fix "..." handling between parser, AST, and CST

go/src/miller/dsl/ast.go

@@ -113,6 +113,21 @@ func NewASTNodeStripDollarPlease(itok interface{}, nodeType TNodeType) (*ASTNode
 	return NewASTNodeNestable(newToken, nodeType), nil
 }
 
+// Likewise for the leading/trailing double quotes on string literals.
+func NewASTNodeStripDoubleQuotePairPlease(
+	itok interface{},
+	nodeType TNodeType,
+) (*ASTNode, error) {
+	oldToken := itok.(*token.Token)
+	n := len(oldToken.Lit)
+	newToken := &token.Token{
+		Type: oldToken.Type,
+		Lit: oldToken.Lit[1:n-1],
+		Pos: oldToken.Pos,
+	}
+	return NewASTNodeNestable(newToken, nodeType), nil
+}
+
 // xxx comment why grammar use
 func NewASTNodeNestable(itok interface{}, nodeType TNodeType) *ASTNode {
 	var tok *token.Token = nil

go/src/miller/parsing/mlr.bnf

@@ -258,6 +258,10 @@ md_token_FILENUM  : 'F' 'I' 'L' 'E' 'N' 'U' 'M' ;
 
 // Note: the parser depends on the dollar sign being here. If this is changed,
 // that needs to be changed as well.
+//
+// Also note: if we omit the '$' here and include it in the parser section
+// below as "$", then we get an LR-1 conflict. So this must be dealt with at
+// the AST level.
 md_token_field_name : '$' _idchar { _idchar } ;
 
 //\$\* {
@@ -2543,6 +2547,11 @@ FieldName
   // ...
 ;
 
+// Note: the field name is "$field" not "field" since md_token_field_name
+// includes the '$'.  If we omit the '$' there and include it in the parser
+// section here as "$", then we get an LR-1 conflict. So this must be dealt
+// with at the AST level. Hence the NewASTNodeStripDollarPlease.
+
 DirectFieldName
   : md_token_field_name
     << dsl.NewASTNodeStripDollarPlease($0, dsl.NodeTypeDirectFieldName) >>
@@ -2641,8 +2650,14 @@ IndirectFieldName
 // xxx split out node-type string vs number so we can track parsed-type through
 // to the lrecs, with mlrvals there! :)
 
+// As with '$' on md_token_field_name, so too for md_token_string_literal
+// we get LR-1 conflicts if we attempt to put the double quotes here.
+// Hence the quote-stripper AST method.
 AtomOrFunction
-  : md_token_string_literal  << dsl.NewASTNode($0, dsl.NodeTypeStringLiteral) >>
+  : md_token_string_literal  << dsl.NewASTNodeStripDoubleQuotePairPlease(
+    $0,
+    dsl.NodeTypeStringLiteral,
+  ) >>
   | md_token_int_literal     << dsl.NewASTNode($0, dsl.NodeTypeIntLiteral) >>
   | md_token_float_literal   << dsl.NewASTNode($0, dsl.NodeTypeFloatLiteral) >>
   | md_token_boolean_literal << dsl.NewASTNode($0, dsl.NodeTypeBoolLiteral) >>

go/src/miller/parsing/parser/productionstable.go

@@ -841,13 +841,19 @@ var productionsTable = ProdTab{
 		},
 	},
 	ProdTabEntry{
-		String: `AtomOrFunction : md_token_string_literal	<< dsl.NewASTNode(X[0], dsl.NodeTypeStringLiteral) >>`,
+		String: `AtomOrFunction : md_token_string_literal	<< dsl.NewASTNodeStripDoubleQuotePairPlease(
+    X[0],
+    dsl.NodeTypeStringLiteral,
+  ) >>`,
 		Id:         "AtomOrFunction",
 		NTType:     22,
 		Index:      82,
 		NumSymbols: 1,
 		ReduceFunc: func(X []Attrib) (Attrib, error) {
-			return dsl.NewASTNode(X[0], dsl.NodeTypeStringLiteral)
+			return dsl.NewASTNodeStripDoubleQuotePairPlease(
+    X[0],
+    dsl.NodeTypeStringLiteral,
+  )
 		},
 	},
 	ProdTabEntry{