// This also implements the preprocessor
//
// Author: Miguel de Icaza (miguel@gnu.org)
-// Marek Safar (marek.safar@seznam.cz)
+// Marek Safar (marek.safar@gmail.com)
//
// Dual licensed under the terms of the MIT X11 or GNU GPL
//
// Copyright 2001, 2002 Ximian, Inc (http://www.ximian.com)
// Copyright 2004-2008 Novell, Inc
-//
+// Copyright 2011 Xamarin, Inc (http://www.xamarin.com)
//
using System;
using System.Text;
-using System.Collections;
-using System.IO;
+using System.Collections.Generic;
using System.Globalization;
-using System.Reflection;
+using System.Diagnostics;
namespace Mono.CSharp
{
public class Tokenizer : yyParser.yyInput
{
+ class KeywordEntry<T>
+ {
+ public readonly T Token;
+ public KeywordEntry<T> Next;
+ public readonly char[] Value;
+
+ public KeywordEntry (string value, T token)
+ {
+ this.Value = value.ToCharArray ();
+ this.Token = token;
+ }
+ }
+
+ sealed class IdentifiersComparer : IEqualityComparer<char[]>
+ {
+ readonly int length;
+
+ public IdentifiersComparer (int length)
+ {
+ this.length = length;
+ }
+
+ public bool Equals (char[] x, char[] y)
+ {
+ for (int i = 0; i < length; ++i)
+ if (x [i] != y [i])
+ return false;
+
+ return true;
+ }
+
+ public int GetHashCode (char[] obj)
+ {
+ int h = 0;
+ for (int i = 0; i < length; ++i)
+ h = (h << 5) - h + obj [i];
+
+ return h;
+ }
+ }
+
+ //
+ // This class has to be used in the parser only, it reuses token
+ // details after each parse
+ //
+ public class LocatedToken
+ {
+ int row, column;
+ string value;
+
+ static LocatedToken[] buffer;
+ static int pos;
+
+ private LocatedToken ()
+ {
+ }
+
+ public static LocatedToken Create (int row, int column)
+ {
+ return Create (null, row, column);
+ }
+
+ public static LocatedToken Create (string value, Location loc)
+ {
+ return Create (value, loc.Row, loc.Column);
+ }
+
+ public static LocatedToken Create (string value, int row, int column)
+ {
+ //
+ // TODO: I am not very happy about the logic but it's the best
+ // what I could come up with for now.
+ // Ideally we should be using just tiny buffer (256 elements) which
+ // is enough to hold all details for currect stack and recycle elements
+ // poped from the stack but there is a trick needed to recycle
+ // them properly.
+ //
+ LocatedToken entry;
+ if (pos >= buffer.Length) {
+ entry = new LocatedToken ();
+ } else {
+ entry = buffer [pos];
+ if (entry == null) {
+ entry = new LocatedToken ();
+ buffer [pos] = entry;
+ }
+
+ ++pos;
+ }
+ entry.value = value;
+ entry.row = row;
+ entry.column = column;
+ return entry;
+ }
+
+ //
+ // Used for token not required by expression evaluator
+ //
+ [Conditional ("FULL_AST")]
+ public static void CreateOptional (int row, int col, ref object token)
+ {
+ token = Create (row, col);
+ }
+
+ public static void Initialize ()
+ {
+ if (buffer == null)
+ buffer = new LocatedToken [10000];
+ pos = 0;
+ }
+
+ public Location Location {
+ get { return new Location (row, column); }
+ }
+
+ public string Value {
+ get { return value; }
+ }
+ }
+
+ public enum PreprocessorDirective
+ {
+ Invalid = 0,
+
+ Region = 1,
+ Endregion = 2,
+ If = 3 | RequiresArgument,
+ Endif = 4,
+ Elif = 5 | RequiresArgument,
+ Else = 6,
+ Define = 7 | RequiresArgument,
+ Undef = 8 | RequiresArgument,
+ Error = 9,
+ Warning = 10,
+ Pragma = 11 | CustomArgumentsParsing,
+ Line = 12,
+
+ CustomArgumentsParsing = 1 << 10,
+ RequiresArgument = 1 << 11
+ }
+
SeekableStreamReader reader;
SourceFile ref_name;
- CompilationUnit file_name;
+ CompilationSourceFile file_name;
+ CompilerContext context;
bool hidden = false;
int ref_line = 1;
int line = 1;
int col = 0;
int previous_col;
int current_token;
+ int tab_size;
bool handle_get_set = false;
bool handle_remove_add = false;
bool handle_where = false;
bool handle_typeof = false;
bool lambda_arguments_parsing;
- Location current_comment_location = Location.Null;
- ArrayList escaped_identifiers;
+ List<Location> escaped_identifiers;
int parsing_generic_less_than;
+ readonly bool doc_processing;
//
// Used mainly for parser optimizations. Some expressions for instance
// Set when parsing generic declaration (type or method header)
//
public bool parsing_generic_declaration;
+ public bool parsing_generic_declaration_doc;
//
// The value indicates that we have not reach any declaration or
//
public int parsing_declaration;
+ public bool parsing_attribute_section;
+
+ public bool parsing_modifiers;
+
//
- // The special character to inject on streams to trigger the EXPRESSION_PARSE
- // token to be returned. It just happens to be a Unicode character that
- // would never be part of a program (can not be an identifier).
+ // The special characters to inject on streams to run the unit parser
+ // in the special expression mode. Using private characters from
+ // Plane Sixteen (U+100000 to U+10FFFD)
//
// This character is only tested just before the tokenizer is about to report
// an error; So on the regular operation mode, this addition will have no
// impact on the tokenizer's performance.
//
- public const int EvalStatementParserCharacter = 0x2190; // Unicode Left Arrow
- public const int EvalCompilationUnitParserCharacter = 0x2191; // Unicode Arrow
- public const int EvalUsingDeclarationsParserCharacter = 0x2192; // Unicode Arrow
+ public const int EvalStatementParserCharacter = 0x100000;
+ public const int EvalCompilationUnitParserCharacter = 0x100001;
+ public const int EvalUsingDeclarationsParserCharacter = 0x100002;
+ public const int DocumentationXref = 0x100003;
//
// XML documentation buffer. The save point is used to divide
//
bool tokens_seen = false;
+ //
+ // Set to true once the GENERATE_COMPLETION token has bee
+ // returned. This helps produce one GENERATE_COMPLETION,
+ // as many COMPLETE_COMPLETION as necessary to complete the
+ // AST tree and one final EOF.
+ //
+ bool generated;
+
//
// Whether a token has been seen on the file
// This is needed because `define' is not allowed to be used
// after a token has been seen.
//
- bool any_token_seen = false;
+ bool any_token_seen;
+
+ //
+ // Class variables
+ //
+ static readonly KeywordEntry<int>[][] keywords;
+ static readonly KeywordEntry<PreprocessorDirective>[][] keywords_preprocessor;
+ static readonly Dictionary<string, object> keyword_strings; // TODO: HashSet
+ static readonly NumberStyles styles;
+ static readonly NumberFormatInfo csharp_format_info;
+
+ // Pragma arguments
+ static readonly char[] pragma_warning = "warning".ToCharArray ();
+ static readonly char[] pragma_warning_disable = "disable".ToCharArray ();
+ static readonly char[] pragma_warning_restore = "restore".ToCharArray ();
+ static readonly char[] pragma_checksum = "checksum".ToCharArray ();
static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
get { return handle_typeof; }
set { handle_typeof = value; }
}
+
+ public int TabSize {
+ get { return tab_size; }
+ set { tab_size = value; }
+ }
public XmlCommentState doc_state {
get { return xml_doc_state; }
}
}
- void AddEscapedIdentifier (LocatedToken lt)
+ //
+ // This is used to trigger completion generation on the parser
+ public bool CompleteOnEOF;
+
+ void AddEscapedIdentifier (Location loc)
{
if (escaped_identifiers == null)
- escaped_identifiers = new ArrayList ();
+ escaped_identifiers = new List<Location> ();
- escaped_identifiers.Add (lt);
+ escaped_identifiers.Add (loc);
}
- public bool IsEscapedIdentifier (Location loc)
+ public bool IsEscapedIdentifier (ATypeNameExpression name)
{
- if (escaped_identifiers != null) {
- foreach (LocatedToken lt in escaped_identifiers)
- if (lt.Location.Equals (loc))
- return true;
- }
-
- return false;
+ return escaped_identifiers != null && escaped_identifiers.Contains (name.Location);
}
- //
- // Class variables
- //
- static CharArrayHashtable[] keywords;
- static Hashtable keyword_strings;
- static NumberStyles styles;
- static NumberFormatInfo csharp_format_info;
-
//
// Values for the associated token returned
//
internal int putback_char; // Used by repl only
- Object val;
+ object val;
//
// Pre-processor
//
// pre-processor if stack state:
//
- Stack ifstack;
+ Stack<int> ifstack;
static System.Text.StringBuilder string_builder;
const int max_id_size = 512;
- static char [] id_builder = new char [max_id_size];
+ static readonly char [] id_builder = new char [max_id_size];
- static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
+ public static Dictionary<char[], string>[] identifiers = new Dictionary<char[], string>[max_id_size + 1];
const int max_number_size = 512;
static char [] number_builder = new char [max_number_size];
static int number_pos;
-
- //
- // Details about the error encoutered by the tokenizer
- //
- string error_details;
-
- public string error {
- get {
- return error_details;
- }
- }
-
+
+ static char[] value_builder = new char[256];
+
public int Line {
get {
return ref_line;
// on its own to deamiguate a token in behalf of the
// parser.
