New, more validating, parser for URI.

tests/corelib/uri_test.dart

 // Copyright (c) 2012, the Dart project authors.  Please see the AUTHORS file
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 
 library uriTest;
 
 import "package:expect/expect.dart";
 import 'dart:convert';
 
 testUri(String uri, bool isAbsolute) {
@@ skipping 106 matching lines @@
                       base.resolve("../g;p/h;s").toString());
 }
 
 void testResolvePath(String expected, String path) {
   Expect.equals(expected, new Uri().resolveUri(new Uri(path: path)).path);
   Expect.equals(
       "http://localhost$expected",
       Uri.parse("http://localhost").resolveUri(new Uri(path: path)).toString());
 }
 
+void testValidCharacters() {
+  const ALPHA = r"abcdefghijklmnopqrstuvwxuzABCDEFGHIJKLMNOPQRSTUVWXUZ";
+  const DIGIT = r"0123456789";
+  const PERCENT_ENCODED = "%00%ff";
+  const SUBDELIM = r"!$&'()*+,;=";
+
+  const SCHEMECHAR = "$ALPHA$DIGIT+-.";
+  const UNRESERVED = "$ALPHA$DIGIT-._~";
+  const REGNAMECHAR = "$UNRESERVED$SUBDELIM$PERCENT_ENCODED";
+  const USERINFOCHAR = "$REGNAMECHAR:";
+
+  const PCHAR_NC = "$UNRESERVED$SUBDELIM$PERCENT_ENCODED@";
+  const PCHAR = "$PCHAR_NC:";
+  const QUERYCHAR = "$PCHAR/?";
+
+  void checkInvalid(uri) {
+    try {
+      var result = Uri.parse(uri);
+      Expect.fail("Invalid URI `$uri` parsed to $result\n"
+                  "  Scheme:    ${result.scheme}\n"
+                  "  User-info: ${result.userInfo}\n"
+                  "  Host:      ${result.host}\n"
+                  "  Port:      ${result.port}\n"
+                  "  Path:      ${result.path}\n"
+                  "  Query:     ${result.query}\n"
+                  "  Fragment:  ${result.fragment}\n");
+    } on FormatException {
+      // Success.
+    }
+  }
+
+  // test that all valid characters are accepted.

[kevmoo, 2014/06/10 21:29:08]

split valid test into its own test method?

+
+  for (var scheme in ["", "$SCHEMECHAR$SCHEMECHAR:"]) {
+    for (var userinfo in ["", "@", "$USERINFOCHAR$USERINFOCHAR@",
+                          "$USERINFOCHAR:$DIGIT@"]) {
+      for (var host in ["", "$REGNAMECHAR$REGNAMECHAR",
+                        "255.255.255.256",  // valid reg-name.
+                        "[ffff::ffff:ffff]", "[ffff::255.255.255.255]"]) {
+        for (var port in ["", ":$DIGIT$DIGIT"]) {
+          var auth = "$userinfo$host$port";
+          if (auth.isNotEmpty) auth = "//$auth";
+          var paths = ["/", "/$PCHAR", "/$PCHAR/"];
+          if (auth.isNotEmpty) {
+            // Initial segemnt may be empty.
+            paths.add("//$PCHAR");
+          } else {
+            // Path may begin with non-slash.
+            paths.add("");
+            if (scheme.isEmpty) {
+              // Initial segment must not contain colon.
+              paths..add("$PCHAR_NC/$PCHAR")..add("$PCHAR_NC/$PCHAR/");
+            } else {
+              paths..add("$PCHAR")..add("$PCHAR/");
+            }
+          }
+          for (var path in paths) {
+            for (var query in ["", "?", "?$QUERYCHAR"]) {
+              for (var fragment in ["", "#", "#$QUERYCHAR"]) {
+                var uri = "$scheme$auth$path$query$fragment";
+                // Should not throw.
+                var result = Uri.parse(uri);
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+
+  // Invalid characters. The characters must be rejected, even if normalizing
+  // the input would cause them to be valid (normalization happens after
+  // validation).
+  var invalidChars = [
+    "\xe7",      // Arbitrary non-ASCII letter
+    " ",         // Space, not allowed anywhere.
+    '"',         // Quote, not allowed anywhere
+    "\x7f",      // DEL, not allowed anywhere
+    "%1g",       // Invalid escape.
+    "\xdf",      // German lower-case scharf-S. Becomes ASCII when upper-cased.
+    "\u0130"     // Latin capital dotted I, becomes ASCII lower-case in Turkish.
+    "%\uFB03",   // % + Ligature ffi, becomes ASCII when upper-cased,
+                 // should not be read as "%FFI".
+    "\u212a",    // Kelvin sign. Becomes ASCII when lower-cased.
+  ];
+  for (var invalid in invalidChars) {
+    checkInvalid("$SCHEMECHAR${invalid}$SCHEMECHAR:///");
+    checkInvalid("${invalid}$SCHEMECHAR:///");
+    checkInvalid("s://user${invalid}info@x.x/");
+    checkInvalid("s://reg${invalid}name/");
+    checkInvalid("s://regname:12${invalid}45/");
+    checkInvalid("s://regname/p${invalid}ath/");
+    checkInvalid("/p${invalid}ath/");
+    checkInvalid("p${invalid}ath/");
+    checkInvalid("s://regname/path/?x${invalid}x");
+    checkInvalid("s://regname/path/#x${invalid}x");
+    checkInvalid("s://regname/path/?#x${invalid}x");
+  }
+  checkInvalid("s%41://x.x/");      // No escapes in scheme,
+                                    // and no colon before slash in path.
+  checkInvalid("1a://x.x/");        // Scheme must start with letter,
+                                    // and no colon before slash in path.
+  checkInvalid(".a://x.x/");        // Scheme must start with letter,
+                                    // and no colon before slash in path.
+  checkInvalid(":");                // Scheme must start with letter,
+                                    // and no colon before slash in path.
+  checkInvalid("s://x@x@x.x/");     // At most one @ in userinfo.
+  checkInvalid("s://x@x:x/");       // No colon in host except before a port.
+  checkInvalid("s://x@x:9:9/");     // At most one port.
+  checkInvalid("s://x/x#foo#bar");  // At most one #.
+  checkInvalid("s://:/");           // Colon in host requires port,
+                                    // and path may not start with //.
+  checkInvalid("s@://x:9/x?x#x");   // @ not allowed in scheme.
+}

