Add ada-url dependency, initial impl of jsg::Url

This will serve as the new underlying implementation of the
url-standard.{h|c++} class while also supporting handling of
module import specifiers as URLs.

.bazelrc

@@ -190,6 +190,7 @@ build:windows --per_file_copt='external/ssl/src/crypto/poly1305/poly1305_vec\.c@
 build:windows --per_file_copt='external/v8/src/objects/literal-objects\.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include' --host_per_file_copt='external/v8/src/objects/literal-objects.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include'
 build:windows --per_file_copt='external/v8/src/runtime/runtime-object\.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include' --host_per_file_copt='external/v8/src/runtime/runtime-object.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include'
 build:windows --per_file_copt='external/v8/src/objects/swiss-name-dictionary\.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include' --host_per_file_copt='external/v8/src/objects/swiss-name-dictionary.cc@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include'
+build:windows --per_file_copt='external/ada-url/ada\.cpp@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include' --host_per_file_copt='external/ada-url/ada\.cpp@-imsvcC:\\Program Files\\LLVM\\lib\\clang\\16.0.6\\include'
 
 # enable clang coverage: https://clang.llvm.org/docs/SourceBasedCodeCoverage.html
 build:clang-coverage --copt="-fprofile-instr-generate" --linkopt="-fprofile-instr-generate"

.github/secret_scanning.yml

@@ -1,3 +1,4 @@
 paths-ignore:
   - "src/workerd/api/node/crypto_keys-test.js"
   - "src/workerd/api/node/crypto_dh-test.js"
+  - "src/workerd/jsg/url-test-corpus-success.h"

WORKSPACE

@@ -76,6 +76,16 @@ http_archive(
     urls = ["https://github.com/google/brotli/tarball/ec107cf015139c791f79afac0f96c3a2c45e157f"],
 )
 
+http_archive(
+    name = "ada-url",
+    build_file = "//:build/BUILD.ada-url",
+    sha256 = "d6be6a559745a79be191bc63c1190015c702a30bacad10028d32b479644a0785",
+    type = "zip",
+    url = "https://github.com/ada-url/ada/releases/download/v2.7.0/singleheader.zip",
+    patches = [],
+    patch_args = ["-p1"],
+)
+
 # ========================================================================================
 # Dawn
 #

build/BUILD.ada-url

@@ -0,0 +1,12 @@
+
+cc_library(
+    name = "ada-url",
+    hdrs = ["ada.h", "ada_c.h"],
+    srcs = ["ada.cpp"],
+    visibility = ["//visibility:public"],
+    include_prefix = ".",
+    copts = ["-w"],
+    defines = [
+      "ADA_SSE2=1"
+    ]
+)

compile_flags.txt

@@ -4,9 +4,11 @@
 -nostdinc
 -Ibazel-bin/external/com_googlesource_chromium_base_trace_event_common/_virtual_includes/trace_event_common
 -Ibazel-bin/external/dawn/include
+-Ibazel-bin/external/ada-url/_virtual_includes/ada-url/
 -Ibazel-bin/external/com_cloudflare_lol_html/_virtual_includes/lolhtml
 -Iexternal/com_google_benchmark/include/
 -Iexternal/dawn/include
+-Iexternal/ada-url/
 -Isrc
 -isystem/usr/include
 -isystem/usr/include/x86_64-linux-gnu

src/workerd/jsg/BUILD.bazel

@@ -26,6 +26,7 @@ wd_cc_library(
         ":exception",
         ":modules_capnp",
         ":observer",
+        ":url",
         "//src/workerd/util",
         "//src/workerd/util:sentry",
         "//src/workerd/util:thread-scopes",
@@ -34,6 +35,17 @@ wd_cc_library(
     ],
 )
 
+wd_cc_library(
+    name = "url",
+    srcs = ["url.c++"],
+    hdrs = ["url.h"],
+    visibility = ["//visibility:public"],
+    deps = [
+        "@capnp-cpp//src/kj",
+        "@ada-url",
+    ],
+)
+
 wd_cc_library(
     name = "exception",
     srcs = ["exception.c++"],

src/workerd/jsg/jsg.h

@@ -2333,3 +2333,4 @@ inline v8::Local<v8::Context> JsContext<T>::getHandle(Lock& js) {
 #include "function.h"
 #include "iterator.h"
 #include "jsvalue.h"
+#include "url.h"