//
- Stack position_stack = new Stack (2);
+ Stack<Position> position_stack = new Stack<Position> (2);
+
class Position {
public int position;
public int line;
public bool hidden;
public int putback_char;
public int previous_col;
- public Stack ifstack;
+ public Stack<int> ifstack;
public int parsing_generic_less_than;
public int current_token;
+ public object val;
public Position (Tokenizer t)
{
hidden = t.hidden;
putback_char = t.putback_char;
previous_col = t.previous_col;
- if (t.ifstack != null && t.ifstack.Count != 0)
- ifstack = (Stack)t.ifstack.Clone ();
+ if (t.ifstack != null && t.ifstack.Count != 0) {
+ // There is no simple way to clone Stack<T> all
+ // methods reverse the order
+ var clone = t.ifstack.ToArray ();
+ Array.Reverse (clone);
+ ifstack = new Stack<int> (clone);
+ }
parsing_generic_less_than = t.parsing_generic_less_than;
current_token = t.current_token;
+ val = t.val;
}
}
+
+ public Tokenizer (SeekableStreamReader input, CompilationSourceFile file, CompilerContext ctx)
+ {
+ this.ref_name = file;
+ this.file_name = file;
+ this.context = ctx;
+ reader = input;
+
+ putback_char = -1;
+
+ xml_comment_buffer = new StringBuilder ();
+ doc_processing = ctx.Settings.DocumentationFile != null;
+
+ if (Environment.OSVersion.Platform == PlatformID.Win32NT)
+ tab_size = 4;
+ else
+ tab_size = 8;
+
+ Mono.CSharp.Location.Push (file, file);
+ }
public void PushPosition ()
{
public void PopPosition ()
{
- Position p = (Position) position_stack.Pop ();
+ Position p = position_stack.Pop ();
reader.Position = p.position;
ref_line = p.ref_line;
ifstack = p.ifstack;
parsing_generic_less_than = p.parsing_generic_less_than;
current_token = p.current_token;
+ val = p.val;
}
// Do not reset the position, ignore it.
static void AddKeyword (string kw, int token)
{
- keyword_strings.Add (kw, kw);
- if (keywords [kw.Length] == null) {
- keywords [kw.Length] = new CharArrayHashtable (kw.Length);
+ keyword_strings.Add (kw, null);
+
+ AddKeyword (keywords, kw, token);
+ }
+
+ static void AddPreprocessorKeyword (string kw, PreprocessorDirective directive)
+ {
+ AddKeyword (keywords_preprocessor, kw, directive);
+ }
+
+ static void AddKeyword<T> (KeywordEntry<T>[][] keywords, string kw, T token)
+ {
+ int length = kw.Length;
+ if (keywords[length] == null) {
+ keywords[length] = new KeywordEntry<T>['z' - '_' + 1];
}
- keywords [kw.Length] [kw.ToCharArray ()] = token;
+
+ int char_index = kw[0] - '_';
+ var kwe = keywords[length][char_index];
+ if (kwe == null) {
+ keywords[length][char_index] = new KeywordEntry<T> (kw, token);
+ return;
+ }
+
+ while (kwe.Next != null) {
+ kwe = kwe.Next;
+ }
+
+ kwe.Next = new KeywordEntry<T> (kw, token);
}
- static void InitTokens ()
+ //
+ // Class initializer
+ //
+ static Tokenizer ()
{
- keyword_strings = new Hashtable ();
- keywords = new CharArrayHashtable [64];
+ keyword_strings = new Dictionary<string, object> ();
+
+ // 11 is the length of the longest keyword for now
+ keywords = new KeywordEntry<int>[11][];
AddKeyword ("__arglist", Token.ARGLIST);
+ AddKeyword ("__makeref", Token.MAKEREF);
+ AddKeyword ("__reftype", Token.REFTYPE);
+ AddKeyword ("__refvalue", Token.REFVALUE);
AddKeyword ("abstract", Token.ABSTRACT);
AddKeyword ("as", Token.AS);
AddKeyword ("add", Token.ADD);
AddKeyword ("ascending", Token.ASCENDING);
AddKeyword ("descending", Token.DESCENDING);
AddKeyword ("into", Token.INTO);
- }
- //
- // Class initializer
- //
- static Tokenizer ()
- {
- Reset ();
- }
+ // Contextual async keywords
+ AddKeyword ("async", Token.ASYNC);
+ AddKeyword ("await", Token.AWAIT);
+
+ keywords_preprocessor = new KeywordEntry<PreprocessorDirective>[10][];
+
+ AddPreprocessorKeyword ("region", PreprocessorDirective.Region);
+ AddPreprocessorKeyword ("endregion", PreprocessorDirective.Endregion);
+ AddPreprocessorKeyword ("if", PreprocessorDirective.If);
+ AddPreprocessorKeyword ("endif", PreprocessorDirective.Endif);
+ AddPreprocessorKeyword ("elif", PreprocessorDirective.Elif);
+ AddPreprocessorKeyword ("else", PreprocessorDirective.Else);
+ AddPreprocessorKeyword ("define", PreprocessorDirective.Define);
+ AddPreprocessorKeyword ("undef", PreprocessorDirective.Undef);
+ AddPreprocessorKeyword ("error", PreprocessorDirective.Error);
+ AddPreprocessorKeyword ("warning", PreprocessorDirective.Warning);
+ AddPreprocessorKeyword ("pragma", PreprocessorDirective.Pragma);
+ AddPreprocessorKeyword ("line", PreprocessorDirective.Line);
- public static void Reset ()
- {
- InitTokens ();
csharp_format_info = NumberFormatInfo.InvariantInfo;
styles = NumberStyles.Float;
int GetKeyword (char[] id, int id_len)
{
- /*
- * Keywords are stored in an array of hashtables grouped by their
- * length.