[kevmoo, 2014/06/10 21:32:04]

Include examples provided in the bug -
https://code.google.com/p/dart/issues/detail?id=16081

Uri.parse("http://www.example.org/red%09ros\u00E9#red");
Uri.parse("http://r\u00E9sum\u00E9.example.org");

Just for sanity and to give a nod to the bug reporter

+
 main() {
   testUri("http:", true);
   testUri("file://", true);
   testUri("file", false);
   testUri("http://user@example.com:8080/fisk?query=89&hest=silas", true);
   testUri("http://user@example.com:8080/fisk?query=89&hest=silas#fragment",
           false);
   Expect.stringEquals("http://user@example.com/a/b/c?query#fragment",
                       new Uri(
                           scheme: "http",
                           userInfo: "user",
                           host: "example.com",
                           port: 80,
                           path: "/a/b/c",
                           query: "query",
                           fragment: "fragment").toString());
   Expect.stringEquals("//null@null/a/b/c/",
                       new Uri(
                           scheme: null,
                           userInfo: null,
                           host: null,
                           port: 0,
                           path: "/a/b/c/",
                           query: null,
                           fragment: null).toString());
   Expect.stringEquals("file://", Uri.parse("file:").toString());
-
   testResolvePath("/a/g", "/a/b/c/./../../g");
   testResolvePath("/a/g", "/a/b/c/./../../g");
   testResolvePath("/mid/6", "mid/content=5/../6");
   testResolvePath("/a/b/e", "a/b/c/d/../../e");
   testResolvePath("/a/b/e", "../a/b/c/d/../../e");
   testResolvePath("/a/b/e", "./a/b/c/d/../../e");
   testResolvePath("/a/b/e", "../a/b/./c/d/../../e");
   testResolvePath("/a/b/e", "./a/b/./c/d/../../e");
   testResolvePath("/a/b/e/", "./a/b/./c/d/../../e/.");
   testResolvePath("/a/b/e/", "./a/b/./c/d/../../e/./.");
@@ skipping 90 matching lines @@
   testEncodeDecodeComponent(":/@',;?&=+\$", "%3A%2F%40'%2C%3B%3F%26%3D%2B%24");
   testEncodeDecodeComponent(s, "%F0%90%80%80");
   testEncodeDecodeQueryComponent("A + B", "A+%2B+B", "A+%2B+B", "A+%2B+B");
   testEncodeDecodeQueryComponent(
       "æ ø å", "%C3%A6+%C3%B8+%C3%A5", "%E6+%F8+%E5", null);
 
   // Invalid URI - : and @ is swapped, port ("host") should be numeric.
   Expect.throws(
       () => Uri.parse("file://user@password:host/path"),
       (e) => e is FormatException);
+
+  testValidCharacters();
 }