Merge pull request #22 from brettlangdon/dev/semicolon.optional.sqwished

Make semicolons optional, require newline instead

README.md

@@ -27,11 +27,14 @@ global = "global value";
 primary {
   string = "primary string value";
   single = 'single quotes are allowed too';
-  integer = 500;
-  float = 80.80;
-  boolean = true;
-  negative = FALSE;
-  nothing = NULL;
+
+  # Semicolons are optional
+  integer = 500
+  float = 80.80
+  boolean = true
+  negative = FALSE
+  nothing = NULL
+
   # Include external files
   include "./include*.cfg";
   # Primary-sub section

forge.go

@@ -27,6 +27,7 @@
 //
 //     IDENTIFIER: [_a-zA-Z]+
 //     NUMBERS: [0-9]+
+//     END: ';' | '\n'
 //
 //     BOOL: 'true' | 'false'
 //     NULL: 'null'
@@ -36,10 +37,10 @@
 //     REFERENCE: (IDENTIFIER)? ('.' IDENTIFIER)+
 //     VALUE: BOOL | NULL | INTEGER | FLOAT | STRING | REFERENCE
 //
-//     INCLUDE: 'include ' STRING ';'
-//     DIRECTIVE: (IDENTIFIER '=' VALUE | INCLUDE) ';'
+//     INCLUDE: 'include ' STRING END
+//     DIRECTIVE: (IDENTIFIER '=' VALUE | INCLUDE) END
 //     SECTION: IDENTIFIER '{' (DIRECTIVE | SECTION)* '}'
-//     COMMENT: '#' .* NEWLINE '\n'
+//     COMMENT: '#' .* '\n'
 //
 //     CONFIG_FILE: (COMMENT | DIRECTIVE | SECTION)*
 //
@@ -70,7 +71,7 @@
 //      until after the newline.
 //  * Directive:
 //      A directive is a setting, a identifier and a value. They are in the format '<identifier> = <value>;'
-//      All directives must end in a semicolon. The value can be any of the types defined above.
+//      All directives must end in either a semicolon or newline. The value can be any of the types defined above.
 //  * Section:
 //      A section is a grouping of directives under a common name. They are in the format '<section_name> { <directives> }'.
 //      All sections must be wrapped in brackets ('{', '}') and must all have a name. They do not end in a semicolon.

forge_test.go

@@ -17,12 +17,15 @@ primary {
   string_with_quote = "some \"quoted\" str\\ing";
   single = 'hello world';
   single_with_quote = '\'hello\' "world"';
-  integer = 500;
-  float = 80.80;
-  negative = -50;
-  boolean = true;
-  not_true = FALSE;
-  nothing = NULL;
+
+  # Semicolons are optional
+  integer = 500
+  float = 80.80
+  negative = -50
+  boolean = true
+  not_true = FALSE
+  nothing = NULL
+
   # Reference secondary._under (which hasn't been defined yet)
   sec_ref = secondary._under;
    # Primary-sub stuff

parser.go

@@ -11,6 +11,10 @@ import (
 	"github.com/brettlangdon/forge/token"
 )
 
+func isSemicolonOrNewline(id token.TokenID) bool {
+	return id == token.SEMICOLON || id == token.NEWLINE
+}
+
 // Parser is a struct to hold data necessary for parsing a config from a scanner
 type Parser struct {
 	files      []string
@@ -88,10 +92,10 @@ func (parser *Parser) parseReference(startingSection *Section, period bool) (Val
 			}
 			name += parser.curTok.Literal
 			period = false
-		} else if parser.curTok.ID == token.SEMICOLON {
+		} else if isSemicolonOrNewline(parser.curTok.ID) {
 			break
 		} else {
-			msg := fmt.Sprintf("expected ';' instead found '%s'", parser.curTok.Literal)
+			msg := fmt.Sprintf("expected ';' or '\n' instead found '%s'", parser.curTok.Literal)
 			return nil, parser.syntaxError(msg)
 		}
 	}
@@ -159,8 +163,8 @@ func (parser *Parser) parseSetting(name string) error {
 	if readNext {
 		parser.readToken()
 	}
-	if parser.curTok.ID != token.SEMICOLON {
-		msg := fmt.Sprintf("expected ';' instead found '%s'", parser.curTok.Literal)
+	if isSemicolonOrNewline(parser.curTok.ID) == false {
+		msg := fmt.Sprintf("expected ';' or '\n' instead found '%s'", parser.curTok.Literal)
 		return parser.syntaxError(msg)
 	}
 	parser.readToken()
@@ -177,8 +181,8 @@ func (parser *Parser) parseInclude() error {
 	pattern := parser.curTok.Literal
 