- */
+ //
+ // Keywords are stored in an array of arrays grouped by their
+ // length and then by the first character
+ //
+ if (id_len >= keywords.Length || keywords [id_len] == null)
+ return -1;
- if ((id_len >= keywords.Length) || (keywords [id_len] == null))
+ int first_index = id [0] - '_';
+ if (first_index > 'z' - '_')
return -1;
- object o = keywords [id_len] [id];
- if (o == null)
+ var kwe = keywords [id_len] [first_index];
+ if (kwe == null)
+ return -1;
+
+ int res;
+ do {
+ res = kwe.Token;
+ for (int i = 1; i < id_len; ++i) {
+ if (id [i] != kwe.Value [i]) {
+ res = 0;
+ kwe = kwe.Next;
+ break;
+ }
+ }
+ } while (res == 0 && kwe != null);
+
+ if (res == 0)
return -1;
int next_token;
- int res = (int) o;
switch (res) {
case Token.GET:
case Token.SET:
res = Token.FROM_FIRST;
query_parsing = true;
- if (RootContext.Version <= LanguageVersion.ISO_2)
- Report.FeatureIsNotAvailable (Location, "query expressions");
+ if (context.Settings.Version <= LanguageVersion.ISO_2)
+ Report.FeatureIsNotAvailable (context, Location, "query expressions");
break;
case Token.VOID:
- Expression.Error_VoidInvalidInTheContext (Location);
+ Expression.Error_VoidInvalidInTheContext (Location, Report);
break;
default:
PopPosition ();
if (ok) {
if (next_token == Token.VOID) {
- if (RootContext.Version == LanguageVersion.ISO_1 ||
- RootContext.Version == LanguageVersion.ISO_2)
- Report.FeatureIsNotAvailable (Location, "partial methods");
- } else if (RootContext.Version == LanguageVersion.ISO_1)
- Report.FeatureIsNotAvailable (Location, "partial types");
+ if (context.Settings.Version <= LanguageVersion.ISO_2)
+ Report.FeatureIsNotAvailable (context, Location, "partial methods");
+ } else if (context.Settings.Version == LanguageVersion.ISO_1)
+ Report.FeatureIsNotAvailable (context, Location, "partial types");
return res;
}
res = -1;
break;
+
+ case Token.ASYNC:
+ if (parsing_modifiers) {
+ //
+ // Skip attributes section or constructor called async
+ //
+ if (parsing_attribute_section || peek_token () == Token.OPEN_PARENS) {
+ res = -1;
+ } else {
+ // async is keyword
+ }
+ } else if (parsing_block > 0) {
+ switch (peek_token ()) {
+ case Token.DELEGATE:
+ case Token.OPEN_PARENS_LAMBDA:
+ // async is keyword
+ break;
+ case Token.IDENTIFIER:
+ PushPosition ();
+ xtoken ();
+ if (xtoken () != Token.ARROW)
+ res = -1;
+
+ PopPosition ();
+ break;
+ default:
+ res = -1;
+ break;
+ }
+ } else {
+ res = -1;
+ }
+
+ if (res == Token.ASYNC && context.Settings.Version <= LanguageVersion.V_4) {
+ Report.FeatureIsNotAvailable (context, Location, "asynchronous functions");
+ }
+
+ break;
+
+ case Token.AWAIT:
+ if (parsing_block == 0)
+ res = -1;
+
+ break;
}
+
+ return res;
+ }
+
+ static PreprocessorDirective GetPreprocessorDirective (char[] id, int id_len)
+ {
+ //
+ // Keywords are stored in an array of arrays grouped by their
+ // length and then by the first character
+ //
+ if (id_len >= keywords_preprocessor.Length || keywords_preprocessor[id_len] == null)
+ return PreprocessorDirective.Invalid;
+
+ int first_index = id[0] - '_';
+ if (first_index > 'z' - '_')
+ return PreprocessorDirective.Invalid;
+
+ var kwe = keywords_preprocessor[id_len][first_index];
+ if (kwe == null)
+ return PreprocessorDirective.Invalid;
+
+ PreprocessorDirective res = PreprocessorDirective.Invalid;
+ do {
+ res = kwe.Token;
+ for (int i = 1; i < id_len; ++i) {
+ if (id[i] != kwe.Value[i]) {
+ res = 0;
+ kwe = kwe.Next;
+ break;
+ }
+ }
+ } while (res == PreprocessorDirective.Invalid && kwe != null);
+
return res;
}
}
}
- public Tokenizer (SeekableStreamReader input, CompilationUnit file)
- {
- this.ref_name = file;
- this.file_name = file;
- reader = input;
-
- putback_char = -1;
-
- xml_comment_buffer = new StringBuilder ();
-
- //
- // FIXME: This could be `Location.Push' but we have to
- // find out why the MS compiler allows this
- //
- Mono.CSharp.Location.Push (file, file);
- }
-
static bool is_identifier_start_character (int c)
{
return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter ((char)c);
static bool is_identifier_part_character (char c)
{
- return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9') ||
- Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
+ if (c >= 'a' && c <= 'z')
+ return true;
+
+ if (c >= 'A' && c <= 'Z')
+ return true;
+
+ if (c == '_' || (c >= '0' && c <= '9'))
+ return true;
+
+ if (c < 0x80)
+ return false;
+
+ return Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
}
public static bool IsKeyword (string s)
{
- return keyword_strings [s] != null;
+ return keyword_strings.ContainsKey (s);
}
//
// Open parens micro parser. Detects both lambda and cast ambiguity.
- //
-
+ //
int TokenizeOpenParens ()
{
int ptoken;
//
// Expression inside parens is lambda, (int i) =>
//
- if (current_token == Token.ARROW) {
- if (RootContext.Version <= LanguageVersion.ISO_2)
- Report.FeatureIsNotAvailable (Location, "lambda expressions");
-
+ if (current_token == Token.ARROW)
return Token.OPEN_PARENS_LAMBDA;
- }
//
// Expression inside parens is single type, (int[])
case Token.BANG:
case Token.TILDE:
case Token.IDENTIFIER:
- case Token.LITERAL_INTEGER:
- case Token.LITERAL_FLOAT:
- case Token.LITERAL_DOUBLE:
- case Token.LITERAL_DECIMAL:
- case Token.LITERAL_CHARACTER:
- case Token.LITERAL_STRING:
+ case Token.LITERAL:
case Token.BASE:
case Token.CHECKED:
case Token.DELEGATE:
case Token.IDENTIFIER:
switch (ptoken) {
case Token.DOT:
+ if (bracket_level == 0) {
+ is_type = false;
+ can_be_type = true;
+ }
+
+ continue;
case Token.OP_GENERICS_LT:
case Token.COMMA:
case Token.DOUBLE_COLON:
the_token = token ();
} while (the_token != Token.CLOSE_BRACKET);
the_token = token ();
+ } else if (the_token == Token.IN || the_token == Token.OUT) {
+ the_token = token ();
}
switch (the_token) {
case Token.IDENTIFIER:
case Token.CHAR:
case Token.VOID:
break;
-
case Token.OP_GENERICS_GT:
+ case Token.IN:
+ case Token.OUT:
return true;
default:
case Token.TRUE:
case Token.FALSE:
case Token.NULL:
- case Token.LITERAL_INTEGER:
- case Token.LITERAL_STRING:
+ case Token.LITERAL:
return Token.INTERR;
}
current_token = Token.NONE;
int next_token;
switch (xtoken ()) {
- case Token.LITERAL_INTEGER:
- case Token.LITERAL_STRING:
- case Token.LITERAL_CHARACTER:
- case Token.LITERAL_DECIMAL:
- case Token.LITERAL_DOUBLE:
- case Token.LITERAL_FLOAT:
+ case Token.LITERAL:
case Token.TRUE:
case Token.FALSE:
case Token.NULL:
case Token.CLOSE_PARENS:
case Token.OPEN_BRACKET:
case Token.OP_GENERICS_GT:
+ case Token.INTERR:
next_token = Token.INTERR_NULLABLE;
break;
return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
}
- static int real_type_suffix (int c)
+ static TypeCode real_type_suffix (int c)
{
- int t;
-
switch (c){
case 'F': case 'f':
- t = Token.LITERAL_FLOAT;
- break;
+ return TypeCode.Single;
case 'D': case 'd':
- t = Token.LITERAL_DOUBLE;
- break;
+ return TypeCode.Double;
case 'M': case 'm':
- t= Token.LITERAL_DECIMAL;
- break;
+ return TypeCode.Decimal;
default:
- return Token.NONE;
+ return TypeCode.Empty;
}
- return t;
}
- int integer_type_suffix (ulong ul, int c)
+ ILiteralConstant integer_type_suffix (ulong ul, int c, Location loc)
{
bool is_unsigned = false;
bool is_long = false;
//
Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
}
- //
- // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
- // work around that.
- //
- //goto case 'L';
- if (is_long)
- scanning = false;
- is_long = true;
- get_char ();
- break;
+
+ goto case 'L';
case 'L':
if (is_long)
}
if (is_long && is_unsigned){
- val = ul;
- return Token.LITERAL_INTEGER;
- } else if (is_unsigned){
+ return new ULongLiteral (context.BuiltinTypes, ul, loc);
+ }
+
+ if (is_unsigned){
// uint if possible, or ulong else.
