?? fetchdns.html
字號:
<a name="226" href="#226">226</a> <strong>if</strong> (logger.isLoggable(Level.WARNING)) {<a name="227" href="#227">227</a> logger.warning(<span class="string">"Unnecessary DNS CrawlURI created: "</span> + curi);<a name="228" href="#228">228</a> }<a name="229" href="#229">229</a> <strong>try</strong> {<a name="230" href="#230">230</a> targetHost.setIP(InetAddress.getByAddress(dnsName, <strong>new</strong> byte[] {<a name="231" href="#231">231</a> (byte) (<strong>new</strong> Integer(matcher.group(1)).intValue()),<a name="232" href="#232">232</a> (byte) (<strong>new</strong> Integer(matcher.group(2)).intValue()),<a name="233" href="#233">233</a> (byte) (<strong>new</strong> Integer(matcher.group(3)).intValue()),<a name="234" href="#234">234</a> (byte) (<strong>new</strong> Integer(matcher.group(4)).intValue()) }),<a name="235" href="#235">235</a> CrawlHost.IP_NEVER_EXPIRES); <em class="comment">// Never expire numeric IPs</em><a name="236" href="#236">236</a> curi.setFetchStatus(S_DNS_SUCCESS);<a name="237" href="#237">237</a> } <strong>catch</strong> (UnknownHostException e) {<a name="238" href="#238">238</a> logger.log(Level.SEVERE, <span class="string">"Should never be "</span> + e.getMessage(), e);<a name="239" href="#239">239</a> setUnresolvable(curi, targetHost);<a name="240" href="#240">240</a> }<a name="241" href="#241">241</a> <strong>return</strong> result;<a name="242" href="#242">242</a> }<a name="243" href="#243">243</a> <a name="244" href="#244">244</a> <strong>protected</strong> <strong>void</strong> recordDNS(<strong>final</strong> <a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi, <strong>final</strong> Record[] rrecordSet)<a name="245" href="#245">245</a> throws IOException {<a name="246" href="#246">246</a> <strong>final</strong> byte[] dnsRecord =<a name="247" href="#247">247</a> getDNSRecord(curi.getLong(A_FETCH_BEGAN_TIME), rrecordSet);<a name="248" href="#248">248</a> HttpRecorder rec = HttpRecorder.getHttpRecorder();<a name="249" href="#249">249</a> <a name="250" href="#250">250</a> <em class="comment">// Shall we get a digest on the content downloaded?</em><a name="251" href="#251">251</a> <strong>boolean</strong> digestContent = ((Boolean)getUncheckedAttribute(curi,<a name="252" href="#252">252</a> FetchHTTP.ATTR_DIGEST_CONTENT)).booleanValue();<a name="253" href="#253">253</a> String algorithm = <strong>null</strong>; <a name="254" href="#254">254</a> <strong>if</strong> (digestContent) {<a name="255" href="#255">255</a> algorithm = ((String)getUncheckedAttribute(curi,<a name="256" href="#256">256</a> FetchHTTP.ATTR_DIGEST_ALGORITHM));<a name="257" href="#257">257</a> rec.getRecordedInput().setDigest(algorithm);<a name="258" href="#258">258</a> } <strong>else</strong> {<a name="259" href="#259">259</a> <em class="comment">// clear</em><a name="260" href="#260">260</a> rec.getRecordedInput().setDigest((MessageDigest)<strong>null</strong>);<a name="261" href="#261">261</a> }<a name="262" href="#262">262</a> <a name="263" href="#263">263</a> curi.setHttpRecorder(rec);<a name="264" href="#264">264</a> InputStream is = curi.getHttpRecorder().inputWrap(<a name="265" href="#265">265</a> <strong>new</strong> ByteArrayInputStream(dnsRecord));<a name="266" href="#266">266</a> <strong>if</strong>(digestContent) {<a name="267" href="#267">267</a> rec.getRecordedInput().startDigest();<a name="268" href="#268">268</a> }<a name="269" href="#269">269</a> <em class="comment">// Reading from the wrapped stream, behind the scenes, will write</em><a name="270" href="#270">270</a> <em class="comment">// files into scratch space</em><a name="271" href="#271">271</a> <strong>try</strong> {<a name="272" href="#272">272</a> <strong>while</strong> (is.read(<strong>this</strong>.reusableBuffer) != -1) {<a name="273" href="#273">273</a> <strong>continue</strong>;<a name="274" href="#274">274</a> }<a name="275" href="#275">275</a> } <strong>finally</strong> {<a name="276" href="#276">276</a> is.close();<a name="277" href="#277">277</a> rec.closeRecorders();<a name="278" href="#278">278</a> }<a name="279" href="#279">279</a> curi.setContentSize(dnsRecord.length);<a name="280" href="#280">280</a> <strong>if</strong> (digestContent) {<a name="281" href="#281">281</a> curi.setContentDigest(algorithm,<a