[clang-tools-extra] [clang-doc] Add standalone Markdown parsing library (PR #202991)

[Neil Nair via cfe-commits]

================
@@ -0,0 +1,304 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "support/Markdown.h"
+#include "llvm/Support/Allocator.h"
+#include "llvm/Support/Casting.h"
+#include "gtest/gtest.h"
+
+using namespace clang::doc::markdown;
+using namespace llvm;
+
+namespace {
+
+struct MarkdownParserTest : public ::testing::Test {
+  llvm::BumpPtrAllocator Arena;
+};
+
+TEST_F(MarkdownParserTest, EmptyInput) {
+  auto Nodes = parseMarkdown("", Arena);
+  EXPECT_TRUE(Nodes.empty());
+}
+
+TEST_F(MarkdownParserTest, WhitespaceOnlyInput) {
+  auto Nodes = parseMarkdown("   \n  \n", Arena);
+  EXPECT_TRUE(Nodes.empty());
+}
+
+TEST_F(MarkdownParserTest, PlainText) {
+  auto Nodes = parseMarkdown("hello world", Arena);
+  ASSERT_EQ(Nodes.size(), 1u);
+  auto *N = cast<TextNode>(Nodes[0]);
+  EXPECT_EQ(N->Text, "hello world");
+}
+
+TEST_F(MarkdownParserTest, FencedCodeBlock) {
+  auto Nodes = parseMarkdown(R"(```cpp
+int x = 0;
+````````)",
+                             Arena);
+  ASSERT_EQ(Nodes.size(), 1u);
+  auto *N = cast<FencedCodeNode>(Nodes[0]);
+  EXPECT_EQ(N->Lang, "cpp");
+  ASSERT_EQ(N->Lines.size(), 1u);
+}
+
+TEST_F(MarkdownParserTest, FencedCodeBlockNoLang) {
+  auto Nodes = parseMarkdown(R"(```
+some code
+```````)",
+                             Arena);
+  ASSERT_EQ(Nodes.size(), 1u);
+  auto *N = cast<FencedCodeNode>(Nodes[0]);
+  EXPECT_TRUE(N->Lang.empty());
+}
+
+TEST_F(MarkdownParserTest, UnterminatedFenceReturnsEmpty) {
+  auto Nodes = parseMarkdown(R"(```cpp
+int x = 0;)",
+                             Arena);
+  // Unterminated fence should not crash and should produce a code node
+  // with whatever lines were found.
+  EXPECT_FALSE(Nodes.empty());
+}
+
+TEST_F(MarkdownParserTest, PipeTable) {
+  auto Nodes = parseMarkdown(R"(| A | B |
+|---|---|
+| 1 | 2 |)",
+                             Arena);
+  ASSERT_EQ(Nodes.size(), 1u);
+  EXPECT_TRUE(isa<TableNode>(Nodes[0]));
+}
+
+TEST_F(MarkdownParserTest, PipeCharacterWithoutSepRowIsPlainText) {
+  auto Nodes = parseMarkdown(R"(a | b
+c | d)",
+                             Arena);
+  // No separator row so should not be parsed as a table.
+  for (const auto *Node : Nodes)
+    EXPECT_FALSE(isa<TableNode>(Node));
+}
+
+TEST_F(MarkdownParserTest, UnorderedList) {
+  auto Nodes = parseMarkdown(R"(- foo
+- bar
+- baz)",
+                             Arena);
+  ASSERT_EQ(Nodes.size(), 1u);
+  auto *N = cast<UnorderedListNode>(Nodes[0]);
+  ASSERT_EQ(N->Items.size(), 3u);
+  EXPECT_EQ(cast<TextNode>(N->Items[0]->Children[0])->Text, "foo");
+  EXPECT_EQ(cast<TextNode>(N->Items[1]->Children[0])->Text, "bar");
+  EXPECT_EQ(cast<TextNode>(N->Items[2]->Children[0])->Text, "baz");
+}
+
+TEST_F(MarkdownParserTest, MixedContent) {
+  auto Nodes = parseMarkdown(R"(some text
+```````
+code
+````````
+- item)",
+                             Arena);
+  EXPECT_EQ(Nodes.size(), 3u);
----------------
Neil-N4 wrote:

Will add node type checks for each of the 3 nodes

https://github.com/llvm/llvm-project/pull/202991