if ((ul & 0xffffffff00000000) == 0)
- val = (uint) ul;
+ return new UIntLiteral (context.BuiltinTypes, (uint) ul, loc);
else
- val = ul;
+ return new ULongLiteral (context.BuiltinTypes, ul, loc);
} else if (is_long){
// long if possible, ulong otherwise
if ((ul & 0x8000000000000000) != 0)
- val = ul;
+ return new ULongLiteral (context.BuiltinTypes, ul, loc);
else
- val = (long) ul;
+ return new LongLiteral (context.BuiltinTypes, (long) ul, loc);
} else {
// int, uint, long or ulong in that order
if ((ul & 0xffffffff00000000) == 0){
uint ui = (uint) ul;
if ((ui & 0x80000000) != 0)
- val = ui;
+ return new UIntLiteral (context.BuiltinTypes, ui, loc);
else
- val = (int) ui;
+ return new IntLiteral (context.BuiltinTypes, (int) ui, loc);
} else {
if ((ul & 0x8000000000000000) != 0)
- val = ul;
+ return new ULongLiteral (context.BuiltinTypes, ul, loc);
else
- val = (long) ul;
+ return new LongLiteral (context.BuiltinTypes, (long) ul, loc);
}
}
- return Token.LITERAL_INTEGER;
}
//
// we need to convert to a special type, and then choose
// the best representation for the integer
//
- int adjust_int (int c)
+ ILiteralConstant adjust_int (int c, Location loc)
{
try {
if (number_pos > 9){
for (int i = 1; i < number_pos; i++){
ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
}
- return integer_type_suffix (ul, c);
+
+ return integer_type_suffix (ul, c, loc);
} else {
uint ui = (uint) (number_builder [0] - '0');
for (int i = 1; i < number_pos; i++){
ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
}
- return integer_type_suffix (ui, c);
+
+ return integer_type_suffix (ui, c, loc);
}
} catch (OverflowException) {
- error_details = "Integral constant is too large";
- Report.Error (1021, Location, error_details);
- val = 0ul;
- return Token.LITERAL_INTEGER;
+ Error_NumericConstantTooLong ();
+ return new IntLiteral (context.BuiltinTypes, 0, loc);
}
catch (FormatException) {
Report.Error (1013, Location, "Invalid number");
- val = 0ul;
- return Token.LITERAL_INTEGER;
+ return new IntLiteral (context.BuiltinTypes, 0, loc);
}
}
- int adjust_real (int t)
+ ILiteralConstant adjust_real (TypeCode t, Location loc)
{
- string s = new String (number_builder, 0, number_pos);
+ string s = new string (number_builder, 0, number_pos);
const string error_details = "Floating-point constant is outside the range of type `{0}'";
switch (t){
- case Token.LITERAL_DECIMAL:
+ case TypeCode.Decimal:
try {
- val = System.Decimal.Parse (s, styles, csharp_format_info);
+ return new DecimalLiteral (context.BuiltinTypes, decimal.Parse (s, styles, csharp_format_info), loc);
} catch (OverflowException) {
- val = 0m;
Report.Error (594, Location, error_details, "decimal");
+ return new DecimalLiteral (context.BuiltinTypes, 0, loc);
}
- break;
- case Token.LITERAL_FLOAT:
+ case TypeCode.Single:
try {
- val = float.Parse (s, styles, csharp_format_info);
+ return new FloatLiteral (context.BuiltinTypes, float.Parse (s, styles, csharp_format_info), loc);
} catch (OverflowException) {
- val = 0.0f;
Report.Error (594, Location, error_details, "float");
+ return new FloatLiteral (context.BuiltinTypes, 0, loc);
}
- break;
-
- case Token.LITERAL_DOUBLE:
- case Token.NONE:
- t = Token.LITERAL_DOUBLE;
+ default:
try {
- val = System.Double.Parse (s, styles, csharp_format_info);
+ return new DoubleLiteral (context.BuiltinTypes, double.Parse (s, styles, csharp_format_info), loc);
} catch (OverflowException) {
- val = 0.0;
- Report.Error (594, Location, error_details, "double");
+ Report.Error (594, loc, error_details, "double");
+ return new DoubleLiteral (context.BuiltinTypes, 0, loc);
}
- break;
}
- return t;
}
- int handle_hex ()
+ ILiteralConstant handle_hex (Location loc)
{
int d;
ulong ul;
}
string s = new String (number_builder, 0, number_pos);
+
try {
if (number_pos <= 8)
ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
else
ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
+
+ return integer_type_suffix (ul, peek_char (), loc);
} catch (OverflowException){
- error_details = "Integral constant is too large";
- Report.Error (1021, Location, error_details);
- val = 0ul;
- return Token.LITERAL_INTEGER;
+ Error_NumericConstantTooLong ();
+ return new IntLiteral (context.BuiltinTypes, 0, loc);
}
catch (FormatException) {
Report.Error (1013, Location, "Invalid number");
- val = 0ul;
- return Token.LITERAL_INTEGER;
+ return new IntLiteral (context.BuiltinTypes, 0, loc);
}
-
- return integer_type_suffix (ul, peek_char ());
}
//
//
int is_number (int c)
{
- bool is_real = false;
- int type;
+ ILiteralConstant res;
+#if FULL_AST
+ int read_start = reader.Position - 1;
+#endif
number_pos = 0;
+ var loc = Location;
if (c >= '0' && c <= '9'){
if (c == '0'){
int peek = peek_char ();
- if (peek == 'x' || peek == 'X')
- return handle_hex ();
+ if (peek == 'x' || peek == 'X') {
+ val = res = handle_hex (loc);
+#if FULL_AST
+ res.ParsedValue = reader.ReadChars (read_start, reader.Position - 1);
+#endif
+
+ return Token.LITERAL;
+ }
}
decimal_digits (c);
c = get_char ();
// We need to handle the case of
// "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
//
+ bool is_real = false;
if (c == '.'){
if (decimal_digits ('.')){
is_real = true;
} else {
putback ('.');
number_pos--;
- return adjust_int (-1);
+ val = res = adjust_int (-1, loc);
+
+#if FULL_AST
+ res.ParsedValue = reader.ReadChars (read_start, reader.Position - 1);
+#endif
+ return Token.LITERAL;
}
}
is_real = true;
if (number_pos == max_number_size)
Error_NumericConstantTooLong ();
- number_builder [number_pos++] = 'e';
+ number_builder [number_pos++] = (char) c;
c = get_char ();
if (c == '+'){
c = get_char ();
}
- type = real_type_suffix (c);
- if (type == Token.NONE && !is_real){
+ var type = real_type_suffix (c);
+ if (type == TypeCode.Empty && !is_real) {
putback (c);
- return adjust_int (c);
- } else
+ res = adjust_int (c, loc);
+ } else {
is_real = true;
- if (type == Token.NONE){
- putback (c);
+ if (type == TypeCode.Empty) {
+ putback (c);
+ }
+
+ res = adjust_real (type, loc);
}
-
- if (is_real)
- return adjust_real (type);
- Console.WriteLine ("This should not be reached");
- throw new Exception ("Is Number should never reach this point");
+ val = res;
+
+#if FULL_AST
+ res.ParsedValue = reader.ReadChars (read_start, reader.Position - (type == TypeCode.Empty ? 1 : 0));
+#endif
+
+ return Token.LITERAL;
}
//
if (putback_char != -1) {
x = putback_char;
putback_char = -1;
- } else
+ } else {
x = reader.Read ();
- if (x == '\n') {
+ }
+
+ if (x == '\r') {
+ if (peek_char () == '\n') {
+ putback_char = -1;
+ }
+
+ x = '\n';
+ advance_line ();
+ } else if (x == '\n') {
advance_line ();
} else {
col++;
public bool advance ()
{
- return peek_char () != -1;
+ return peek_char () != -1 || CompleteOnEOF;
}
public Object Value {
return current_token;
}
- static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
-
- void get_cmd_arg (out string cmd, out string arg)
+ int TokenizePreprocessorIdentifier (out int c)
{
- int c;
-
- tokens_seen = false;
- arg = "";
-
// skip over white space
do {
c = get_char ();
- } while (c == '\r' || c == ' ' || c == '\t');
+ } while (c == ' ' || c == '\t');
+
- static_cmd_arg.Length = 0;
- while (c != -1 && is_identifier_part_character ((char)c)) {
- static_cmd_arg.Append ((char)c);
+ int pos = 0;
+ while (c != -1 && c >= 'a' && c <= 'z') {
+ id_builder[pos++] = (char) c;
c = get_char ();
if (c == '\\') {
int peek = peek_char ();
int surrogate;
c = EscapeUnicode (c, out surrogate);
if (surrogate != 0) {
- if (is_identifier_part_character ((char) c))
- static_cmd_arg.Append ((char) c);
+ if (is_identifier_part_character ((char) c)) {
+ id_builder[pos++] = (char) c;
+ }
c = surrogate;
}
}
}
}
- cmd = static_cmd_arg.ToString ();
+ return pos;
+ }
+
+ PreprocessorDirective get_cmd_arg (out string arg)
+ {
+ int c;
+
+ tokens_seen = false;
+ arg = "";
+
+ var cmd = GetPreprocessorDirective (id_builder, TokenizePreprocessorIdentifier (out c));
+
+ if ((cmd & PreprocessorDirective.CustomArgumentsParsing) != 0)
+ return cmd;
// skip over white space
- while (c == '\r' || c == ' ' || c == '\t')
+ while (c == ' ' || c == '\t')
c = get_char ();
- static_cmd_arg.Length = 0;
- int has_identifier_argument = 0;
+ int has_identifier_argument = (int)(cmd & PreprocessorDirective.RequiresArgument);
+ int pos = 0;
- while (c != -1 && c != '\n' && c != '\r') {
+ while (c != -1 && c != '\n') {
if (c == '\\' && has_identifier_argument >= 0) {
- if (has_identifier_argument != 0 || (cmd == "define" || cmd == "if" || cmd == "elif" || cmd == "undef")) {
+ if (has_identifier_argument != 0) {
has_identifier_argument = 1;
int peek = peek_char ();
int surrogate;
c = EscapeUnicode (c, out surrogate);
if (surrogate != 0) {
- if (is_identifier_part_character ((char) c))
- static_cmd_arg.Append ((char) c);
+ if (is_identifier_part_character ((char) c)) {
+ if (pos == value_builder.Length)
+ Array.Resize (ref value_builder, pos * 2);
+
+ value_builder[pos++] = (char) c;
+ }
c = surrogate;
}
}
} else {
has_identifier_argument = -1;
}
+ } else if (c == '/' && peek_char () == '/') {
+ //
+ // Eat single-line comments
+ //
+ get_char ();
+ do {
+ c = get_char ();
+ } while (c != -1 && c != '\n');
+
+ break;
}
- static_cmd_arg.Append ((char) c);
+
+ if (pos == value_builder.Length)
+ Array.Resize (ref value_builder, pos * 2);
+
+ value_builder[pos++] = (char) c;
c = get_char ();
}
- if (static_cmd_arg.Length != 0)
- arg = static_cmd_arg.ToString ();
+ if (pos != 0) {
+ if (pos > max_id_size)
+ arg = new string (value_builder, 0, pos);
+ else
+ arg = InternIdentifier (value_builder, pos);
+
+ // Eat any trailing whitespaces
+ arg = arg.Trim (simple_whitespaces);
+ }
+
+ return cmd;
}
//
char [] quotes = { '\"' };
string name = arg.Substring (pos). Trim (quotes);
- ref_name = Location.LookupFile (file_name, name);
- file_name.AddFile (ref_name);
+ ref_name = context.LookupFile (file_name, name);
+ file_name.AddIncludeFile (ref_name);
hidden = false;
Location.Push (file_name, ref_name);
} else {
//
// #define ident
//
- if (RootContext.IsConditionalDefined (ident))
+ if (context.Settings.IsConditionalSymbolDefined (ident))
return;
file_name.AddDefine (ident);
}
}
- static byte read_hex (string arg, int pos, out bool error)
+ byte read_hex (out bool error)
{
- error = false;
-
int total;
- char c = arg [pos];
+ int c = get_char ();
if ((c >= '0') && (c <= '9'))
total = (int) c - (int) '0';
}
total *= 16;
- c = arg [pos+1];
+ c = get_char ();
if ((c >= '0') && (c <= '9'))
total += (int) c - (int) '0';
return 0;
}
+ error = false;
return (byte) total;
}
- /// <summary>
- /// Handles #pragma checksum
- /// </summary>
- bool PreProcessPragmaChecksum (string arg)
+ //
+ // Parses #pragma checksum
+ //
+ bool ParsePragmaChecksum ()
{
- if ((arg [0] != ' ') && (arg [0] != '\t'))
- return false;
+ //
+ // The syntax is ` "foo.txt" "{guid}" "hash"'
+ //
+ int c = get_char ();
- arg = arg.Trim (simple_whitespaces);
- if ((arg.Length < 2) || (arg [0] != '"'))
+ if (c != '"')
return false;
- StringBuilder file_sb = new StringBuilder ();
-
- int pos = 1;
- char ch;
- while ((ch = arg [pos++]) != '"') {
- if (pos >= arg.Length)
- return false;
-
- if (ch == '\\') {
- if (pos+1 >= arg.Length)
- return false;
- ch = arg [pos++];
+ string_builder.Length = 0;
+ while (c != -1 && c != '\n') {
+ c = get_char ();
+ if (c == '"') {
+ c = get_char ();
+ break;
}
- file_sb.Append (ch);
+ string_builder.Append ((char) c);
+ }
+
+ if (string_builder.Length == 0) {
+ Report.Warning (1709, 1, Location, "Filename specified for preprocessor directive is empty");
}
- if ((pos+2 >= arg.Length) || ((arg [pos] != ' ') && (arg [pos] != '\t')))
+ // TODO: Any white-spaces count
+ if (c != ' ')
return false;
- arg = arg.Substring (pos).Trim (simple_whitespaces);
- if ((arg.Length < 42) || (arg [0] != '"') || (arg [1] != '{') ||
- (arg [10] != '-') || (arg [15] != '-') || (arg [20] != '-') ||
- (arg [25] != '-') || (arg [38] != '}') || (arg [39] != '"'))
+ SourceFile file = context.LookupFile (file_name, string_builder.ToString ());
+
+ if (get_char () != '"' || get_char () != '{')
return false;
bool error;
byte[] guid_bytes = new byte [16];
+ int i = 0;
- for (int i = 0; i < 4; i++) {
- guid_bytes [i] = read_hex (arg, 2+2*i, out error);
+ for (; i < 4; i++) {
+ guid_bytes [i] = read_hex (out error);
if (error)
return false;
}
- for (int i = 0; i < 2; i++) {
- guid_bytes [i+4] = read_hex (arg, 11+2*i, out error);
+
+ if (get_char () != '-')
+ return false;
+
+ for (; i < 10; i++) {
+ guid_bytes [i] = read_hex (out error);
if (error)
return false;
- guid_bytes [i+6] = read_hex (arg, 16+2*i, out error);
+
+ guid_bytes [i++] = read_hex (out error);
if (error)
return false;
- guid_bytes [i+8] = read_hex (arg, 21+2*i, out error);
- if (error)
+
+ if (get_char () != '-')
return false;
}
- for (int i = 0; i < 6; i++) {
- guid_bytes [i+10] = read_hex (arg, 26+2*i, out error);
+ for (; i < 16; i++) {
+ guid_bytes [i] = read_hex (out error);
if (error)
return false;
}
- arg = arg.Substring (40).Trim (simple_whitespaces);
- if ((arg.Length < 34) || (arg [0] != '"') || (arg [33] != '"'))
+ if (get_char () != '}' || get_char () != '"')
+ return false;
+
+ // TODO: Any white-spaces count
+ c = get_char ();
+ if (c != ' ')
+ return false;
+
+ if (get_char () != '"')
return false;
- byte[] checksum_bytes = new byte [16];
- for (int i = 0; i < 16; i++) {
- checksum_bytes [i] = read_hex (arg, 1+2*i, out error);
+ // Any length of checksum
+ List<byte> checksum_bytes = new List<byte> (16);
+
+ c = peek_char ();
+ while (c != '"' && c != -1) {
+ checksum_bytes.Add (read_hex (out error));
if (error)
return false;
+
+ c = peek_char ();
}
- arg = arg.Substring (34).Trim (simple_whitespaces);
- if (arg.Length > 0)
+ if (c == '/') {
+ ReadSingleLineComment ();
+ } else if (get_char () != '"') {
return false;
+ }
- SourceFile file = Location.LookupFile (file_name, file_sb.ToString ());
- file.SetChecksum (guid_bytes, checksum_bytes);
+ file.SetChecksum (guid_bytes, checksum_bytes.ToArray ());
ref_name.AutoGenerated = true;
return true;
}
+ bool IsTokenIdentifierEqual (char[] identifier)
+ {
+ for (int i = 0; i < identifier.Length; ++i) {
+ if (identifier[i] != id_builder[i])
+ return false;
+ }
+
+ return true;
+ }
+
+ int TokenizePragmaNumber (ref int c)
+ {
+ number_pos = 0;
+
+ int number;
+
+ if (c >= '0' && c <= '9') {
+ decimal_digits (c);
+ uint ui = (uint) (number_builder[0] - '0');
+
+ try {
+ for (int i = 1; i < number_pos; i++) {
+ ui = checked ((ui * 10) + ((uint) (number_builder[i] - '0')));
+ }
+
+ number = (int) ui;
+ } catch (OverflowException) {
+ Error_NumericConstantTooLong ();
+ number = -1;
+ }
+
+
+ c = get_char ();
+
+ // skip over white space
+ while (c == ' ' || c == '\t')
+ c = get_char ();
+
+ if (c == ',') {
+ c = get_char ();
+ }
+
+ // skip over white space
+ while (c == ' ' || c == '\t')
+ c = get_char ();
+ } else {
+ number = -1;
+ if (c == '/') {
+ ReadSingleLineComment ();
+ } else {
+ Report.Warning (1692, 1, Location, "Invalid number");
+
+ // Read everything till the end of the line or file
+ do {
+ c = get_char ();
+ } while (c != -1 && c != '\n');
+ }
+ }
+
+ return number;
+ }
+
+ void ReadSingleLineComment ()
+ {
+ if (peek_char () != '/')
+ Report.Warning (1696, 1, Location, "Single-line comment or end-of-line expected");
+
+ // Read everything till the end of the line or file
+ int c;
+ do {
+ c = get_char ();
+ } while (c != -1 && c != '\n');
+ }
+
/// <summary>
/// Handles #pragma directive
/// </summary>
- void PreProcessPragma (string arg)
+ void ParsePragmaDirective (string arg)
{
- const string warning = "warning";
- const string w_disable = "warning disable";
- const string w_restore = "warning restore";
- const string checksum = "checksum";
+ int c;
+ int length = TokenizePreprocessorIdentifier (out c);
+ if (length == pragma_warning.Length && IsTokenIdentifierEqual (pragma_warning)) {
+ length = TokenizePreprocessorIdentifier (out c);
- if (arg == w_disable) {
- Report.RegisterWarningRegion (Location).WarningDisable (Location.Row);
- return;
- }
+ //
+ // #pragma warning disable
+ // #pragma warning restore
+ //
+ if (length == pragma_warning_disable.Length) {
+ bool disable = IsTokenIdentifierEqual (pragma_warning_disable);
+ if (disable || IsTokenIdentifierEqual (pragma_warning_restore)) {
+ // skip over white space
+ while (c == ' ' || c == '\t')
+ c = get_char ();
- if (arg == w_restore) {
- Report.RegisterWarningRegion (Location).WarningEnable (Location.Row);
- return;
- }
+ var loc = Location;
- if (arg.StartsWith (w_disable)) {
- int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
- foreach (int code in codes) {
- if (code != 0)
- Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
- }
- return;
- }
+ if (c == '\n' || c == '/') {
+ if (c == '/')
+ ReadSingleLineComment ();
- if (arg.StartsWith (w_restore)) {
- int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
- Hashtable w_table = Report.warning_ignore_table;
- foreach (int code in codes) {