name="282" href="#282">282</a> rec.getRecordedInput().getDigestValue());<a name="283" href="#283">283</a> }<a name="284" href="#284">284</a> }<a name="285" href="#285">285</a> <a name="286" href="#286">286</a> <strong>protected</strong> byte [] getDNSRecord(<strong>final</strong> <strong>long</strong> fetchStart,<a name="287" href="#287">287</a> <strong>final</strong> Record[] rrecordSet)<a name="288" href="#288">288</a> throws IOException {<a name="289" href="#289">289</a> ByteArrayOutputStream baos = <strong>new</strong> ByteArrayOutputStream();<a name="290" href="#290">290</a> <em class="comment">// Start the record with a 14-digit date per RFC 2540</em><a name="291" href="#291">291</a> byte[] fetchDate = ArchiveUtils.get14DigitDate(fetchStart).getBytes();<a name="292" href="#292">292</a> baos.write(fetchDate);<a name="293" href="#293">293</a> <em class="comment">// Don't forget the newline</em><a name="294" href="#294">294</a> baos.write(<span class="string">"\n"</span>.getBytes());<a name="295" href="#295">295</a> <strong>int</strong> recordLength = fetchDate.length + 1;<a name="296" href="#296">296</a> <strong>if</strong> (rrecordSet != <strong>null</strong>) {<a name="297" href="#297">297</a> <strong>for</strong> (<strong>int</strong> i = 0; i < rrecordSet.length; i++) {<a name="298" href="#298">298</a> byte[] record = rrecordSet[i].toString().getBytes();<a name="299" href="#299">299</a> recordLength += record.length;<a name="300" href="#300">300</a> baos.write(record);<a name="301" href="#301">301</a> <em class="comment">// Add the newline between records back in</em><a name="302" href="#302">302</a> baos.write(<span class="string">"\n"</span>.getBytes());<a name="303" href="#303">303</a> recordLength += 1;<a name="304" href="#304">304</a> }<a name="305" href="#305">305</a> }<a name="306" href="#306">306</a> <strong>return</strong> baos.toByteArray();<a name="307" href="#307">307</a> }<a name="308" href="#308">308</a> <a name="309" href="#309">309</a> <strong>protected</strong> <strong>void</strong> setUnresolvable(<a href="../../../../org/archive/crawler/datamodel/CrawlURI.html">CrawlURI</a> curi, <a href="../../../../org/archive/crawler/datamodel/CrawlHost.html">CrawlHost</a> host) {<a name="310" href="#310">310</a> host.setIP(<strong>null</strong>, 0);<a name="311" href="#311">311</a> curi.setFetchStatus(S_DOMAIN_UNRESOLVABLE); <a name="312" href="#312">312</a> }<a name="313" href="#313">313</a> <a name="314" href="#314">314</a> <strong>protected</strong> ARecord getFirstARecord(Record[] rrecordSet) {<a name="315" href="#315">315</a> ARecord arecord = <strong>null</strong>;<a name="316" href="#316">316</a> <strong>if</strong> (rrecordSet == <strong>null</strong> || rrecordSet.length == 0) {<a name="317" href="#317">317</a> <strong>if</strong> (logger.isLoggable(Level.FINEST)) {<a name="318" href="#318">318</a> logger.finest(<span class="string">"rrecordSet is null or zero length: "</span> +<a name="319" href="#319">319</a> rrecordSet);<a name="320" href="#320">320</a> }<a name="321" href="#321">321</a> <strong>return</strong> arecord;<a name="322" href="#322">322</a> }<a name="323" href="#323">323</a> <strong>for</strong> (<strong>int</strong> i = 0; i < rrecordSet.length; i++) {<a name="324" href="#324">324</a> <strong>if</strong> (rrecordSet[i].getType() != Type.A) {<a name="325" href="#325">325</a> <strong>if</strong> (logger.isLoggable(Level.FINEST)) {<a name="326" href="#326">326</a> logger.finest(<span class="string">"Record "</span> + Integer.toString(i) +<a name="327" href="#327">327</a> <span class="string">" is not A type but "</span> + rrecordSet[i].getType());<a name="328" href="#328">328</a> }<a name="329" href="#329">329</a> <strong>continue</strong>;<a name="330" href="#330">330</a> }<a name="331" href="#331">331</a> arecord = (ARecord) rrecordSet[i];<a name="332" href="#332">332</a> <strong>break</strong>;<a name="333" href="#333">333</a> }<a name="334" href="#334">334</a> <strong>return</strong> arecord;<a name="335" href="#335">335</a> }<a name="336" href="#336">336</a> }</pre><hr/><div id="footer">This page was automatically generated by <a href="http://maven.apache.org/">Maven</a></div></body></html>
?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -