From 71157122d6b403763ffcc677352ab7189f2ddc46 Mon Sep 17 00:00:00 2001
From: Sasha Koshka <sashakoshka@tebibyte.media>
Date: Thu, 11 Aug 2022 12:39:49 -0500
Subject: [PATCH] Added support for hex, octal, and unicode escape sequences

---
 lexer/lexer_test.go |   2 +
 lexer/text.go       | 130 +++++++++++++++++++++++++++++++++++---------
 tests/lexer/text    |   1 +
 3 files changed, 106 insertions(+), 27 deletions(-)

diff --git a/lexer/lexer_test.go b/lexer/lexer_test.go
index 2dca3f8..4885e1a 100644
--- a/lexer/lexer_test.go
+++ b/lexer/lexer_test.go
@@ -133,5 +133,7 @@ func TestTokenizeText (test *testing.T) {
 		Token { kind: TokenKindRune, value: '"'  },
 		Token { kind: TokenKindRune, value: '\\' },
 		Token { kind: TokenKindNewline },
+		Token { kind: TokenKindString, value: "hello world \x40\u0040\U00000040!" },
+		Token { kind: TokenKindNewline },
 	}, test)
 }
diff --git a/lexer/text.go b/lexer/text.go
index e349581..31f3e3a 100644
--- a/lexer/text.go
+++ b/lexer/text.go
@@ -1,20 +1,9 @@
 package lexer
 
+import "strconv"
 import "github.com/sashakoshka/arf/file"
 
-var escapeSequenceMap = map[rune] rune {
-        'a':  '\x07',
-        'b':  '\x08',
-        'f':  '\x0c',
-        'n':  '\x0a',
-        'r':  '\x0d',
-        't':  '\x09',
-        'v':  '\x0b',
-        '\'': '\'',
-        '"':  '"',
-        '\\': '\\',
-}
-
+// tokenizeString tokenizes a string or rune literal.
 func (lexer *LexingOperation) tokenizeString (isRuneLiteral bool) (err error) {
 	err = lexer.nextRune()
 	if err != nil { return }
@@ -26,23 +15,18 @@ func (lexer *LexingOperation) tokenizeString (isRuneLiteral bool) (err error) {
 		if lexer.char == '\\' {
 			err = lexer.nextRune()
 			if err != nil { return }
-	
-			actual, exists := escapeSequenceMap[lexer.char]
-			if exists {
-				got += string(actual)
-			} else {
-				err = file.NewError (
-					lexer.file.Location(), 1,
-					"unknown escape character " +
-					string(lexer.char), file.ErrorKindError)
-				return
-			}
+
+			var actual rune
+			actual, err = lexer.getEscapeSequence()
+			if err != nil { return }
+
+			got += string(actual)
 		} else {
 			got += string(lexer.char)
+			
+			err = lexer.nextRune()
+			if err != nil { return }
 		}
-		
-		err = lexer.nextRune()
-		if err != nil { return }
 
 		if isRuneLiteral {
 			if lexer.char == '\'' { break }
@@ -75,3 +59,95 @@ func (lexer *LexingOperation) tokenizeString (isRuneLiteral bool) (err error) {
 	lexer.addToken(token)
 	return
 }
+
+// escapeSequenceMap contains basic escape sequences and how they map to actual
+// runes.
+var escapeSequenceMap = map[rune] rune {
+        'a':  '\x07',
+        'b':  '\x08',
+        'f':  '\x0c',
+        'n':  '\x0a',
+        'r':  '\x0d',
+        't':  '\x09',
+        'v':  '\x0b',
+        '\'': '\'',
+        '"':  '"',
+        '\\': '\\',
+}
+
+// getEscapeSequence reads an escape sequence in a string or rune literal.
+func (lexer *LexingOperation) getEscapeSequence () (result rune, err error) {
+	result, exists := escapeSequenceMap[lexer.char]
+	if exists {
+                err = lexer.nextRune()
+                return
+	} else if lexer.char >= '0' && lexer.char <= '7' {
+                // octal escape sequence
+                number := string(lexer.char)
+        
+                err = lexer.nextRune()
+                if err != nil { return }
+                
+                for len(number) < 3 {
+                        if lexer.char < '0' || lexer.char > '7' { break }
+
+                        number += string(lexer.char)
+                        
+	                err = lexer.nextRune()
+	                if err != nil { return }
+                }
+                
+                if len(number) < 3 {
+			err = file.NewError (
+				lexer.file.Location(), 1,
+				"octal escape sequence too short",
+				file.ErrorKindError)
+			return
+                }
+
+                parsedNumber, _ := strconv.ParseInt(number, 8, 8)
+                result = rune(parsedNumber)
+                
+        } else if lexer.char == 'x' || lexer.char == 'u' || lexer.char == 'U' {
+                // hexidecimal escape sequence
+                want := 2
+                if lexer.char == 'u' { want = 4 }
+                if lexer.char == 'U' { want = 8 }
+        
+                number := ""
+
+                err = lexer.nextRune()
+                if err != nil { return }
+                
+                for len(number) < want {
+                        notLower := lexer.char < 'a' || lexer.char > 'f'
+                        notUpper := lexer.char < 'A' || lexer.char > 'F'
+                        notNum   := lexer.char < '0' || lexer.char > '9'
+                        if notLower && notUpper && notNum { break }
+                        
+                        number += string(lexer.char)
+                        
+			err = lexer.nextRune()
+			if err != nil { return }
+                }
+                
+                if len(number) < want {
+			err = file.NewError (
+				lexer.file.Location(), 1,
+				"hex escape sequence too short ",
+				file.ErrorKindError)
+			return
+                }
+
+                parsedNumber, _ := strconv.ParseInt(number, 16, want * 4)
+                result = rune(parsedNumber)
+	} else {
+		err = file.NewError (
+			lexer.file.Location(), 1,
+			"unknown escape character " +
+			string(lexer.char), file.ErrorKindError)
+		return
+	}
+
+	return
+}
diff --git a/tests/lexer/text b/tests/lexer/text
index 3515bfc..be5b8c6 100644
--- a/tests/lexer/text
+++ b/tests/lexer/text
@@ -1,2 +1,3 @@
 "hello world!\a\b\f\n\r\t\v\'\"\\"
 '\a' '\b' '\f' '\n' '\r' '\t' '\v' '\'' '\"' '\\'
+"hello world \x40\u0040\U00000040!"