- if (w_table != null && w_table.Contains (code))
- Report.Warning (1635, 1, Location, String.Format ("Cannot restore warning `CS{0:0000}' because it was disabled globally", code));
- Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
+ //
+ // Disable/Restore all warnings
+ //
+ if (disable) {
+ Report.RegisterWarningRegion (loc).WarningDisable (loc.Row);
+ } else {
+ Report.RegisterWarningRegion (loc).WarningEnable (loc.Row);
+ }
+ } else {
+ //
+ // Disable/Restore a warning or group of warnings
+ //
+ int code;
+ do {
+ code = TokenizePragmaNumber (ref c);
+ if (code > 0) {
+ if (disable) {
+ Report.RegisterWarningRegion (loc).WarningDisable (loc, code, Report);
+ } else {
+ Report.RegisterWarningRegion (loc).WarningEnable (loc, code, Report);
+ }
+ }
+ } while (code >= 0 && c != '\n' && c != -1);
+ }
+
+ return;
+ }
}
- return;
- }
- if (arg.StartsWith (warning)) {
Report.Warning (1634, 1, Location, "Expected disable or restore");
return;
}
- if (arg.StartsWith (checksum)) {
- if (!PreProcessPragmaChecksum (arg.Substring (checksum.Length)))
- Warning_InvalidPragmaChecksum ();
+ //
+ // #pragma checksum
+ //
+ if (length == pragma_checksum.Length && IsTokenIdentifierEqual (pragma_checksum)) {
+ if (c != ' ' || !ParsePragmaChecksum ()) {
+ Report.Warning (1695, 1, Location,
+ "Invalid #pragma checksum syntax. Expected \"filename\" \"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
+ }
+
return;
}
Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
}
- int[] ParseNumbers (string text)
- {
- string[] string_array = text.Split (',');
- int[] values = new int [string_array.Length];
- int index = 0;
- foreach (string string_code in string_array) {
- try {
- values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
- }
- catch (FormatException) {
- Report.Warning (1692, 1, Location, "Invalid number");
- }
- }
- return values;
- }
-
bool eval_val (string s)
{
if (s == "true")
if (s == "false")
return false;
- return file_name.IsConditionalDefined (s);
+ return file_name.IsConditionalDefined (context, s);
}
bool pp_primary (ref string s)
void Error_NumericConstantTooLong ()
{
- Report.Error (1021, Location, "Numeric constant too long");
+ Report.Error (1021, Location, "Integral constant is too large");
}
void Error_InvalidDirective ()
{
Report.Error (1025, Location, "Single-line comment or end-of-line expected");
}
-
- void Warning_InvalidPragmaChecksum ()
+
+ //
+ // Raises a warning when tokenizer found documentation comment
+ // on unexpected place
+ //
+ void WarningMisplacedComment (Location loc)
{
- Report.Warning (1695, 1, Location,
- "Invalid #pragma checksum syntax; should be " +
- "#pragma checksum \"filename\" " +
- "\"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
+ if (doc_state != XmlCommentState.Error) {
+ doc_state = XmlCommentState.Error;
+ Report.Warning (1587, 2, loc, "XML comment is not placed on a valid language element");
+ }
}
+
//
// if true, then the code continues processing the code
// if false, the code stays in a loop until another directive is
// reached.
// When caller_is_taking is false we ignore all directives except the ones
// which can help us to identify where the #if block ends
- bool handle_preprocessing_directive (bool caller_is_taking)
+ bool ParsePreprocessingDirective (bool caller_is_taking)
{
- string cmd, arg;
+ string arg;
bool region_directive = false;
- get_cmd_arg (out cmd, out arg);
-
- // Eat any trailing whitespaces and single-line comments
- if (arg.IndexOf ("//") != -1)
- arg = arg.Substring (0, arg.IndexOf ("//"));
- arg = arg.Trim (simple_whitespaces);
+ var directive = get_cmd_arg (out arg);
//
// The first group of pre-processing instructions is always processed
//
- switch (cmd){
- case "region":
+ switch (directive) {
+ case PreprocessorDirective.Region:
region_directive = true;
arg = "true";
- goto case "if";
+ goto case PreprocessorDirective.If;
- case "endregion":
+ case PreprocessorDirective.Endregion:
if (ifstack == null || ifstack.Count == 0){
Error_UnexpectedDirective ("no #region for this #endregion");
return true;
}
- int pop = (int) ifstack.Pop ();
+ int pop = ifstack.Pop ();
if ((pop & REGION) == 0)
Report.Error (1027, Location, "Expected `#endif' directive");
return caller_is_taking;
- case "if":
+ case PreprocessorDirective.If:
if (ifstack == null)
- ifstack = new Stack (2);
+ ifstack = new Stack<int> (2);
int flags = region_directive ? REGION : 0;
if (ifstack.Count == 0){
flags |= PARENT_TAKING;
} else {
- int state = (int) ifstack.Peek ();
+ int state = ifstack.Peek ();
if ((state & TAKING) != 0) {
flags |= PARENT_TAKING;
}
}
- if (caller_is_taking && eval (arg)) {
+ if (eval (arg) && caller_is_taking) {
ifstack.Push (flags | TAKING);
return true;
}
ifstack.Push (flags);
return false;
-
- case "endif":
+
+ case PreprocessorDirective.Endif:
if (ifstack == null || ifstack.Count == 0){
Error_UnexpectedDirective ("no #if for this #endif");
return true;
} else {
- pop = (int) ifstack.Pop ();
+ pop = ifstack.Pop ();
if ((pop & REGION) != 0)
Report.Error (1038, Location, "#endregion directive expected");
if (ifstack.Count == 0)
return true;
- int state = (int) ifstack.Peek ();
+ int state = ifstack.Peek ();
return (state & TAKING) != 0;
}
- case "elif":
+ case PreprocessorDirective.Elif:
if (ifstack == null || ifstack.Count == 0){
Error_UnexpectedDirective ("no #if for this #elif");
return true;
} else {
- int state = (int) ifstack.Pop ();
+ int state = ifstack.Pop ();
if ((state & REGION) != 0) {
Report.Error (1038, Location, "#endregion directive expected");
return false;
}
- case "else":
+ case PreprocessorDirective.Else:
if (ifstack == null || ifstack.Count == 0){
Error_UnexpectedDirective ("no #if for this #else");
return true;
} else {
- int state = (int) ifstack.Peek ();
+ int state = ifstack.Peek ();
if ((state & REGION) != 0) {
Report.Error (1038, Location, "#endregion directive expected");
return ret;
}
- case "define":
+ case PreprocessorDirective.Define:
if (any_token_seen){
Error_TokensSeen ();
return caller_is_taking;
PreProcessDefinition (true, arg, caller_is_taking);
return caller_is_taking;
- case "undef":
+ case PreprocessorDirective.Undef:
if (any_token_seen){
Error_TokensSeen ();
return caller_is_taking;
}
PreProcessDefinition (false, arg, caller_is_taking);
return caller_is_taking;
+
+ case PreprocessorDirective.Invalid:
+ Report.Error (1024, Location, "Wrong preprocessor directive");
+ return true;
}
//
if (!caller_is_taking)
return false;
- switch (cmd){
- case "error":
+ switch (directive){
+ case PreprocessorDirective.Error:
Report.Error (1029, Location, "#error: '{0}'", arg);
return true;
- case "warning":
+ case PreprocessorDirective.Warning:
Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
return true;
- case "pragma":
- if (RootContext.Version == LanguageVersion.ISO_1) {
- Report.FeatureIsNotAvailable (Location, "#pragma");
- return true;
+ case PreprocessorDirective.Pragma:
+ if (context.Settings.Version == LanguageVersion.ISO_1) {
+ Report.FeatureIsNotAvailable (context, Location, "#pragma");
}
- PreProcessPragma (arg);
+ ParsePragmaDirective (arg);
return true;
- case "line":
+ case PreprocessorDirective.Line:
if (!PreProcessLine (arg))
Report.Error (
1576, Location,
return caller_is_taking;
}
- Report.Error (1024, Location, "Wrong preprocessor directive");
- return true;
-
+ throw new NotImplementedException (directive.ToString ());
}
private int consume_string (bool quoted)
{
int c;
- string_builder.Length = 0;
+ int pos = 0;
+ Location start_location = Location;
+ if (quoted)
+ start_location = start_location - 1;
+
+#if FULL_AST
+ int reader_pos = reader.Position;
+#endif
+
+ while (true){
+ c = get_char ();
+ if (c == '"') {
+ if (quoted && peek_char () == '"') {
+ if (pos == value_builder.Length)
+ Array.Resize (ref value_builder, pos * 2);
- while ((c = get_char ()) != -1){
- if (c == '"'){
- if (quoted && peek_char () == '"'){
- string_builder.Append ((char) c);
+ value_builder[pos++] = (char) c;
get_char ();
continue;
- } else {
- val = string_builder.ToString ();
- return Token.LITERAL_STRING;
}
- }
- if (c == '\n'){
- if (!quoted)
- Report.Error (1010, Location, "Newline in constant");
+ string s;
+ if (pos == 0)
+ s = string.Empty;
+ else if (pos <= 4)
+ s = InternIdentifier (value_builder, pos);
+ else
+ s = new string (value_builder, 0, pos);
+
+ ILiteralConstant res = new StringLiteral (context.BuiltinTypes, s, start_location);
+ val = res;
+#if FULL_AST
+ res.ParsedValue = quoted ?
