uurifactory.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="412" href="#412">412</a>         
<a name="413" href="#413">413</a>         <em class="comment">// Kill newlines etc</em>
<a name="414" href="#414">414</a>         uri = TextUtils.replaceAll(NEWLINE, uri, EMPTY_STRING);
<a name="415" href="#415">415</a>         
<a name="416" href="#416">416</a>         <em class="comment">// Test for the case of more than two slashes after the http(s) scheme.</em>
<a name="417" href="#417">417</a>         <em class="comment">// Replace with two slashes as mozilla does if found.</em>
<a name="418" href="#418">418</a>         <em class="comment">// See [ 788219 ] URI Syntax Errors stop page parsing.</em>
<a name="419" href="#419">419</a>         Matcher matcher = HTTP_SCHEME_SLASHES.matcher(uri);
<a name="420" href="#420">420</a>         <strong>if</strong> (matcher.matches()) {
<a name="421" href="#421">421</a>             uri = matcher.group(1) + matcher.group(2);
<a name="422" href="#422">422</a>         }
<a name="423" href="#423">423</a> 
<a name="424" href="#424">424</a>         <em class="comment">// now, minimally escape any whitespace</em>
<a name="425" href="#425">425</a>         uri = escapeWhitespace(uri);
<a name="426" href="#426">426</a>         
<a name="427" href="#427">427</a>         <em class="comment">// For further processing, get uri elements.  See the RFC2396REGEX</em>
<a name="428" href="#428">428</a>         <em class="comment">// comment above for explaination of group indices used in the below.</em>
<a name="429" href="#429">429</a>         matcher = RFC2396REGEX.matcher(uri);
<a name="430" href="#430">430</a>         <strong>if</strong> (!matcher.matches()) {
<a name="431" href="#431">431</a>             <strong>throw</strong> <strong>new</strong> URIException(<span class="string">"Failed parse of "</span> + uri);
<a name="432" href="#432">432</a>         }
<a name="433" href="#433">433</a>         String uriScheme = checkUriElementAndLowerCase(matcher.group(2));
<a name="434" href="#434">434</a>         String uriSchemeSpecificPart = checkUriElement(matcher.group(3));
<a name="435" href="#435">435</a>         String uriAuthority = checkUriElement(matcher.group(5));
<a name="436" href="#436">436</a>         String uriPath = checkUriElement(matcher.group(6));
<a name="437" href="#437">437</a>         String uriQuery = checkUriElement(matcher.group(8));
<a name="438" href="#438">438</a>         <em class="comment">// UNUSED String uriFragment = checkUriElement(matcher.group(10));</em>
<a name="439" href="#439">439</a>         
<a name="440" href="#440">440</a>         <em class="comment">// If a scheme, is it a supported scheme?</em>
<a name="441" href="#441">441</a>         <strong>if</strong> (uriScheme != <strong>null</strong> &amp;&amp; uriScheme.length() > 0 &amp;&amp;
<a name="442" href="#442">442</a>                 <strong>this</strong>.schemes != <strong>null</strong>) {
<a name="443" href="#443">443</a>             <strong>if</strong> (!(Arrays.binarySearch(schemes,uriScheme)>=0)) {
<a name="444" href="#444">444</a>                 <em class="comment">// unsupported; see if silently ignored</em>
<a name="445" href="#445">445</a>                 <strong>if</strong>((Arrays.binarySearch(ignoredSchemes,uriScheme)>=0)) {
<a name="446" href="#446">446</a>                     <strong>throw</strong> <strong>new</strong> URIException(
<a name="447" href="#447">447</a>                             IGNORED_SCHEME, <span class="string">"Ignored scheme: "</span> + uriScheme);
<a name="448" href="#448">448</a>                 } <strong>else</strong> {
<a name="449" href="#449">449</a>                     <strong>throw</strong> <strong>new</strong> URIException(<span class="string">"Unsupported scheme: "</span> + uriScheme);
<a name="450" href="#450">450</a>                 }
<a name="451" href="#451">451</a>             }
<a name="452" href="#452">452</a>         }
<a name="453" href="#453">453</a>         
<a name="454" href="#454">454</a>         <em class="comment">// Test if relative URI. If so, need a base to resolve against.</em>
<a name="455" href="#455">455</a>         <strong>if</strong> (uriScheme == <strong>null</strong> || uriScheme.length() &lt;= 0) {
<a name="456" href="#456">456</a>             <strong>if</strong> (base == <strong>null</strong>) {
<a name="457" href="#457">457</a>                 <strong>throw</strong> <strong>new</strong> URIException(<span class="string">"Relative URI but no base: "</span> + uri);
