LibCpp: Add regression tests for the parser

For each .cpp file in the test suite data, there is a .ast file that
represents the "known good" baseline of the parser result.

Each .cpp file goes through the parser, and the result of
invoking `ASTNode::dump()` on the root node is compared to the
baseline to find regressions.

We also check that there were no parser errors when parsing the .cpp
files.

Tests/LibCpp/test-cpp-parser.cpp

@@ -0,0 +1,80 @@
+/*
+ * Copyright (c) 2021, Itamar S. <itamar8910@gmail.com>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <AK/LexicalPath.h>
+#include <LibCore/DirIterator.h>
+#include <LibCore/File.h>
+#include <LibCpp/Parser.h>
+#include <LibTest/TestCase.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+
+constexpr char TESTS_ROOT_DIR[] = "/home/anon/cpp-tests/parser";
+
+static String read_all(const String& path)
+{
+    auto result = Core::File::open(path, Core::OpenMode::ReadOnly);
+    VERIFY(!result.is_error());
+    auto content = result.value()->read_all();
+    return { reinterpret_cast<const char*>(content.data()), content.size() };
+}
+
+TEST_CASE(test_regression)
+{
+    Core::DirIterator directory_iterator(TESTS_ROOT_DIR, Core::DirIterator::Flags::SkipDots);
+
+    while (directory_iterator.has_next()) {
+        auto file_path = directory_iterator.next_full_path();
+
+        auto path = LexicalPath { file_path };
+        if (!path.has_extension(".cpp"))
+            continue;
+
+        outln("Checking {}...", path.basename());
+
+        auto ast_file_path = String::formatted("{}.ast", file_path.substring(0, file_path.length() - sizeof(".cpp") + 1));
+
+        auto source = read_all(file_path);
+        auto target_ast = read_all(ast_file_path);
+
+        StringView source_view(source);
+        ::Cpp::Parser parser(source_view, file_path);
+        auto root = parser.parse();
+
+        EXPECT(parser.errors().is_empty());
+
+        int pipefd[2] = {};
+        if (pipe(pipefd) < 0) {
+            perror("pipe");
+            exit(1);
+        }
+
+        FILE* input_stream = fdopen(pipefd[0], "r");
+        FILE* output_stream = fdopen(pipefd[1], "w");
+
+        root->dump(output_stream);
+
+        fclose(output_stream);
+
+        ByteBuffer buffer;
+        while (!feof(input_stream)) {
+            char chunk[4096];
+            size_t size = fread(chunk, sizeof(char), sizeof(chunk), input_stream);
+            if (size == 0)
+                break;
+            buffer.append(chunk, size);
+        }
+
+        fclose(input_stream);
+
+        String content { reinterpret_cast<const char*>(buffer.data()), buffer.size() };
+
+        auto equal = content == target_ast;
+        EXPECT(equal);
+    }
+}