From 224012b708fa8ee6f0617c0ef95f2b5c20cad14e Mon Sep 17 00:00:00 2001
From: Miha Korenjak <miha@korenjak.si>
Date: Thu, 10 Nov 2022 20:33:38 +0000
Subject: [PATCH] add escape sequences in character strings (#38)

* add escape sequences in character strings

Standard escape sequences (those used in java) and ability to write
bytes directly (\xbb).

* remove unneeded line in escape sequence parsing

* fix handling of unterminated byte strings

* add double quoted strings

To avoid breaking backwards compatibility, escaped strings need to be
enclosed in double quotes.
Moved code for lexing to Lexer.
---
 src/sic/asm/parsing/Lexer.java | 48 ++++++++++++++++++++++++++++++++++
 src/sic/ast/data/DataChr.java  | 13 +++++++--
 2 files changed, 59 insertions(+), 2 deletions(-)

diff --git a/src/sic/asm/parsing/Lexer.java b/src/sic/asm/parsing/Lexer.java
index 747f439..9604e62 100644
--- a/src/sic/asm/parsing/Lexer.java
+++ b/src/sic/asm/parsing/Lexer.java
@@ -218,4 +218,52 @@ public double readFloat() throws AsmError {
         return num;
     }
 
+    public String readEscapedString(char terminator) throws AsmError {
+        StringBuilder buf = new StringBuilder();
+        for (char c = advance(); c != terminator; c = advance()) {
+            if (!ready() || c == '\n') {
+                throw new AsmError(loc(), "Unterminated byte string");
+            }
+            if (c == '\\') {
+                c = advance();
+                switch (c) {
+                    case '\"', '\\':
+                        break;
+                    case 'n':
+                        c = '\n';
+                        break;
+                    case 'r':
+                        c = '\r';
+                        break;
+                    case 't':
+                        c = '\t';
+                        break;
+                    case 'b':
+                        c = '\b';
+                        break;
+                    case 'f':
+                        c = '\f';
+                        break;
+                    case '0':
+                        c = '\0';
+                        break;
+                    case 'x':
+                        int mark = pos();
+                        advance(2);
+                        String str = extract(mark);
+                        try {
+                            c = (char) Integer.parseInt(str, 16);
+                        } catch (NumberFormatException e) {
+                            throw new AsmError(loc(), "Hexadecimal byte expected");
+                        }
+                        break;
+                    default:
+                        throw new AsmError(loc(), "Unknown escape sequence '\\%c'", c);
+                }
+            }
+            buf.append(c);
+        }
+        return buf.toString();
+    }
+
 }
diff --git a/src/sic/ast/data/DataChr.java b/src/sic/ast/data/DataChr.java
index eba4439..f3cf2fb 100644
--- a/src/sic/ast/data/DataChr.java
+++ b/src/sic/ast/data/DataChr.java
@@ -26,8 +26,17 @@ public String toString() {
     @Override
     public void parse(Parser parser, boolean allowList) throws AsmError {
         parser.advance('C');
-        parser.advance('\'');
-        data = parser.readUntil('\'').getBytes();
+        char quote = parser.advance();
+        switch (quote) {
+            case '\'':
+                data = parser.readUntil('\'').getBytes();
+                break;
+            case '"':
+                data = parser.readEscapedString('"').getBytes();
+                break;
+            default:
+                throw new AsmError(parser.loc(), "Expected quote");
+        }
         if (allowList) super.parse(parser, true);
     }