throws IOException {<a name="line.116"></a> +<span class="sourceLineNo">117</span> if (endReached) {<a name="line.117"></a> +<span class="sourceLineNo">118</span> return -1;<a name="line.118"></a> +<span class="sourceLineNo">119</span> }<a name="line.119"></a> +<span class="sourceLineNo">120</span> int r = readOnce(b, off, len);<a name="line.120"></a> +<span class="sourceLineNo">121</span> if (r == -1) {<a name="line.121"></a> +<span class="sourceLineNo">122</span> nextBlock();<a name="line.122"></a> +<span class="sourceLineNo">123</span> if (!endReached) {<a name="line.123"></a> +<span class="sourceLineNo">124</span> r = readOnce(b, off, len);<a name="line.124"></a> +<span class="sourceLineNo">125</span> }<a name="line.125"></a> +<span class="sourceLineNo">126</span> }<a name="line.126"></a> +<span class="sourceLineNo">127</span> if (expectContentChecksum && r != -1) {<a name="line.127"></a> +<span class="sourceLineNo">128</span> contentHash.update(b, off, r);<a name="line.128"></a> +<span class="sourceLineNo">129</span> }<a name="line.129"></a> +<span class="sourceLineNo">130</span> return r;<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> private void readSignature() throws IOException {<a name="line.133"></a> +<span class="sourceLineNo">134</span> final byte[] b = new byte[4];<a name="line.134"></a> +<span class="sourceLineNo">135</span> final int read = IOUtils.readFully(in, b);<a name="line.135"></a> +<span class="sourceLineNo">136</span> count(read);<a name="line.136"></a> +<span class="sourceLineNo">137</span> if (4 != read || !matches(b, 4)) {<a name="line.137"></a> +<span class="sourceLineNo">138</span> throw new IOException("Not a LZ4 frame stream");<a name="line.138"></a> +<span class="sourceLineNo">139</span> }<a name="line.139"></a> +<span class="sourceLineNo">140</span> }<a name="line.140"></a> +<span class="sourceLineNo">141</span><a name="line.141"></a> +<span class="sourceLineNo">142</span> private void readFrameDescriptor() throws IOException {<a name="line.142"></a> +<span class="sourceLineNo">143</span> int flags = readOneByte();<a name="line.143"></a> +<span class="sourceLineNo">144</span> if (flags == -1) {<a name="line.144"></a> +<span class="sourceLineNo">145</span> throw new IOException("Premature end of stream while reading frame flags");<a name="line.145"></a> +<span class="sourceLineNo">146</span> }<a name="line.146"></a> +<span class="sourceLineNo">147</span> contentHash.update(flags);<a name="line.147"></a> +<span class="sourceLineNo">148</span> if ((flags & VERSION_MASK) != SUPPORTED_VERSION) {<a name="line.148"></a> +<span class="sourceLineNo">149</span> throw new IOException("Unsupported version " + (flags >> 6));<a name="line.149"></a> +<span class="sourceLineNo">150</span> }<a name="line.150"></a> +<span class="sourceLineNo">151</span> if ((flags & BLOCK_INDEPENDENCE_MASK) == 0) {<a name="line.151"></a> +<span class="sourceLineNo">152</span> throw new IOException("Block dependence is not supported");<a name="line.152"></a> +<span class="sourceLineNo">153</span> }<a name="line.153"></a> +<span class="sourceLineNo">154</span> expectBlockChecksum = (flags & BLOCK_CHECKSUM_MASK) != 0;<a name="line.154"></a> +<span class="sourceLineNo">155</span> expectContentSize = (flags & CONTENT_SIZE_MASK) != 0;<a name="line.155"></a> +<span class="sourceLineNo">156</span> expectContentChecksum = (flags & CONTENT_CHECKSUM_MASK) != 0;<a name="line.156"></a> +<span class="sourceLineNo">157</span> int bdByte = readOneByte();<a name="line.157"></a> +<span class="sourceLineNo">158</span> if (bdByte == -1) { // max size is irrelevant for this implementation<a name="line.158"></a> +<span class="sourceLineNo">159</span> throw new IOException("Premature end of stream while reading frame BD byte");<a name="line.159"></a> +<span class="sourceLineNo">160</span> }<a name="line.160"></a> +<span class="sourceLineNo">161</span> contentHash.update(bdByte);<a name="line.161"></a> +<span class="sourceLineNo">162</span> if (expectContentSize) { // for now we don't care, contains the uncompressed size<a name="line.162"></a> +<span class="sourceLineNo">163</span> byte[] contentSize = new byte[8];<a name="line.163"></a> +<span class="sourceLineNo">164</span> int skipped = (int) IOUtils.readFully(in, contentSize);<a name="line.164"></a> +<span class="sourceLineNo">165</span> count(skipped);<a name="line.165"></a> +<span class="sourceLineNo">166</span> if (8 != skipped) {<a name="line.166"></a> +<span class="sourceLineNo">167</span> throw new IOException("Premature end of stream while reading content size");<a name="line.167"></a> +<span class="sourceLineNo">168</span> }<a name="line.168"></a> +<span class="sourceLineNo">169</span> contentHash.update(contentSize, 0, contentSize.length);<a name="line.169"></a> +<span class="sourceLineNo">170</span> }<a name="line.170"></a> +<span class="sourceLineNo">171</span> int headerHash = readOneByte();<a name="line.171"></a> +<span class="sourceLineNo">172</span> if (headerHash == -1) { // partial hash of header.