Backport string escapes and many op-assign

3 files changed, 210 insertions, 11 deletions

diff --git a/chapter22/src/main/java/com/seaofnodes/simple/Parser.java b/chapter22/src/main/java/com/seaofnodes/simple/Parser.java
index 23e9ddee..c26b9064 100644
--- a/chapter22/src/main/java/com/seaofnodes/simple/Parser.java
+++ b/chapter22/src/main/java/com/seaofnodes/simple/Parser.java
@@ -1172,11 +1172,17 @@ public class Parser {
         // flt + flt ==>> use float op
         Node op = switch(ch) {
         case 1, (char)-1,
-             '+' -> new AddNode(lhs,rhs);
-        case '-' -> new SubNode(lhs,rhs);
-        case '*' -> new MulNode(lhs,rhs);
-        case '/' -> new DivNode(lhs,rhs);
-        default  -> throw Utils.TODO();
+             '+'        -> new AddNode(lhs,rhs);
+        case '-'        -> new SubNode(lhs,rhs);
+        case '*'        -> new MulNode(lhs,rhs);
+        case '/'        -> new DivNode(lhs,rhs);
+        case '&'        -> new AndNode(loc(),lhs,rhs);
+        case '|'        -> new  OrNode(loc(),lhs,rhs);
+        case '^'        -> new XorNode(loc(),lhs,rhs);
+        case Lexer.SHL  -> new ShlNode(loc(),lhs,rhs);
+        case Lexer.SAR  -> new SarNode(loc(),lhs,rhs);
+        case Lexer.SHR  -> new ShrNode(loc(),lhs,rhs);
+        default         -> throw Utils.TODO();
         };
         // Convert to float ops, or narrow int types; error if not declared type.
         // Also, if postfix LHS is still keep()
@@ -1546,16 +1552,44 @@ public class Parser {
     private String parseString() {
         if( !peek('"') ) return null;
         _lexer.inc();
-        int start = pos();
-        while( !_lexer.isEOF() && _lexer.nextChar()!= '"' ) ;
-        if( _lexer.isEOF() )
-            throw error("Unclosed string");
-        return new String(_lexer._input,start,pos()-start-1);
+        StringBuilder sb = new StringBuilder();
+        while( !_lexer.isEOF() ) {
+            char c = _lexer.nextChar();
+            if( c=='"' ) return sb.toString();
+            if( c!='\\' ) {
+                sb.append(c);
+                continue;
+            }
+            if( _lexer.isEOF() )
+                throw error("Unclosed string");
+            sb.append(escapedChar("string"));
+        }
+        throw error("Unclosed string");
+    }
+
+    private char escapedChar(String kind) {
+        char esc = _lexer.nextChar();
+        return switch( esc ) {
+        case '0'  -> '\u0000';
+        case 'n'  -> '\n';
+        case 't'  -> '\t';
+        case 'r'  -> '\r';
+        case '\\' -> '\\';
+        case '"'  -> '"';
+        case '\'' -> '\'';
+        default -> throw error("Unknown "+kind+" escape \\\\"+esc);
+        };
     }
 
     // Already parsed "'"
     private Node parseChar() {
-        return require(con(TypeInteger.constant(_lexer.nextChar())),"'");
+        if( _lexer.isEOF() ) throw error("Unclosed character");
+        char c = _lexer.nextChar();
+        if( c=='\\' ) {
+            if( _lexer.isEOF() ) throw error("Unclosed character");
+            c = escapedChar("character");
+        }
+        return require(con(TypeInteger.constant(c)),"'");
     }
 