 	parser.readToken()
-	if parser.curTok.ID != token.SEMICOLON {
-		msg := fmt.Sprintf("expected ';' instead found '%s'", parser.curTok.Literal)
+	if isSemicolonOrNewline(parser.curTok.ID) == false {
+		msg := fmt.Sprintf("expected ';' or '\n' instead found '%s'", parser.curTok.Literal)
 		return parser.syntaxError(msg)
 	}
 
@@ -260,6 +264,9 @@ func (parser *Parser) parse() error {
 			if err != nil {
 				return err
 			}
+		case token.NEWLINE:
+			// Ignore extra newlines
+			continue
 		default:
 			return parser.syntaxError(fmt.Sprintf("unexpected token %s", tok))
 		}

scanner.go

@@ -18,8 +18,8 @@ func isDigit(ch rune) bool {
 	return ('0' <= ch && ch <= '9')
 }
 
-func isWhitespace(ch rune) bool {
-	return (ch == ' ' || ch == '\t' || ch == '\n' || ch == '\r')
+func isNonNewlineWhitespace(ch rune) bool {
+	return (ch == ' ' || ch == '\t' || ch == '\r')
 }
 
 func isBoolean(str string) bool {
@@ -171,10 +171,10 @@ func (scanner *Scanner) parseComment() {
 	scanner.readRune()
 }
 
-func (scanner *Scanner) skipWhitespace() {
+func (scanner *Scanner) skipNonNewlineWhitespace() {
 	for {
 		scanner.readRune()
-		if !isWhitespace(scanner.curCh) {
+		if !isNonNewlineWhitespace(scanner.curCh) {
 			break
 		}
 	}
@@ -182,8 +182,8 @@ func (scanner *Scanner) skipWhitespace() {
 
 // NextToken will read in the next valid token from the Scanner
 func (scanner *Scanner) NextToken() token.Token {
-	if isWhitespace(scanner.curCh) {
-		scanner.skipWhitespace()
+	if isNonNewlineWhitespace(scanner.curCh) {
+		scanner.skipNonNewlineWhitespace()
 	}
 
 	scanner.curTok = token.Token{
@@ -217,6 +217,8 @@ func (scanner *Scanner) NextToken() token.Token {
 			scanner.curTok.ID = token.RBRACKET
 		case ';':
 			scanner.curTok.ID = token.SEMICOLON
+		case '\n':
+			scanner.curTok.ID = token.NEWLINE
 		case '.':
 			scanner.curTok.ID = token.PERIOD
 		case '-':

test.cfg

@@ -6,12 +6,15 @@ primary {
   string_with_quote = "some \"quoted\" str\\ing";
   single = 'hello world';
   single_with_quote = '\'hello\' "world"';
-  integer = 500;
-  float = 80.80;
-  negative = -50;
-  boolean = true;
-  not_true = FALSE;
-  nothing = NULL;
+
+  # Semicolons are optional
+  integer = 500
+  float = 80.80
+  negative = -50
+  boolean = true
+  not_true = FALSE
+  nothing = NULL
+
   # Reference secondary._under (which hasn't been defined yet)
   sec_ref = secondary._under;
   # Primary-sub stuff

token/token.go

@@ -26,6 +26,7 @@ const (
 	RBRACKET
 	EQUAL
 	SEMICOLON
+	NEWLINE
 	PERIOD
 
 	IDENTIFIER
@@ -45,6 +46,7 @@ var tokenNames = [...]string{
 	RBRACKET:   "RBRACKET",
 	EQUAL:      "EQUAL",
 	SEMICOLON:  "SEMICOLON",
+	NEWLINE:    "NEWLINE",
 	PERIOD:     "PERIOD",
 	IDENTIFIER: "IDENTIFIER",
 	BOOLEAN:    "BOOLEAN",