[FIRParser] Parse RawString ExtModule Parameters (#1340)

docs/RationaleFIRRTL.md

@@ -46,10 +46,9 @@ support is planned, but not implemented.
 
 There are some exceptions to the above:
 
-1) We don't support the `'raw string'` syntax for strings.
-2) We don't support the `Fixed` types for fixed point numbers, and some
+1) We don't support the `Fixed` types for fixed point numbers, and some
    primitives associated with them.
-3) We don't support `Interval` types
+2) We don't support `Interval` types
 
 Some of these may be research efforts that didn't gain broad adoption, in which
 case we don't want to support them.  However, if there is a good reason and a

lib/Dialect/FIRRTL/Import/FIRLexer.cpp

@@ -66,7 +66,7 @@ std::string FIRToken::getStringValue(StringRef spelling) {
 
   std::string result;
   result.reserve(bytes.size());
-  for (unsigned i = 0, e = bytes.size(); i != e;) {
+  for (size_t i = 0, e = bytes.size(); i != e;) {
     auto c = bytes[i++];
     if (c != '\\') {
       result.push_back(c);
@@ -76,17 +76,27 @@ std::string FIRToken::getStringValue(StringRef spelling) {
     assert(i + 1 <= e && "invalid string should be caught by lexer");
     auto c1 = bytes[i++];
     switch (c1) {
-    case '"':
     case '\\':
+    case '"':
+    case '\'':
       result.push_back(c1);
       continue;
+    case 'b':
+      result.push_back('\b');
+      continue;
     case 'n':
       result.push_back('\n');
       continue;
     case 't':
       result.push_back('\t');
       continue;
-      // TODO: Handle the rest of the escapes.
+    case 'f':
+      result.push_back('\f');
+      continue;
+    case 'r':
+      result.push_back('\r');
+      continue;
+      // TODO: Handle the rest of the escapes (octal and unicode).
     default:
       break;
     }
@@ -101,6 +111,38 @@ std::string FIRToken::getStringValue(StringRef spelling) {
   return result;
 }
 
+/// Given a token containing a raw string, return its value, including removing
+/// the quote characters and unescaping the quotes of the string. The lexer has
+/// already verified that this token is valid.
+std::string FIRToken::getRawStringValue() const {
+  assert(getKind() == raw_string);
+  return getRawStringValue(getSpelling());
+}
+
+std::string FIRToken::getRawStringValue(StringRef spelling) {
+  // Start by dropping the quotes.
+  StringRef bytes = spelling.drop_front().drop_back();
+
+  std::string result;
+  result.reserve(bytes.size());
+  for (size_t i = 0, e = bytes.size(); i != e;) {
+    auto c = bytes[i++];
+    if (c != '\\') {
+      result.push_back(c);
+      continue;
+    }
+
+    assert(i + 1 <= e && "invalid string should be caught by lexer");
+    auto c1 = bytes[i++];
+    if (c1 != '\'') {
+      result.push_back(c);
+    }
+    result.push_back(c1);
+  }
+
+  return result;
+}
+
 //===----------------------------------------------------------------------===//
 // FIRLexer
 //===----------------------------------------------------------------------===//
@@ -243,7 +285,9 @@ FIRToken FIRLexer::lexTokenImpl() {
       continue;
 
     case '"':
-      return lexString(tokStart);
+      return lexString(tokStart, /*isRaw=*/false);
+    case '\'':
+      return lexString(tokStart, /*isRaw=*/true);
 