+ reader.ReadChars (reader_pos - 2, reader.Position - 1) :
+ reader.ReadChars (reader_pos - 1, reader.Position);
+#endif
+
+ return Token.LITERAL;
}
- if (!quoted){
+ if (c == '\n') {
+ if (!quoted) {
+ Report.Error (1010, Location, "Newline in constant");
+ val = new StringLiteral (context.BuiltinTypes, new string (value_builder, 0, pos), start_location);
+ return Token.LITERAL;
+ }
+ } else if (c == '\\' && !quoted) {
int surrogate;
c = escape (c, out surrogate);
if (c == -1)
return Token.ERROR;
if (surrogate != 0) {
- string_builder.Append ((char) c);
+ if (pos == value_builder.Length)
+ Array.Resize (ref value_builder, pos * 2);
+
+ value_builder[pos++] = (char) c;
c = surrogate;
}
+ } else if (c == -1) {
+ Report.Error (1039, Location, "Unterminated string literal");
+ return Token.EOF;
}
- string_builder.Append ((char) c);
- }
- Report.Error (1039, Location, "Unterminated string literal");
- return Token.EOF;
+ if (pos == value_builder.Length)
+ Array.Resize (ref value_builder, pos * 2);
+
+ value_builder[pos++] = (char) c;
+ }
}
private int consume_identifier (int s)
return res;
}
- private int consume_identifier (int c, bool quoted)
+ int consume_identifier (int c, bool quoted)
{
+ //
+ // This method is very performance sensitive. It accounts
+ // for approximately 25% of all parser time
+ //
+
int pos = 0;
+ int column = col;
+ if (quoted)
+ --column;
if (c == '\\') {
int surrogate;
}
id_builder [pos++] = (char) c;
- Location loc = Location;
- while ((c = get_char ()) != -1) {
- loop:
- if (is_identifier_part_character ((char) c)){
- if (pos == max_id_size){
- Report.Error (645, loc, "Identifier too long (limit is 512 chars)");
- return Token.ERROR;
+ try {
+ while (true) {
+ c = reader.Read ();
+
+ if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9')) {
+ id_builder [pos++] = (char) c;
+ continue;
}
-
- id_builder [pos++] = (char) c;
- } else if (c == '\\') {
- int surrogate;
- c = escape (c, out surrogate);
- if (surrogate != 0) {
- if (is_identifier_part_character ((char) c))
- id_builder [pos++] = (char) c;
- c = surrogate;
+
+ if (c < 0x80) {
+ if (c == '\\') {
+ int surrogate;
+ c = escape (c, out surrogate);
+ if (is_identifier_part_character ((char) c))
+ id_builder[pos++] = (char) c;
+
+ if (surrogate != 0) {
+ c = surrogate;
+ }
+
+ continue;
+ }
+ } else if (Char.IsLetter ((char) c) || Char.GetUnicodeCategory ((char) c) == UnicodeCategory.ConnectorPunctuation) {
+ id_builder [pos++] = (char) c;
+ continue;
}
- goto loop;
- } else {
- putback (c);
+
+ putback_char = c;
break;
}
+ } catch (IndexOutOfRangeException) {
+ Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
+ --pos;
+ col += pos;
}
+ col += pos - 1;
+
//
// Optimization: avoids doing the keyword lookup
// on uppercase letters
if (id_builder [0] >= '_' && !quoted) {
int keyword = GetKeyword (id_builder, pos);
if (keyword != -1) {
- // TODO: No need to store location for keyword, required location cleanup
- val = loc;
+ val = LocatedToken.Create (keyword == Token.AWAIT ? "await" : null, ref_line, column);
return keyword;
}
}
+ string s = InternIdentifier (id_builder, pos);
+ val = LocatedToken.Create (s, ref_line, column);
+ if (quoted && parsing_attribute_section)
+ AddEscapedIdentifier (((LocatedToken) val).Location);
+
+ return Token.IDENTIFIER;
+ }
+
+ static string InternIdentifier (char[] charBuffer, int length)
+ {
//
// Keep identifiers in an array of hashtables to avoid needless
// allocations
//
- CharArrayHashtable identifiers_group = identifiers [pos];
+ var identifiers_group = identifiers[length];
+ string s;
if (identifiers_group != null) {
- val = identifiers_group [id_builder];
- if (val != null) {
- val = new LocatedToken (loc, (string) val);
- if (quoted)
- AddEscapedIdentifier ((LocatedToken) val);
- return Token.IDENTIFIER;
+ if (identifiers_group.TryGetValue (charBuffer, out s)) {
+ return s;
}
} else {
- identifiers_group = new CharArrayHashtable (pos);
- identifiers [pos] = identifiers_group;
+ // TODO: this should be number of files dependant
+ // corlib compilation peaks at 1000 and System.Core at 150
+ int capacity = length > 20 ? 10 : 100;
+ identifiers_group = new Dictionary<char[], string> (capacity, new IdentifiersComparer (length));
+ identifiers[length] = identifiers_group;
}
- char [] chars = new char [pos];
- Array.Copy (id_builder, chars, pos);
-
- val = new String (id_builder, 0, pos);
- identifiers_group.Add (chars, val);
+ char[] chars = new char[length];
+ Array.Copy (charBuffer, chars, length);
- if (RootContext.Version == LanguageVersion.ISO_1) {
- for (int i = 1; i < chars.Length; i += 3) {
- if (chars [i] == '_' && (chars [i - 1] == '_' || chars [i + 1] == '_')) {
- Report.Error (1638, loc,
- "`{0}': Any identifier with double underscores cannot be used when ISO language version mode is specified", val.ToString ());
- }
- }
- }
-
- val = new LocatedToken (loc, (string) val);
- if (quoted)
- AddEscapedIdentifier ((LocatedToken) val);
- return Token.IDENTIFIER;
+ s = new string (charBuffer, 0, length);
+ identifiers_group.Add (chars, s);
+ return s;
}
public int xtoken ()
while ((c = get_char ()) != -1) {
switch (c) {
case '\t':
- col = ((col + 8) / 8) * 8;
+ col = ((col - 1 + tab_size) / tab_size) * tab_size;
continue;
case ' ':
}
break;
*/
- case '\r':
- if (peek_char () != '\n')
- advance_line ();
- else
- get_char ();
-
- any_token_seen |= tokens_seen;
- tokens_seen = false;
- comments_seen = false;
- continue;
-
case '\\':
tokens_seen = true;
return consume_identifier (c);
case '{':
- val = Location;
+ val = LocatedToken.Create (ref_line, col);
return Token.OPEN_BRACE;
case '}':
- val = Location;
+ val = LocatedToken.Create (ref_line, col);
return Token.CLOSE_BRACE;
case '[':
// To block doccomment inside attribute declaration.