     //////////////////////////////////
@@ -1834,8 +1868,14 @@ public class Parser {
 
         // Next oper= character, or 0.
         // As a convenience, mark "++" as a char 1 and "--" as char -1 (65535)
+        // Distinct tags for the multi-character compound operators.
+        static final char SHL = 2, SAR = 3, SHR = 4;
+
         public char matchOperAssign() {
             skipWhiteSpace();
+            if( match("<<=" ) ) return SHL;
+            if( match(">>>=") ) return SHR;
+            if( match(">>=" ) ) return SAR;
             if( _position+2 >= _input.length ) return 0;
             char ch0 = (char)_input[_position];
             if( "+-/*&|^".indexOf(ch0) == -1 ) return 0;
diff --git a/chapter22/src/test/java/com/seaofnodes/simple/CompoundAssignmentTest.java b/chapter22/src/test/java/com/seaofnodes/simple/CompoundAssignmentTest.java
new file mode 100644
index 00000000..0d67d3c5
--- /dev/null
+++ b/chapter22/src/test/java/com/seaofnodes/simple/CompoundAssignmentTest.java
@@ -0,0 +1,94 @@
+package com.seaofnodes.simple;
+
+import com.seaofnodes.simple.codegen.CodeGen;
+import org.junit.Test;
+import static org.junit.Assert.*;
+
+public class CompoundAssignmentTest {
+    private static final String[] OPS = {"&=", "|=", "^=", "<<=", ">>=", ">>>="};
+
+    private static long apply(String op, long lhs, long rhs) {
+        return switch (op) {
+            case "&=" -> lhs & rhs;
+            case "|=" -> lhs | rhs;
+            case "^=" -> lhs ^ rhs;
+            case "<<=" -> lhs << rhs;
+            case ">>=" -> lhs >> rhs;
+            case ">>>=" -> lhs >>> rhs;
+            default -> throw new AssertionError(op);
+        };
+    }
+
+    @Test public void testLocals() {
+        for (String op : OPS)
+            for (long arg : new long[]{-8, 0, 5, 64}) {
+                long expected = apply(op, arg, 3);
+                check("int x=arg; int r=(x " + op + " 3); return x*10+r;",
+                      arg, expected*11);
+            }
+    }
+
+    @Test public void testRhsOnce() {
+        for (String op : OPS) {
+            long expected = apply(op, 5, 1);
+            check("int x=arg; int count=0; int r=(x " + op +
+                  " (count=count+1)); return count*1000+x*10+r;", 5, 1000+expected*11);
+        }
+    }
+
+    @Test public void testNarrowingAndChaining() {
+        check("u8 x=255; x <<= 1; return x;", 0, 254);
+        check("u8 x=255; x ^= 256; return x;", 0, 255);
+        check("int x=1; int y=3; x |= y <<= 2; return x*100+y;", 0, 1312);
+        check("int x=-8; x >>= 65; return x;", 0, -4);
+        check("int x=-8; x >>>= 65; return x;", 0, Long.MAX_VALUE-3);
+    }
+
+    @Test public void testExistingOperators() {
+        check("int x=8; x+=2; x-=1; x*=3; x/=9; return x;", 0, 3);
+        check("int x=arg; int y=x++; int z=x--; return x*100+y*10+z;", 5, 556);
+        check("return (arg << 2) + (arg >> 1) + (arg >>> 1);", 8, 40);
+        check("return arg <= 8;", 8, 1);
+        check("return arg >= 8;", 8, 1);
+    }
+
+    @Test public void testRejectedAssignments() {
+        for (String op : OPS) {
+            rejected("val x=3; x " + op + " 1; return x;");
+            rejected("int x=3; x " + op + " 1.5; return x;");
+        }
+    }
+
+    private static void rejected(String src) {
+        try {
+            check(src, 0, 0);
+            fail("Accepted invalid assignment: " + src);
+        } catch (RuntimeException e) {
+            assertTrue("Must report a language error: " + e, e instanceof Parser.ParseException);
+            assertNotNull(e.getMessage());
+            assertFalse("Must report a language error: " + e, e.getMessage().contains("Not yet implemented"));
+        }
+    }
+
+    @Test public void testFields() {
+        for (String op : OPS) {
+            long expected = apply(op, -8, 3);