<a name="458" href="#458">458</a>             }
<a name="459" href="#459">459</a>         }
<a name="460" href="#460">460</a>         
<a name="461" href="#461">461</a>         <em class="comment">// fixup authority portion: lowercase/IDN-punycode any domain; </em>
<a name="462" href="#462">462</a>         <em class="comment">// remove stray trailing spaces</em>
<a name="463" href="#463">463</a>         uriAuthority = fixupAuthority(uriAuthority);
<a name="464" href="#464">464</a> 
<a name="465" href="#465">465</a>         <em class="comment">// Do some checks if absolute path.</em>
<a name="466" href="#466">466</a>         <strong>if</strong> (uriSchemeSpec<strong>if</strong>icPart != <strong>null</strong> &amp;&amp;
<a name="467" href="#467">467</a>                 uriSchemeSpecificPart.startsWith(SLASH)) {
<a name="468" href="#468">468</a>             <strong>if</strong> (uriPath != <strong>null</strong>) {
<a name="469" href="#469">469</a>                 <em class="comment">// Eliminate '..' if its first thing in the path.  IE does this.</em>
<a name="470" href="#470">470</a>                 uriPath = TextUtils.replaceFirst(SLASHDOTDOTSLASH, uriPath,
<a name="471" href="#471">471</a>                     SLASH);
<a name="472" href="#472">472</a>             }
<a name="473" href="#473">473</a>             <em class="comment">// Ensure root URLs end with '/': browsers always send "/"</em>
<a name="474" href="#474">474</a>             <em class="comment">// on the request-line, so we should consider "http://host"</em>
<a name="475" href="#475">475</a>             <em class="comment">// to be "http://host/".</em>
<a name="476" href="#476">476</a>             <strong>if</strong> (uriPath == <strong>null</strong> || EMPTY_STRING.equals(uriPath)) {
<a name="477" href="#477">477</a>                 uriPath = SLASH;
<a name="478" href="#478">478</a>             }
<a name="479" href="#479">479</a>         }
<a name="480" href="#480">480</a> 
<a name="481" href="#481">481</a>         <strong>if</strong> (uriAuthority != <strong>null</strong>) {
<a name="482" href="#482">482</a>             <strong>if</strong> (uriScheme != <strong>null</strong> &amp;&amp; uriScheme.length() > 0 &amp;&amp;
<a name="483" href="#483">483</a>                     uriScheme.equals(HTTP)) {
<a name="484" href="#484">484</a>                 uriAuthority = checkPort(uriAuthority);
<a name="485" href="#485">485</a>                 uriAuthority = stripTail(uriAuthority, HTTP_PORT);
<a name="486" href="#486">486</a>             } <strong>else</strong> <strong>if</strong> (uriScheme != <strong>null</strong> &amp;&amp; uriScheme.length() > 0 &amp;&amp;
<a name="487" href="#487">487</a>                     uriScheme.equals(HTTPS)) {
<a name="488" href="#488">488</a>                 uriAuthority = checkPort(uriAuthority);
<a name="489" href="#489">489</a>                 uriAuthority = stripTail(uriAuthority, HTTPS_PORT);
<a name="490" href="#490">490</a>             }
<a name="491" href="#491">491</a>             <em class="comment">// Strip any prefix dot or tail dots from the authority.</em>
<a name="492" href="#492">492</a>             uriAuthority = stripTail(uriAuthority, DOT);
<a name="493" href="#493">493</a>             uriAuthority = stripPrefix(uriAuthority, DOT);
<a name="494" href="#494">494</a>         } <strong>else</strong> {
<a name="495" href="#495">495</a>             <em class="comment">// no authority; may be relative. consider stripping scheme</em>
<a name="496" href="#496">496</a>             <em class="comment">// to work-around org.apache.commons.httpclient.URI bug</em>
<a name="497" href="#497">497</a>             <em class="comment">// ( http://issues.apache.org/jira/browse/HTTPCLIENT-587 )</em>
<a name="498" href="#498">498</a>             <strong>if</strong> (uriScheme != <strong>null</strong> &amp;&amp; base != <strong>null</strong>
<a name="499" href="#499">499</a>                     &amp;&amp; uriScheme.equals(base.getScheme())) {
<a name="500" href="#500">500</a>                 <em class="comment">// uriScheme redundant and will only confound httpclient.URI</em>
<a name="501" href="#501">501</a>                 uriScheme = <strong>null</strong>; 
<a name="502" href="#502">502</a>             }
<a name="503" href="#503">503</a>         }
<a name="504" href="#504">504</a>         
<a name="505" href="#505">505</a>         <em class="comment">// Ensure minimal escaping. Use of 'lax' URI and URLCodec </em>
