[llvm] [TableGen] Implement a preprocessor (PR #120113)

[Jay Foad via llvm-commits]

================
@@ -68,12 +68,65 @@ static cl::opt<bool> NoWarnOnUnusedTemplateArgs(
     "no-warn-on-unused-template-args",
     cl::desc("Disable unused template argument warnings."));
 
+static cl::opt<bool> Preprocess("E", cl::desc("Write preprocessed output"));
+
 static int reportError(const char *ProgName, Twine Msg) {
   errs() << ProgName << ": " << Msg;
   errs().flush();
   return 1;
 }
 
+/// Encapsulate file, line and column numbers from SourceMgr.
+struct SMCoords {
+  unsigned Buf = 0;
+  unsigned Line = 0;
+  unsigned Col = 0;
+  SMCoords() = default;
+  SMCoords(const SourceMgr &Mgr, SMLoc Loc) {
+    Buf = Mgr.FindBufferContainingLoc(Loc);
+    // TODO: SourceMgr::getLineAndColumn is not a fast method. Find a better way
+    // to do this. For example we don't need the column number for every token,
+    // only the first token on each output line.
+    std::tie(Line, Col) = Mgr.getLineAndColumn(Loc, Buf);
+  }
+};
+
+/// Create preprocessed output for `-E` option.
+static int preprocessInput(raw_ostream &OS) {
+  TGLexer Lex(SrcMgr, MacroNames);
+  SMCoords Last;
+  bool Any = false;
+  while (true) {
+    Lex.Lex();
+    if (Lex.getCode() == tgtok::Eof || Lex.getCode() == tgtok::Error)
+      break;
+    SMCoords This(SrcMgr, Lex.getLoc());
+    if (This.Buf == Last.Buf && This.Line == Last.Line) {
+      // Add a single space between tokens on the same line. This is overkill in
+      // many cases but at least it will parse correctly.
+      OS << ' ';
+    } else if (Last.Buf) {
+      // Always start a new line when including a new file or popping back out
+      // to the previous file. This is just a heuristic to make the output look
+      // reasonably pretty.
+      OS << '\n';
+      // Indent the first token on a line to its original indentation, to make
+      // the output look pretty.
+      OS.indent(This.Col - 1);
+    }
+
+    const char *Start = Lex.getLoc().getPointer();
+    const char *End = Lex.getLocRange().End.getPointer();
+    OS << StringRef(Start, End - Start);
+    Any = true;
+
+    Last = This;
+  }
+  if (Any)
+    OS << '\n';
+  return Lex.getCode() == tgtok::Error;
----------------
jayfoad wrote:

No, if there is any error then there will be no output. This is the same as any regular tablegen backend and it works because all the output so far has been written to an in-memory buffer.

https://github.com/llvm/llvm-project/pull/120113