+            check("struct S { int x; }; S !s=new S; s.x=arg; int r=(s.x " +
+                  op + " 3); return s.x*10+r;", -8, expected*11);
+        }
+    }
+
+    @Test public void testArrayAddressAndRhsOnce() {
+        for (String op : OPS) {
+            long expected = apply(op, 5, 1);
+            check("int[] !a=new int[2]; a[0]=arg; int i=0; int count=0; " +
+                  "int r=(a[i++] " + op + " (count=count+1)); " +
+                  "return i*1000+count*100+a[0]*10+r;", 5, 1100+expected*11);
+        }
+    }
+
+    private static void check(String src, long arg, long expected) {
+        var code = new CodeGen(src).parse().opto().typeCheck();
+        assertEquals(src, Long.toString(expected), Eval2.eval(code, arg));
+    }
+}
diff --git a/chapter22/src/test/java/com/seaofnodes/simple/LiteralTest.java b/chapter22/src/test/java/com/seaofnodes/simple/LiteralTest.java
new file mode 100644
index 00000000..2375f213
--- /dev/null
+++ b/chapter22/src/test/java/com/seaofnodes/simple/LiteralTest.java
@@ -0,0 +1,65 @@
+package com.seaofnodes.simple;
+
+import com.seaofnodes.simple.codegen.CodeGen;
+import com.seaofnodes.simple.type.TypeConAryB;
+import com.seaofnodes.simple.type.TypeInteger;
+import com.seaofnodes.simple.type.TypeMemPtr;
+import org.junit.Test;
+import static org.junit.Assert.*;
+
+public class LiteralTest {
+    private static final String[] ESCAPES = {"\\0", "\\n", "\\t", "\\r", "\\\\", "\\\"", "\\'"};
+    private static final int[] VALUES = {0, 10, 9, 13, 92, 34, 39};
+
+    @Test public void testStringEscapes() {
+        for (int i = 0; i < ESCAPES.length; i++)
+            assertString(ESCAPES[i], new byte[]{(byte)VALUES[i]});
+        assertString("a\\\"b\\\\c\\0d", new byte[]{97, 34, 98, 92, 99, 0, 100});
+        assertString("", new byte[]{});
+        assertString("plain", new byte[]{112, 108, 97, 105, 110});
+    }
+
+    private static void assertString(String text, byte[] expected) {
+        var code = new CodeGen("return \"" + text + "\";").parse().opto().typeCheck();
+        TypeMemPtr ptr = (TypeMemPtr)code.expr()._type;
+        TypeConAryB bytes = (TypeConAryB)ptr._obj._con;
+        assertArrayEquals(text, expected, bytes._ary);
+    }
+
+    @Test public void testCharacterEscapes() {
+        for (int i = 0; i < ESCAPES.length; i++)
+            assertCharacter(ESCAPES[i], VALUES[i]);
+        assertCharacter("A", 65);
+    }
+
+    private static void assertCharacter(String text, int expected) {
+        var code = new CodeGen("return '" + text + "';").parse().opto().typeCheck();
+        assertEquals(text, expected, ((TypeInteger)code.expr()._type).value());
+    }
+
+    @Test public void testMalformedStrings() {
+        assertBadLiteral("return \"\\q\";", "Unknown string escape");
+        assertBadLiteral("return \"abc", "Unclosed string");
+        assertBadLiteral("return \"abc\\", "Unclosed string");
+        // The quote closes the string at EOF; the missing token is the semicolon.
+        assertBadLiteral("return \"plain\"", ";");
+    }
+
+    @Test public void testMalformedCharacters() {
+        assertBadLiteral("return '\\q';", "Unknown character escape");
+        assertBadLiteral("return '", "Unclosed character");
+        assertBadLiteral("return '\\", "Unclosed character");
+        assertBadLiteral("return '';", "Syntax error");
+        assertBadLiteral("return 'ab';", "Syntax error");
+        assertBadLiteral("return 'a", "Syntax error");
+    }
+
+    private static void assertBadLiteral(String src, String message) {
+        try {
+            new CodeGen(src).parse();
+            fail("Accepted malformed literal: " + src);
+        } catch (Parser.ParseException e) {
+            assertTrue(e.getMessage(), e.getMessage().contains(message));
+        }
+    }
+}