arcreader.html

用JAVA编写的,在做实验的时候留下来的,本来想删的,但是传上来,大家分享吧

<a name="137" href="#137">137</a>         <strong>if</strong> (getIn().available() &lt;= 0) {
<a name="138" href="#138">138</a>             <strong>return</strong>;
<a name="139" href="#139">139</a>         }
<a name="140" href="#140">140</a>         
<a name="141" href="#141">141</a>         <em class="comment">// Remove any trailing LINE_SEPARATOR</em>
<a name="142" href="#142">142</a>         <strong>int</strong> c = -1;
<a name="143" href="#143">143</a>         <strong>while</strong> (getIn().available() > 0) {
<a name="144" href="#144">144</a>             <strong>if</strong> (getIn().markSupported()) {
<a name="145" href="#145">145</a>                 getIn().mark(1);
<a name="146" href="#146">146</a>             }
<a name="147" href="#147">147</a>             c = getIn().read();
<a name="148" href="#148">148</a>             <strong>if</strong> (c != -1) {
<a name="149" href="#149">149</a>                 <strong>if</strong> (c == LINE_SEPARATOR) {
<a name="150" href="#150">150</a>                     <strong>continue</strong>;
<a name="151" href="#151">151</a>                 }
<a name="152" href="#152">152</a>                 <strong>if</strong> (getIn().markSupported()) {
<a name="153" href="#153">153</a>                     <em class="comment">// We've overread.  We're probably in next record.  There is</em>
<a name="154" href="#154">154</a>                     <em class="comment">// no way of telling for sure. It may be dross at end of</em>
<a name="155" href="#155">155</a>                     <em class="comment">// current record. Backup.</em>
<a name="156" href="#156">156</a>                 	getIn().reset();
<a name="157" href="#157">157</a>                     <strong>break</strong>;
<a name="158" href="#158">158</a>                 }
<a name="159" href="#159">159</a>                 <a href="../../../../org/archive/io/ArchiveRecordHeader.html">ArchiveRecordHeader</a> h = (getCurrentRecord() != <strong>null</strong>)?
<a name="160" href="#160">160</a>                     record.getHeader(): <strong>null</strong>;
<a name="161" href="#161">161</a>                 <strong>throw</strong> <strong>new</strong> IOException(<span class="string">"Read "</span> + (<strong>char</strong>)c +
<a name="162" href="#162">162</a>                     <span class="string">" when only "</span> + LINE_SEPARATOR + <span class="string">" expected. "</span> + 
<a name="163" href="#163">163</a>                     getReaderIdentifier() + ((h != <strong>null</strong>)?
<a name="164" href="#164">164</a>                         h.getHeaderFields().toString(): <span class="string">""</span>));
<a name="165" href="#165">165</a>             }
<a name="166" href="#166">166</a>         }
<a name="167" href="#167">167</a>     }
<a name="168" href="#168">168</a>     
<a name="169" href="#169">169</a>     <em>/**<em>*</em></em>
<a name="170" href="#170">170</a> <em>     * Create new arc record.</em>
<a name="171" href="#171">171</a> <em>     *</em>
<a name="172" href="#172">172</a> <em>     * Encapsulate housekeeping that has to do w/ creating a new record.</em>
<a name="173" href="#173">173</a> <em>     *</em>
<a name="174" href="#174">174</a> <em>     * &lt;p>Call this method at end of constructor to read in the</em>
<a name="175" href="#175">175</a> <em>     * arcfile header.  Will be problems reading subsequent arc records</em>
<a name="176" href="#176">176</a> <em>     * if you don't since arcfile header has the list of metadata fields for</em>
<a name="177" href="#177">177</a> <em>     * all records that follow.</em>
<a name="178" href="#178">178</a> <em>     * </em>
<a name="179" href="#179">179</a> <em>     * &lt;p>When parsing through ARCs writing out CDX info, we spend about</em>
<a name="180" href="#180">180</a> <em>     * 38% of CPU in here -- about 30% of which is in getTokenizedHeaderLine</em>
<a name="181" href="#181">181</a> <em>     * -- of which 16% is reading.</em>
<a name="182" href="#182">182</a> <em>     *</em>
<a name="183" href="#183">183</a> <em>     * @param is InputStream to use.</em>
<a name="184" href="#184">184</a> <em>     * @param offset Absolute offset into arc file.</em>
<a name="185" href="#185">185</a> <em>     * @return An arc record.</em>
<a name="186" href="#186">186</a> <em>     * @throws IOException</em>
<a name="187" href="#187">187</a> <em>     */</em>
<a name="188" href="#188">188</a>     <strong>protected</strong> <a href="../../../../org/archive/io/arc/ARCRecord.html">ARCRecord</a> createArchiveRecord(InputStream is, <strong>long</strong> offset)
<a name="189" href="#189">189</a>     throws IOException {