<a name="506" href="#506">506</a>         <em class="comment">// means minimal escaping isn't necessarily complete/consistent.</em>
<a name="507" href="#507">507</a>         <em class="comment">// There is a chance such lax encoding will throw exceptions</em>
<a name="508" href="#508">508</a>         <em class="comment">// later at inconvenient times. </em>
<a name="509" href="#509">509</a>         <em class="comment">//</em>
<a name="510" href="#510">510</a>         <em class="comment">// One reason for these bad escapings -- though not the only --</em>
<a name="511" href="#511">511</a>         <em class="comment">// is that the page is using an encoding other than the ASCII or the</em>
<a name="512" href="#512">512</a>         <em class="comment">// UTF-8 that is our default URI encoding.  In this case the parent</em>
<a name="513" href="#513">513</a>         <em class="comment">// class is burping on the passed URL encoding.  If the page encoding</em>
<a name="514" href="#514">514</a>         <em class="comment">// was passed into this factory, the encoding seems to be parsed</em>
<a name="515" href="#515">515</a>         <em class="comment">// correctly (See the testEscapedEncoding unit test).</em>
<a name="516" href="#516">516</a>         <em class="comment">//</em>
<a name="517" href="#517">517</a>         <em class="comment">// This fixup may cause us to miss content.  There is the charset case</em>
<a name="518" href="#518">518</a>         <em class="comment">// noted above.  TODO: Look out for cases where we fail other than for</em>
<a name="519" href="#519">519</a>         <em class="comment">// the above given reason which will be fixed when we address</em>
<a name="520" href="#520">520</a>         <em class="comment">// '[ 913687 ] Make extractors interrogate for charset'.</em>
<a name="521" href="#521">521</a> 
<a name="522" href="#522">522</a>         uriPath = ensureMinimalEscaping(uriPath, charset);
<a name="523" href="#523">523</a>         uriQuery = ensureMinimalEscaping(uriQuery, charset,
<a name="524" href="#524">524</a>             LaxURLCodec.QUERY_SAFE);
<a name="525" href="#525">525</a> 
<a name="526" href="#526">526</a>         <em class="comment">// Preallocate.  The '1's and '2's in below are space for ':',</em>
<a name="527" href="#527">527</a>         <em class="comment">// '//', etc. URI characters.</em>
<a name="528" href="#528">528</a>         MutableString s = <strong>new</strong> MutableString(
<a name="529" href="#529">529</a>             ((uriScheme != <strong>null</strong>)? uriScheme.length(): 0)
<a name="530" href="#530">530</a>             + 1 <em class="comment">// ';' </em>
<a name="531" href="#531">531</a>             + ((uriAuthority != <strong>null</strong>)? uriAuthority.length(): 0)
<a name="532" href="#532">532</a>             + 2 <em class="comment">// '//'</em>
<a name="533" href="#533">533</a>             + ((uriPath != <strong>null</strong>)? uriPath.length(): 0)
<a name="534" href="#534">534</a>             + 1 <em class="comment">// '?'</em>
<a name="535" href="#535">535</a>             + ((uriQuery != <strong>null</strong>)? uriQuery.length(): 0));
<a name="536" href="#536">536</a>         appendNonNull(s, uriScheme, <span class="string">":"</span>, <strong>true</strong>);
<a name="537" href="#537">537</a>         appendNonNull(s, uriAuthority, <span class="string">"//"</span>, false);
<a name="538" href="#538">538</a>         appendNonNull(s, uriPath, <span class="string">""</span>, false);
<a name="539" href="#539">539</a>         appendNonNull(s, uriQuery, <span class="string">"?"</span>, false);
<a name="540" href="#540">540</a>         <strong>return</strong> s.toString();
<a name="541" href="#541">541</a>     }
<a name="542" href="#542">542</a>     
<a name="543" href="#543">543</a>     <em>/**<em>*</em></em>
<a name="544" href="#544">544</a> <em>     * Fixup 'authority' portion of URI, by removing any stray </em>
<a name="545" href="#545">545</a> <em>     * encoded spaces, lowercasing any domain names, and applying</em>
<a name="546" href="#546">546</a> <em>     * IDN-punycoding to Unicode domains. </em>
<a name="547" href="#547">547</a> <em>     * </em>
<a name="548" href="#548">548</a> <em>     * @param uriAuthority the authority string to fix</em>
<a name="549" href="#549">549</a> <em>     * @return fixed version</em>
<a name="550" href="#550">550</a> <em>     * @throws URIException</em>
<a name="551" href="#551">551</a> <em>     */</em>