     case '+':
     case '-':
@@ -390,16 +434,23 @@ void FIRLexer::skipComment() {
 }
 
 /// StringLit      ::= '"' UnquotedString? '"'
+/// RawString      ::= '\'' UnquotedString? '\''
 /// UnquotedString ::= ( '\\\'' | '\\"' | ~[\r\n] )+?
 ///
-FIRToken FIRLexer::lexString(const char *tokStart) {
+FIRToken FIRLexer::lexString(const char *tokStart, bool isRaw) {
   while (1) {
     switch (*curPtr++) {
     case '"': // This is the end of the string literal.
+      if (isRaw)
+        break;
       return formToken(FIRToken::string, tokStart);
+    case '\'': // This is the end of the raw string.
+      if (!isRaw)
+        break;
+      return formToken(FIRToken::raw_string, tokStart);
     case '\\':
-      // Ignore escaped '"'
-      if (*curPtr == '"')
+      // Ignore escaped '\'' or '"'
+      if (*curPtr == '\'' || *curPtr == '"')
         ++curPtr;
       break;
     case 0:
@@ -409,6 +460,7 @@ FIRToken FIRLexer::lexString(const char *tokStart) {
         break;
       LLVM_FALLTHROUGH;
     case '\n': // Vertical whitespace isn't allowed in a string.
+    case '\r':
     case '\v':
     case '\f':
       return emitError(tokStart, "unterminated string");

lib/Dialect/FIRRTL/Import/FIRLexer.h

@@ -74,6 +74,12 @@ public:
   std::string getStringValue() const;
   static std::string getStringValue(StringRef spelling);
 
+  /// Given a token containing a raw string, return its value, including removing
+  /// the quote characters and unescaping the quotes of the string. The lexer has
+  /// already verified that this token is valid.
+  std::string getRawStringValue() const;
+  static std::string getRawStringValue(StringRef spelling);
+
   // Location processing.
   llvm::SMLoc getLoc() const;
   llvm::SMLoc getEndLoc() const;
@@ -128,7 +134,7 @@ private:
   FIRToken lexNumber(const char *tokStart);
   FIRToken lexFloatingPoint(const char *tokStart);
   void skipComment();
-  FIRToken lexString(const char *tokStart);
+  FIRToken lexString(const char *tokStart, bool isRaw);
 
   const llvm::SourceMgr &sourceMgr;
   mlir::MLIRContext *const context;

lib/Dialect/FIRRTL/Import/FIRParser.cpp

@@ -3295,11 +3295,17 @@ ParseResult FIRCircuitParser::parseModule(CircuitOp circuit,
       break;
     }
     case FIRToken::string: {
-      // Drop the quotes and unescape.
+      // Drop the double quotes and unescape.
       value = builder.getStringAttr(getToken().getStringValue());
       consumeToken(FIRToken::string);
       break;
     }
+    case FIRToken::raw_string: {
+      // Drop the single quotes and unescape the ones inside.
+      value = builder.getStringAttr(getToken().getRawStringValue());
+      consumeToken(FIRToken::raw_string);
+      break;
+    }
 
     case FIRToken::floatingpoint:
       double v;

lib/Dialect/FIRRTL/Import/FIRTokenKinds.def

@@ -52,7 +52,7 @@ TOK_LITERAL(integer)        // 42
 TOK_LITERAL(signed_integer) // -42 and +42
 TOK_LITERAL(floatingpoint)  // 42.0
 TOK_LITERAL(string)         // "foo"
-// TOK_LITERAL(raw_string)    // 'foo'
+TOK_LITERAL(raw_string)     // 'foo'
 
 TOK_LITERAL(fileinfo)
 TOK_LITERAL(inlineannotation) // %[{"foo":"bar"}]

test/Dialect/FIRRTL/parse-basic.fir

@@ -857,3 +857,12 @@ circuit MyModule :     ; CHECK: firrtl.circuit "MyModule" {
     output c: UInt<8>
     c <= mux(sel, a, b)
 
+  ; CHECK-LABEL: firrtl.extmodule @RawStringParam
+  ; CHECK: parameters = {FORMAT = "xyz_timeout=%d\\n",
+  ; CHECK:               MIXED_QUOTES = "\22'\\\22",
+  ; CHECK:               TYPE = "bit"}}
+  extmodule RawStringParam :
+    parameter TYPE = 'bit'
+    parameter FORMAT = 'xyz_timeout=%d\n'
+    parameter MIXED_QUOTES = '"\'\"'
+