<a name="line.172"></a> +<span class="sourceLineNo">173</span> throw new IOException("Premature end of stream while reading frame header checksum");<a name="line.173"></a> +<span class="sourceLineNo">174</span> }<a name="line.174"></a> +<span class="sourceLineNo">175</span> int expectedHash = (int) ((contentHash.getValue() >> 8) & 0xff);<a name="line.175"></a> +<span class="sourceLineNo">176</span> contentHash.reset();<a name="line.176"></a> +<span class="sourceLineNo">177</span> if (headerHash != expectedHash) {<a name="line.177"></a> +<span class="sourceLineNo">178</span> throw new IOException("frame header checksum mismatch.");<a name="line.178"></a> +<span class="sourceLineNo">179</span> }<a name="line.179"></a> +<span class="sourceLineNo">180</span> }<a name="line.180"></a> +<span class="sourceLineNo">181</span><a name="line.181"></a> +<span class="sourceLineNo">182</span> private void nextBlock() throws IOException {<a name="line.182"></a> +<span class="sourceLineNo">183</span> maybeFinishCurrentBlock();<a name="line.183"></a> +<span class="sourceLineNo">184</span> long len = ByteUtils.fromLittleEndian(supplier, 4);<a name="line.184"></a> +<span class="sourceLineNo">185</span> boolean uncompressed = (len & UNCOMPRESSED_FLAG_MASK) != 0;<a name="line.185"></a> +<span class="sourceLineNo">186</span> int realLen = (int) (len & (~UNCOMPRESSED_FLAG_MASK));<a name="line.186"></a> +<span class="sourceLineNo">187</span> if (realLen == 0) {<a name="line.187"></a> +<span class="sourceLineNo">188</span> endReached = true;<a name="line.188"></a> +<span class="sourceLineNo">189</span> verifyContentChecksum();<a name="line.189"></a> +<span class="sourceLineNo">190</span> return;<a name="line.190"></a> +<span class="sourceLineNo">191</span> }<a name="line.191"></a> +<span class="sourceLineNo">192</span> InputStream capped = new BoundedInputStream(in, realLen);<a name="line.192"></a> +<span class="sourceLineNo">193</span> if (uncompressed) {<a name="line.193"></a> +<span class="sourceLineNo">194</span> inUncompressed = true;<a name="line.194"></a> +<span class="sourceLineNo">195</span> currentBlock = capped;<a name="line.195"></a> +<span class="sourceLineNo">196</span> } else {<a name="line.196"></a> +<span class="sourceLineNo">197</span> inUncompressed = false;<a name="line.197"></a> +<span class="sourceLineNo">198</span> currentBlock = new BlockLZ4CompressorInputStream(capped);<a name="line.198"></a> +<span class="sourceLineNo">199</span> }<a name="line.199"></a> +<span class="sourceLineNo">200</span> }<a name="line.200"></a> +<span class="sourceLineNo">201</span><a name="line.201"></a> +<span class="sourceLineNo">202</span> private void maybeFinishCurrentBlock() throws IOException {<a name="line.202"></a> +<span class="sourceLineNo">203</span> if (currentBlock != null) {<a name="line.203"></a> +<span class="sourceLineNo">204</span> currentBlock.close();<a name="line.204"></a> +<span class="sourceLineNo">205</span> currentBlock = null;<a name="line.205"></a> +<span class="sourceLineNo">206</span> if (expectBlockChecksum) {<a name="line.206"></a> +<span class="sourceLineNo">207</span> int skipped = (int) IOUtils.skip(in, 4);<a name="line.207"></a> +<span class="sourceLineNo">208</span> count(skipped);<a name="line.208"></a> +<span class="sourceLineNo">209</span> if (4 != skipped) {<a name="line.209"></a> +<span class="sourceLineNo">210</span> throw new IOException("Premature end of stream while reading block checksum");<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> }<a name="line.214"></a> +<span class="sourceLineNo">215</span><a name="line.215"></a> +<span class="sourceLineNo">216</span> private void verifyContentChecksum() throws IOException {<a name="line.216"></a> +<span class="sourceLineNo">217</span> if (expectContentChecksum) {<a name="line.217"></a> +<span class="sourceLineNo">218</span> byte[] checksum = new byte[4];<a name="line.218"></a> +<span class="sourceLineNo">219</span> int read = IOUtils.readFully(in, checksum);<a name="line.219"></a> +<span class="sourceLineNo">220</span> count(read);<a