<a name="190" href="#190">190</a>         ArrayList&lt;String> firstLineValues = <strong>new</strong> ArrayList&lt;String>(20);
<a name="191" href="#191">191</a>         getTokenizedHeaderLine(is, firstLineValues);
<a name="192" href="#192">192</a>         <strong>int</strong> bodyOffset = 0;
<a name="193" href="#193">193</a>         <strong>if</strong> (offset == 0 &amp;&amp; isAlignedOnFirstRecord()) {
<a name="194" href="#194">194</a>             <em class="comment">// If offset is zero and we were aligned at first record on</em>
<a name="195" href="#195">195</a>             <em class="comment">// creation (See #alignedOnFirstRecord for more on this), then no</em>
<a name="196" href="#196">196</a>             <em class="comment">// records have been read yet and we're reading our first one, the</em>
<a name="197" href="#197">197</a>             <em class="comment">// record of ARC file meta info.  Its special.  In ARC versions</em>
<a name="198" href="#198">198</a>             <em class="comment">// 1.x, first record has three lines of meta info. We've just read</em>
<a name="199" href="#199">199</a>             <em class="comment">// the first line. There are two more.  The second line has misc.</em>
<a name="200" href="#200">200</a>             <em class="comment">// info.  We're only interested in the first field, the version</em>
<a name="201" href="#201">201</a>             <em class="comment">// number.  The third line is the list of field names. Here's what</em>
<a name="202" href="#202">202</a>             <em class="comment">// ARC file version 1.x meta content looks like:</em>
<a name="203" href="#203">203</a>             <em class="comment">//</em>
<a name="204" href="#204">204</a>             <em class="comment">// filedesc://testIsBoundary-JunitIAH200401070157520.arc 0.0.0.0 &#47;&#47;</em>
<a name="205" href="#205">205</a>             <em class="comment">//      20040107015752 text/plain 77</em>
<a name="206" href="#206">206</a>             <em class="comment">// 1 0 InternetArchive</em>
<a name="207" href="#207">207</a>             <em class="comment">// URL IP-address Archive-date Content-type Archive-length</em>
<a name="208" href="#208">208</a>             <em class="comment">//</em>
<a name="209" href="#209">209</a>             ArrayList&lt;String> secondLineValues = <strong>new</strong> ArrayList&lt;String>(20);
<a name="210" href="#210">210</a>             bodyOffset += getTokenizedHeaderLine(is, secondLineValues);
<a name="211" href="#211">211</a>             setVersion((String)secondLineValues.get(0) +
<a name="212" href="#212">212</a>                 <span class="string">"."</span> + (String)secondLineValues.get(1));
<a name="213" href="#213">213</a>             <em class="comment">// Just read over the 3rd line.  We used to parse it and use</em>
<a name="214" href="#214">214</a>             <em class="comment">// values found here but now we just hardcode them to avoid</em>
<a name="215" href="#215">215</a>             <em class="comment">// having to read this 3rd line even for random arc file accesses.</em>
<a name="216" href="#216">216</a>             bodyOffset += getTokenizedHeaderLine(is, <strong>null</strong>);
<a name="217" href="#217">217</a>         }
<a name="218" href="#218">218</a> 
<a name="219" href="#219">219</a>         <strong>try</strong> {
<a name="220" href="#220">220</a>             currentRecord(<strong>new</strong> <a href="../../../../org/archive/io/arc/ARCRecord.html">ARCRecord</a>(is,
<a name="221" href="#221">221</a>                 (<a href="../../../../org/archive/io/ArchiveRecordHeader.html">ArchiveRecordHeader</a>)computeMetaData(<strong>this</strong>.headerFieldNameKeys,
<a name="222" href="#222">222</a>                 	firstLineValues,
<a name="223" href="#223">223</a>                     getVersion(), offset), bodyOffset, isDigest(),
<a name="224" href="#224">224</a>                     isStrict(), isParseHttpHeaders()));
<a name="225" href="#225">225</a>         } <strong>catch</strong> (IOException e) {
<a name="226" href="#226">226</a>             IOException <strong>new</strong>E = <strong>new</strong> IOException(e.getMessage() + <span class="string">" (Offset "</span> +
<a name="227" href="#227">227</a>                     offset + <span class="string">")."</span>);
<a name="228" href="#228">228</a>             newE.setStackTrace(e.getStackTrace());
<a name="229" href="#229">229</a>             <strong>throw</strong> newE;
<a name="230" href="#230">230</a>         }
<a name="231" href="#231">231</a>         <strong>return</strong> (ARCRecord)getCurrentRecord();
<a name="232" href="#232">232</a>     }
<a name="233" href="#233">233</a>     
