[clang-tools-extra] [clangd] Improve Markup Rendering (PR #140498)

[Maksim Ivanov via cfe-commits]

================
@@ -370,40 +497,128 @@ std::unique_ptr<Block> Paragraph::clone() const {
 
 /// Choose a marker to delimit `Text` from a prioritized list of options.
 /// This is more readable than escaping for plain-text.
-llvm::StringRef chooseMarker(llvm::ArrayRef<llvm::StringRef> Options,
-                             llvm::StringRef Text) {
+llvm::StringRef Paragraph::chooseMarker(llvm::ArrayRef<llvm::StringRef> Options,
+                                        llvm::StringRef Text) const {
   // Prefer a delimiter whose characters don't appear in the text.
   for (llvm::StringRef S : Options)
     if (Text.find_first_of(S) == llvm::StringRef::npos)
       return S;
   return Options.front();
 }
 
+bool Paragraph::punctuationIndicatesLineBreak(llvm::StringRef Line) const {
+  constexpr llvm::StringLiteral Punctuation = R"txt(.:,;!?)txt";
+
+  Line = Line.rtrim();
+  return !Line.empty() && Punctuation.contains(Line.back());
+}
+
+bool Paragraph::isHardLineBreakIndicator(llvm::StringRef Rest) const {
+  // '-'/'*' md list, '@'/'\' documentation command, '>' md blockquote,
+  // '#' headings, '`' code blocks, two spaces (markdown force newline)
+  constexpr llvm::StringLiteral LinebreakIndicators = R"txt(-*@\>#`)txt";
+
+  Rest = Rest.ltrim(" \t");
+  if (Rest.empty())
+    return false;
+
+  if (LinebreakIndicators.contains(Rest.front()))
+    return true;
+
+  if (llvm::isDigit(Rest.front())) {
+    llvm::StringRef AfterDigit = Rest.drop_while(llvm::isDigit);
+    if (AfterDigit.starts_with(".") || AfterDigit.starts_with(")"))
+      return true;
+  }
+  return false;
+}
+
+bool Paragraph::isHardLineBreakAfter(llvm::StringRef Line,
+                                     llvm::StringRef Rest) const {
+  // In Markdown, 2 spaces before a line break forces a line break.
+  // Add a line break for plaintext in this case too.
+  // Should we also consider whether Line is short?
+  return Line.ends_with("  ") || punctuationIndicatesLineBreak(Line) ||
+         isHardLineBreakIndicator(Rest);
+}
+
 void Paragraph::renderPlainText(llvm::raw_ostream &OS) const {
   bool NeedsSpace = false;
+  std::string ConcatenatedText;
----------------
emaxx-google wrote:

Thinking about the extra overhead potentially introduced here, maybe we should at least look at avoiding most of reallocations if we `reserve()` to some reasonable value here?

https://github.com/llvm/llvm-project/pull/140498