diff options
Diffstat (limited to 'src/main/java/com/seaofnodes/simple/parser')
4 files changed, 28 insertions, 4 deletions
diff --git a/src/main/java/com/seaofnodes/simple/parser/Expressions.java b/src/main/java/com/seaofnodes/simple/parser/Expressions.java index 247d974e..6486aba8 100644 --- a/src/main/java/com/seaofnodes/simple/parser/Expressions.java +++ b/src/main/java/com/seaofnodes/simple/parser/Expressions.java @@ -400,7 +400,7 @@ final class Expressions { * Parse integer literal * * <pre> - * integerLiteral: [1-9][0-9]* | [0] + * integerLiteral: [1-9][0-9]* | 0[0-7]* | 0[xX][0-9a-fA-F]+ * floatLiteral: [digits].[digits]?[e [digits]]?[f]? * </pre> */ diff --git a/src/main/java/com/seaofnodes/simple/parser/Lexer.java b/src/main/java/com/seaofnodes/simple/parser/Lexer.java index 5e4b7311..9ab678ca 100644 --- a/src/main/java/com/seaofnodes/simple/parser/Lexer.java +++ b/src/main/java/com/seaofnodes/simple/parser/Lexer.java @@ -239,6 +239,7 @@ public class Lexer { // Return a constant Type, either TypeInteger or TypeFloat Type parseNumber() { Lexer start=new Lexer(this); + if( isHex() ) return parseInteger(parseNumberString(),16,2,start); int old = _position; int len = isLongOrDouble(); boolean f32 = peek()=='f'; @@ -250,7 +251,7 @@ public class Lexer { return TypeFloat.constant(Float.parseFloat(literal)); if( len > 0 ) { if( len > 1 && _input[old]=='0' ) - throw Parser.error("Syntax error: integer values cannot start with '0'",start); + return parseInteger(literal,8,0,start); long i = Long.parseLong(literal); return TypeInteger.constant(i); } @@ -262,6 +263,11 @@ public class Lexer { } private String parseNumberString() { int old = _position; + if( isHex() ) { + _position += 2; + while( isIdLetter(peek()) ) inc(); + return new String(_input,old,_position-old); + } int len = Math.abs(isLongOrDouble()); if( peek()=='f' ) { nextChar(); @@ -270,6 +276,24 @@ public class Lexer { return new String(_input,old,len); } + private boolean isHex() { + return peek()=='0' && _position+1<_input.length && (peek(1)=='x' || peek(1)=='X'); + } + + private TypeInteger parseInteger(String literal, int radix, int prefix, Lexer start) { + String kind = radix==16 ? "Hexadecimal" : "Octal"; + if( literal.length()==prefix ) + throw Parser.error(kind+" integer literal requires at least one digit",start); + for( int i=prefix; i<literal.length(); i++ ) + if( Character.digit(literal.charAt(i),radix)<0 ) + throw Parser.error("Invalid "+kind.toLowerCase()+" integer literal '"+literal+"'",start); + try { + return TypeInteger.constant(Long.parseUnsignedLong(literal.substring(prefix),radix)); + } catch( NumberFormatException ex ) { + throw Parser.error(kind+" integer literal exceeds 64 bits",start); + } + } + // Return +len that ends a long // Return -len that ends a double private int isLongOrDouble() { diff --git a/src/main/java/com/seaofnodes/simple/parser/README.md b/src/main/java/com/seaofnodes/simple/parser/README.md index cee1d64a..d0220761 100644 --- a/src/main/java/com/seaofnodes/simple/parser/README.md +++ b/src/main/java/com/seaofnodes/simple/parser/README.md @@ -2,7 +2,7 @@ The parser builds Sea-of-Nodes IR directly. Each component shares the cursor, active scope, and compilation state through its `Parser P` reference; there is no intermediate -AST. The [grammar map](../../../../../../../docs/chapter30/30-grammar.md) connects language +AST. The [grammar map](../../../../../../../docs/chapter31/31-grammar.md) connects language productions to their entry points. | File | Responsibility | diff --git a/src/main/java/com/seaofnodes/simple/parser/package-info.java b/src/main/java/com/seaofnodes/simple/parser/package-info.java index c886f72a..73ebec2e 100644 --- a/src/main/java/com/seaofnodes/simple/parser/package-info.java +++ b/src/main/java/com/seaofnodes/simple/parser/package-info.java @@ -7,7 +7,7 @@ * use-side name/field resolution; Allocations, ControlFlow, and Conversions * contain their shared IR construction. Lexer also supplies location snapshots. * - * <p>See {@code chapter30/docs/30-grammar.md} for the production-to-method map. + * <p>See {@code chapter31/docs/31-grammar.md} for the production-to-method map. * Evaluation order and scope lifetimes are explicit: loop updates and declared * constructor bodies are the two deferred source fragments. There is no AST. */ |