src/workerd/jsg/url-test.c++

@@ -0,0 +1,104 @@
+// Copyright (c) 2017-2022 Cloudflare, Inc.
+// Licensed under the Apache 2.0 license found in the LICENSE file or at:
+//     https://opensource.org/licenses/Apache-2.0
+
+#include "jsg-test.h"
+#include "url.h"
+#include <kj/table.h>
+
+namespace workerd::jsg::test {
+namespace {
+
+KJ_TEST("Basics") {
+  Url theUrl = nullptr;
+  KJ_IF_SOME(url, Url::tryParse("http://example.org:81"_kj)) {
+    KJ_ASSERT(url.getOrigin() == "http://example.org:81"_kj);
+    KJ_ASSERT(url.getHref() == "http://example.org:81/"_kj);
+    KJ_ASSERT(url.getProtocol() == "http:"_kj);
+    KJ_ASSERT(url.getHostname() == "example.org"_kj);
+    KJ_ASSERT(url.getHost() == "example.org:81"_kj);
+    KJ_ASSERT(url.getPort() == "81"_kj);
+    KJ_ASSERT(url.getPathname() == "/"_kj);
+    KJ_ASSERT(url.getSchemeType() == Url::SchemeType::HTTP);
+    KJ_ASSERT(url.getHostType() == Url::HostType::DEFAULT);
+    KJ_ASSERT(url.getUsername() == ""_kj);
+    KJ_ASSERT(url.getPassword() == ""_kj);
+    KJ_ASSERT(url.getHash() == ""_kj);
+    KJ_ASSERT(url.getSearch() == ""_kj);
+
+    theUrl = url.clone();
+    KJ_ASSERT(theUrl == url);
+    theUrl = kj::mv(url);
+
+    auto res = KJ_ASSERT_NONNULL(theUrl.resolve("abc"_kj));
+    KJ_ASSERT(res.getHref() == "http://example.org:81/abc"_kj);
+
+    // jsg::Urls support KJ_STRINGIFY
+    KJ_ASSERT(kj::str(res) == "http://example.org:81/abc");
+
+    // jsg::Urls are suitable to be used as keys in a hashset, hashmap
+    kj::HashSet<Url> urls;
+    urls.insert(res.clone());
+    KJ_ASSERT(urls.contains(res));
+
+    kj::HashMap<Url, int> urlmap;
+    urlmap.insert(res.clone(), 1);
+    KJ_ASSERT(KJ_ASSERT_NONNULL(urlmap.find(res)) == 1);
+  } else {
+    KJ_FAIL_ASSERT("url could not be parsed");
+  }
+
+  KJ_ASSERT(Url::idnToAscii("täst.de"_kj) == "xn--tst-qla.de"_kj);
+  KJ_ASSERT(Url::idnToUnicode("xn--tst-qla.de"_kj) == "täst.de"_kj);
+}
+
+KJ_TEST("Non-special URL") {
+  auto url = KJ_ASSERT_NONNULL(Url::tryParse("abc://123"_kj));
+  KJ_ASSERT(url.getOrigin() == "null"_kj);
+  KJ_ASSERT(url.getProtocol() == "abc:"_kj);
+}
+
+KJ_TEST("Invalid Urls") {
+#include "url-test-corpus-failures.h"
+}
+
+void test(kj::StringPtr input,
+          kj::Maybe<kj::StringPtr> base,
+          kj::StringPtr href) {
+  KJ_ASSERT(Url::canParse(input, base));
+  auto url = KJ_ASSERT_NONNULL(Url::tryParse(input, base));
+  KJ_ASSERT(url.getHref() == href);
+}
+
+KJ_TEST("Valid Urls") {
+  KJ_ASSERT_NONNULL(Url::tryParse(""_kj, "http://example.org"_kj));
+#include "url-test-corpus-success.h"
+}
+
+KJ_TEST("Search params (1)") {
+  UrlSearchParams params;
+  params.append("foo"_kj, "bar"_kj);
+  KJ_ASSERT(params.toStr() == "foo=bar"_kj);
+}
+
+KJ_TEST("Search params (2)") {
+  auto params = KJ_ASSERT_NONNULL(UrlSearchParams::tryParse("foo=bar&a=b&a=c"_kj));
+  KJ_ASSERT(params.has("a"_kj));
+  KJ_ASSERT(params.has("foo"_kj, kj::Maybe("bar"_kj)));
+  KJ_ASSERT(!params.has("foo"_kj, kj::Maybe("baz"_kj)));
+  KJ_ASSERT(KJ_ASSERT_NONNULL(params.get("a"_kj)) == "b"_kj);
+
+  auto all = params.getAll("a"_kj);
+  KJ_ASSERT(all.size() == 2);
+  KJ_ASSERT(all[0] == "b"_kj);
+  KJ_ASSERT(all[1] == "c"_kj);
+
+  params.delete_("foo"_kj);
+  params.delete_("a"_kj, kj::Maybe("c"_kj));
+
+  params.set("a"_kj, "z"_kj);
+  KJ_ASSERT(kj::str(params) == "a=z");
+}
+
+}  // namespace
+}  // namespace workerd::jsg::test