From 824dc49b592e352f0e6d6a9415afc0c5197b3d1f Mon Sep 17 00:00:00 2001
From: Mart Lubbers <mart@martlubbers.net>
Date: Tue, 9 Feb 2021 13:10:02 +0100
Subject: [PATCH] octal escapes

---
 input.txt |  2 +-
 scan.l    | 14 ++++++++---
 util.c    | 73 +++++++++++++++++++++++++++++++++++++++++++------------
 3 files changed, 68 insertions(+), 21 deletions(-)

diff --git a/input.txt b/input.txt
index c1579d6..3f8969e 100644
--- a/input.txt
+++ b/input.txt
@@ -15,7 +15,7 @@ fun(x){
 	"abr";
 	"a\br";
 	"a\br\"";
-	"a\xaar\\";
+	"a\xaar\\\0377\01\xa";
 return 5;
 f(); 
 f(x); f(1, 2, []);
diff --git a/scan.l b/scan.l
index 9337331..38dab0c 100644
--- a/scan.l
+++ b/scan.l
@@ -1,3 +1,9 @@
+D [0-9]
+H [0-9a-fA-F]
+E ([0\\abtnvfr]|x{H}{H}?|0[0-3]{O}{O}|0{O}{O}?)
+I [a-zA-Z_]
+O [0-7]
+
 %option noinput
 %option nounput
 %{
@@ -68,13 +74,13 @@ Void        return TVOID;
 \[\]        return NIL;
 \.          return DOT;
 ,           return COMMA;
-\"([^\\"]|\\[\"0\\abtnvfr]|\\x[0-9a-fA-F][0-9a-fA-F])*\" {
+\"([^\\"]|\\(\"|{E}))*\" {
 	yylval.expr = expr_string(trimquotes(yytext)); return STRING; }
-'([^\\']|\\['0\\abtnvfr]|\\x[0-9a-fA-F][0-9a-fA-F])' {
+'([^\\']|\\('|{E}))' {
 	yylval.expr = expr_char(trimquotes(yytext)); return CHAR; }
-[0-9]+ {
+{D}+ {
 	yylval.expr = expr_int(atoi(yytext)); return INTEGER; }
-[_a-zA-Z][_a-zA-Z0-9]* {
+{I}({I}|{D})* {
 	yylval.ident = safe_strdup(yytext); return IDENT; }
 }
 <IN_COMMENT>{
diff --git a/util.c b/util.c
index 91166fd..09747c9 100644
--- a/util.c
+++ b/util.c
@@ -2,6 +2,7 @@
 #include <stdlib.h>
 #include <stdio.h>
 #include <string.h>
+#include <ctype.h>
 
 #include "util.h"
 
@@ -76,32 +77,72 @@ char *escape_char(char c, char *buf, bool str)
 	case '\'': strcpy(buf, str ? "'" : "\\'"); break;
 	case '"': strcpy(buf, str ? "\\\"" : "\""); break;
 	default:
-		if (c >= ' ' && c < 127) {
+		if (c >= ' ' && c < 127)
 			sprintf(buf, "%c", c);
-		} else {
+		else
 			sprintf(buf, "\\x%02x", (unsigned char)c);
-		}
 		break;
 	}
 	return buf;
 }
 
+bool isodigit(char c)
+{
+	return c >= '0' && c <= '7';
+}
+
+int fromOctal(char c)
+{
+	if (isodigit(c))
+		return c-'0';
+	return -1;
+}
+
 char *unescape_char(char *c)
 {
-	//escape
 	if (c[0] == '\\') {
-		switch (c[1]) {
-		case '0': c[1] = '\0'; break;
-		case '\'': c[1] = '\''; break;
-		case '\\': c[1] = '\\'; break;
-		case '"': c[1] = '"'; break;
-		case 'a': c[1] = '\a'; break;
-		case 'b': c[1] = '\b'; break;
-		case 't': c[1] = '\t'; break;
-		case 'v': c[1] = '\v'; break;
-		case 'f': c[1] = '\f'; break;
-		case 'r': c[1] = '\r'; break;
-		case 'x': c[3] = (fromHex(c[2])<<4)+fromHex(c[3]); c+=2; break;
+		if (c[1] == 'x' && isxdigit(c[2])) {
+			//two hex
+			if (isxdigit(c[3])) {
+				c[3] = (fromHex(c[2])*16)+fromHex(c[3]);
+				c+=2;
+			//one hex
+			} else {
+				c[2] = fromHex(c[2]);
+				c++;
+			}
+		} else if (c[1] == '0' && isodigit(c[2])) {
+			if (isodigit(c[3])) {
+				//three octal
+				if (isodigit(c[4])) {
+					c[4] = fromOctal(c[2])*64
+						+fromOctal(c[3])*8
+						+fromOctal(c[4]);
+					c+=2;
+				//two octal
+				} else {
+					c[3] = fromOctal(c[2])*8
+						+fromOctal(c[3]);
+					c+=2;
+				}
+			// one octal
+			} else {
+				c[2] = fromOctal(c[2]);
+				c++;
+			}
+		} else {
+			switch (c[1]) {
+			case '0': c[1] = '\0'; break;
+			case '\'': c[1] = '\''; break;
+			case '\\': c[1] = '\\'; break;
+			case '"': c[1] = '"'; break;
+			case 'a': c[1] = '\a'; break;
+			case 'b': c[1] = '\b'; break;
+			case 't': c[1] = '\t'; break;
+			case 'v': c[1] = '\v'; break;
+			case 'f': c[1] = '\f'; break;
+			case 'r': c[1] = '\r'; break;
+			}
 		}
 		c++;
 	}
-- 
2.20.1