<a name="234" href="#234">234</a>     <em>/**<em>*</em></em>
<a name="235" href="#235">235</a> <em>     * Returns version of this ARC file.  Usually read from first record of ARC.</em>
<a name="236" href="#236">236</a> <em>     * If we're reading without having first read the first record -- e.g.</em>
<a name="237" href="#237">237</a> <em>     * random access into middle of an ARC -- then version will not have been</em>
<a name="238" href="#238">238</a> <em>     * set.  For now, we return a default, version 1.1.  Later, if more than</em>
<a name="239" href="#239">239</a> <em>     * just one version of ARC, we could look at such as the meta line to see</em>
<a name="240" href="#240">240</a> <em>     * what version of ARC this is.</em>
<a name="241" href="#241">241</a> <em>     * @return Version of this ARC file.</em>
<a name="242" href="#242">242</a> <em>     */</em>
<a name="243" href="#243">243</a>     <strong>public</strong> String getVersion() {
<a name="244" href="#244">244</a>         <strong>return</strong> (<strong>super</strong>.getVersion() == <strong>null</strong>)? <span class="string">"1.1"</span>: <strong>super</strong>.getVersion();
<a name="245" href="#245">245</a>     }
<a name="246" href="#246">246</a> 
<a name="247" href="#247">247</a>     <em>/**<em>*</em></em>
<a name="248" href="#248">248</a> <em>     * Get a record header line as list of tokens.</em>
<a name="249" href="#249">249</a> <em>     *</em>
<a name="250" href="#250">250</a> <em>     * We keep reading till we find a LINE_SEPARATOR or we reach the end</em>
<a name="251" href="#251">251</a> <em>     * of file w/o finding a LINE_SEPARATOR or the line length is crazy.</em>
<a name="252" href="#252">252</a> <em>     *</em>
<a name="253" href="#253">253</a> <em>     * @param stream InputStream to read from.</em>
<a name="254" href="#254">254</a> <em>     * @param list Empty list that gets filled w/ string tokens.</em>
<a name="255" href="#255">255</a> <em>     * @return Count of characters read.</em>
<a name="256" href="#256">256</a> <em>     * @exception IOException If problem reading stream or no line separator</em>
<a name="257" href="#257">257</a> <em>     * found or EOF before EOL or we didn't get minimum header fields.</em>
<a name="258" href="#258">258</a> <em>     */</em>
<a name="259" href="#259">259</a>     <strong>private</strong> <strong>int</strong> getTokenizedHeaderLine(<strong>final</strong> InputStream stream,
<a name="260" href="#260">260</a>             List&lt;String> list) throws IOException {
<a name="261" href="#261">261</a>         <em class="comment">// Preallocate usual line size.</em>
<a name="262" href="#262">262</a>         <em class="comment">// TODO: Replace StringBuffer with more lightweight.  We burn</em>
<a name="263" href="#263">263</a>         <em class="comment">// alot of our parse CPU in this method.</em>
<a name="264" href="#264">264</a>         StringBuffer buffer = <strong>new</strong> StringBuffer(2048 + 20);
<a name="265" href="#265">265</a>         <strong>int</strong> read = 0;
<a name="266" href="#266">266</a>         <strong>for</strong> (<strong>int</strong> c = -1; <strong>true</strong>;) {
<a name="267" href="#267">267</a>             c = stream.read();
<a name="268" href="#268">268</a>             <strong>if</strong> (c == -1) {
<a name="269" href="#269">269</a>                 <strong>throw</strong> <strong>new</strong> <a href="../../../../org/archive/io/RecoverableIOException.html">RecoverableIOException</a>(<span class="string">"Hit EOF before header EOL."</span>);
<a name="270" href="#270">270</a>             }
<a name="271" href="#271">271</a>             c &amp;= 0xff; 
<a name="272" href="#272">272</a>             read++;
<a name="273" href="#273">273</a>             <strong>if</strong> (read > MAX_HEADER_LINE_LENGTH) {
<a name="274" href="#274">274</a>                 <strong>throw</strong> <strong>new</strong> IOException(<span class="string">"Header line longer than max allowed "</span> +
<a name="275" href="#275">275</a>                     <span class="string">" -- "</span> + String.valueOf(MAX_HEADER_LINE_LENGTH) +
<a name="276" href="#276">276</a>                     <span class="string">" -- or passed buffer doesn't contain a line (Read: "</span> +
<a name="277" href="#277">277</a>                     buffer.length() + <span class="string">").  Here's"</span> +
<a name="278" href="#278">278</a>                     <span class="string">" some of what was read: "</span> +
<a name="279" href="#279">279</a>                     buffer.substring(0, Math.min(buffer.length(), 256)));
<a name="280" href="#280">280</a>             }
<a name="281" href="#281">281</a>