name="line.220"></a> +<span class="sourceLineNo">221</span> if (4 != read) {<a name="line.221"></a> +<span class="sourceLineNo">222</span> throw new IOException("Premature end of stream while reading content checksum");<a name="line.222"></a> +<span class="sourceLineNo">223</span> }<a name="line.223"></a> +<span class="sourceLineNo">224</span> long expectedHash = contentHash.getValue();<a name="line.224"></a> +<span class="sourceLineNo">225</span> if (expectedHash != ByteUtils.fromLittleEndian(checksum)) {<a name="line.225"></a> +<span class="sourceLineNo">226</span> throw new IOException("content checksum mismatch.");<a name="line.226"></a> +<span class="sourceLineNo">227</span> }<a name="line.227"></a> +<span class="sourceLineNo">228</span> contentHash.reset();<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><a name="line.231"></a> +<span class="sourceLineNo">232</span> private int readOneByte() throws IOException {<a name="line.232"></a> +<span class="sourceLineNo">233</span> final int b = in.read();<a name="line.233"></a> +<span class="sourceLineNo">234</span> if (b != -1) {<a name="line.234"></a> +<span class="sourceLineNo">235</span> count(1);<a name="line.235"></a> +<span class="sourceLineNo">236</span> return b & 0xFF;<a name="line.236"></a> +<span class="sourceLineNo">237</span> }<a name="line.237"></a> +<span class="sourceLineNo">238</span> return -1;<a name="line.238"></a> +<span class="sourceLineNo">239</span> }<a name="line.239"></a> +<span class="sourceLineNo">240</span><a name="line.240"></a> +<span class="sourceLineNo">241</span> private int readOnce(byte[] b, int off, int len) throws IOException {<a name="line.241"></a> +<span class="sourceLineNo">242</span> if (inUncompressed) {<a name="line.242"></a> +<span class="sourceLineNo">243</span> int cnt = currentBlock.read(b, off, len);<a name="line.243"></a> +<span class="sourceLineNo">244</span> count(cnt);<a name="line.244"></a> +<span class="sourceLineNo">245</span> return cnt;<a name="line.245"></a> +<span class="sourceLineNo">246</span> } else {<a name="line.246"></a> +<span class="sourceLineNo">247</span> BlockLZ4CompressorInputStream l = (BlockLZ4CompressorInputStream) currentBlock;<a name="line.247"></a> +<span class="sourceLineNo">248</span> long before = l.getBytesRead();<a name="line.248"></a> +<span class="sourceLineNo">249</span> int cnt = currentBlock.read(b, off, len);<a name="line.249"></a> +<span class="sourceLineNo">250</span> count(l.getBytesRead() - before);<a name="line.250"></a> +<span class="sourceLineNo">251</span> return cnt;<a name="line.251"></a> +<span class="sourceLineNo">252</span> }<a name="line.252"></a> +<span class="sourceLineNo">253</span> }<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> * Checks if the signature matches what is expected for a .lz4 file.<a name="line.256"></a> +<span class="sourceLineNo">257</span> *<a name="line.257"></a> +<span class="sourceLineNo">258</span> * <p>.lz4 files start with a four byte signature.</p><a name="line.258"></a> +<span class="sourceLineNo">259</span> *<a name="line.259"></a> +<span class="sourceLineNo">260</span> * @param signature the bytes to check<a name="line.260"></a> +<span class="sourceLineNo">261</span> * @param length the number of bytes to check<a name="line.261"></a> +<span class="sourceLineNo">262</span> * @return true if this is a .sz stream, false otherwise<a name="line.262"></a> +<span class="sourceLineNo">263</span> */<a name="line.263"></a> +<span class="sourceLineNo">264</span> public static boolean matches(final byte[] signature, final int length) {<a name="line.264"></a> +<span class="sourceLineNo">265</span><a name="line.265"></a> +<span class="sourceLineNo">266</span> if (length < LZ4_SIGNATURE.length) {<a name="line.266"></a> +<span class="sourceLineNo">267</span> return false;<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> byte[] shortenedSig = signature;<a name="line.270"></a> +<span class="sourceLineNo">271</span> if (signature.length > LZ4_SIGNATURE.length) {<a name="line.271"></a> +<span class="sourceLineNo">272</span> shortenedSig = new byte[LZ4_SIGNATURE.length];<a name="line.272"></a> +<span class="sourceLineNo">273</span> System.arraycopy(signature, 0, shortenedSig, 0, LZ4_SIGNATURE.length);<a name="line.273"></a> +<span class="sourceLineNo">274</span> }<a name="line.274"></a> +<span class="sourceLineNo">275</span><a name="line.275"></a> +<span class="sourceLineNo">276</span> return Arrays.equals(shortenedSig, LZ4_SIGNATURE);<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>