if (doc_state == XmlCommentState.Allowed)
doc_state = XmlCommentState.NotAllowed;
- return Token.OPEN_BRACKET;
+
+ val = LocatedToken.Create (ref_line, col);
+
+ if (parsing_block == 0 || lambda_arguments_parsing)
+ return Token.OPEN_BRACKET;
+
+ int next = peek_char ();
+ switch (next) {
+ case ']':
+ case ',':
+ return Token.OPEN_BRACKET;
+
+ case ' ':
+ case '\f':
+ case '\v':
+ case '\r':
+ case '\n':
+ case '/':
+ next = peek_token ();
+ if (next == Token.COMMA || next == Token.CLOSE_BRACKET)
+ return Token.OPEN_BRACKET;
+
+ return Token.OPEN_BRACKET_EXPR;
+ default:
+ return Token.OPEN_BRACKET_EXPR;
+ }
case ']':
+ LocatedToken.CreateOptional (ref_line, col, ref val);
return Token.CLOSE_BRACKET;
case '(':
- val = Location;
+ val = LocatedToken.Create (ref_line, col);
//
// An expression versions of parens can appear in block context only
//
return Token.OPEN_PARENS;
case ')':
+ LocatedToken.CreateOptional (ref_line, col, ref val);
return Token.CLOSE_PARENS;
case ',':
+ LocatedToken.CreateOptional (ref_line, col, ref val);
return Token.COMMA;
case ';':
+ LocatedToken.CreateOptional (ref_line, col, ref val);
return Token.SEMICOLON;
case '~':
+ val = LocatedToken.Create (ref_line, col);
return Token.TILDE;
case '?':
+ val = LocatedToken.Create (ref_line, col);
return TokenizePossibleNullableType ();
case '<':
+ val = LocatedToken.Create (ref_line, col);
if (parsing_generic_less_than++ > 0)
return Token.OP_GENERICS_LT;
return TokenizeLessThan ();
case '>':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '='){
}
return Token.OP_GT;
-
+
case '+':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '+') {
d = Token.OP_INC;
return d;
case '-':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '-') {
d = Token.OP_DEC;
return d;
case '!':
+ val = LocatedToken.Create (ref_line, col);
if (peek_char () == '='){
get_char ();
return Token.OP_NE;
return Token.BANG;
case '=':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '='){
get_char ();
return Token.ASSIGN;
case '&':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '&'){
get_char ();
return Token.BITWISE_AND;
case '|':
+ val = LocatedToken.Create (ref_line, col);
d = peek_char ();
if (d == '|'){
get_char ();
return Token.BITWISE_OR;
case '*':
+ val = LocatedToken.Create (ref_line, col);
if (peek_char () == '='){
get_char ();
return Token.OP_MULT_ASSIGN;
}
- val = Location;
return Token.STAR;
case '/':
d = peek_char ();
if (d == '='){
+ val = LocatedToken.Create (ref_line, col);
get_char ();
return Token.OP_DIV_ASSIGN;
}
// Handle double-slash comments.
if (d == '/'){
get_char ();
- if (RootContext.Documentation != null && peek_char () == '/') {
- get_char ();
- // Don't allow ////.
- if ((d = peek_char ()) != '/') {
- update_comment_location ();
- if (doc_state == XmlCommentState.Allowed)
- handle_one_line_xml_comment ();
- else if (doc_state == XmlCommentState.NotAllowed)
- warn_incorrect_doc_comment ();
+ if (doc_processing) {
+ if (peek_char () == '/') {
+ get_char ();
+ // Don't allow ////.
+ if ((d = peek_char ()) != '/') {
+ if (doc_state == XmlCommentState.Allowed)
+ handle_one_line_xml_comment ();
+ else if (doc_state == XmlCommentState.NotAllowed)
+ WarningMisplacedComment (Location - 3);
+ }
+ } else {
+ if (xml_comment_buffer.Length > 0)
+ doc_state = XmlCommentState.NotAllowed;
}
}
- while ((d = get_char ()) != -1 && (d != '\n') && d != '\r');
+
+ while ((d = get_char ()) != -1 && d != '\n');
any_token_seen |= tokens_seen;
tokens_seen = false;
} else if (d == '*'){
get_char ();
bool docAppend = false;
- if (RootContext.Documentation != null && peek_char () == '*') {
+ if (doc_processing && peek_char () == '*') {
get_char ();
- update_comment_location ();
// But when it is /**/, just do nothing.
if (peek_char () == '/') {
get_char ();
}
if (doc_state == XmlCommentState.Allowed)
docAppend = true;
- else if (doc_state == XmlCommentState.NotAllowed)
- warn_incorrect_doc_comment ();
+ else if (doc_state == XmlCommentState.NotAllowed) {
+ WarningMisplacedComment (Location - 2);
+ }
}
int current_comment_start = 0;
update_formatted_doc_comment (current_comment_start);
continue;
}
+ val = LocatedToken.Create (ref_line, col);
return Token.DIV;
case '%':
+ val = LocatedToken.Create (ref_line, col);
if (peek_char () == '='){
get_char ();
return Token.OP_MOD_ASSIGN;
return Token.PERCENT;
case '^':
+ val = LocatedToken.Create (ref_line, col);
if (peek_char () == '='){
get_char ();
return Token.OP_XOR_ASSIGN;
return Token.CARRET;
case ':':
+ val = LocatedToken.Create (ref_line, col);
if (peek_char () == ':') {
get_char ();
return Token.DOUBLE_COLON;
d = peek_char ();
if (d >= '0' && d <= '9')
return is_number (c);
+
+ LocatedToken.CreateOptional (ref_line, col, ref val);
return Token.DOT;
case '#':
return Token.ERROR;
}
- if (handle_preprocessing_directive (true))
+ if (ParsePreprocessingDirective (true))
continue;
bool directive_expected = false;
continue;
}
- if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v' )
+ if (c == ' ' || c == '\t' || c == '\n' || c == '\f' || c == '\v' )
continue;
if (c == '#') {
- if (handle_preprocessing_directive (false))
+ if (ParsePreprocessingDirective (false))
break;
}
directive_expected = false;
return Token.EVAL_COMPILATION_UNIT_PARSER;
case EvalUsingDeclarationsParserCharacter:
return Token.EVAL_USING_DECLARATIONS_UNIT_PARSER;
+ case DocumentationXref:
+ return Token.DOC_SEE;
}
if (is_identifier_start_character (c)) {
return consume_identifier (c);
}
- error_details = ((char)c).ToString ();
- return Token.ERROR;
+ if (char.IsWhiteSpace ((char) c))
+ continue;
+
+ Report.Error (1056, Location, "Unexpected character `{0}'", ((char) c).ToString ());
+ }
+
+ if (CompleteOnEOF){
+ if (generated)
+ return Token.COMPLETE_COMPLETION;
+
+ generated = true;
+ return Token.GENERATE_COMPLETION;
}
+
return Token.EOF;
}
int TokenizeBackslash ()
{
+#if FULL_AST
+ int read_start = reader.Position;
+#endif
+ Location start_location = Location;
int c = get_char ();
tokens_seen = true;
if (c == '\'') {
- error_details = "Empty character literal";
- Report.Error (1011, Location, error_details);
- return Token.ERROR;
+ val = new CharLiteral (context.BuiltinTypes, (char) c, start_location);
+ Report.Error (1011, start_location, "Empty character literal");
+ return Token.LITERAL;
}
- if (c == '\r' || c == '\n') {
- Report.Error (1010, Location, "Newline in constant");
+
+ if (c == '\n') {
+ Report.Error (1010, start_location, "Newline in constant");
return Token.ERROR;
}
if (d != 0)
throw new NotImplementedException ();
- val = (char) c;
+ ILiteralConstant res = new CharLiteral (context.BuiltinTypes, (char) c, start_location);
+ val = res;
c = get_char ();
if (c != '\'') {
- Report.Error (1012, Location, "Too many characters in character literal");
+ Report.Error (1012, start_location, "Too many characters in character literal");
// Try to recover, read until newline or next "'"
while ((c = get_char ()) != -1) {
if (c == '\n' || c == '\'')
break;
}
- return Token.ERROR;
}
- return Token.LITERAL_CHARACTER;
+#if FULL_AST
+ res.ParsedValue = reader.ReadChars (read_start - 1, reader.Position);
+#endif
+
+ return Token.LITERAL;
}
int TokenizeLessThan ()
// Save current position and parse next token.
PushPosition ();
if (parse_less_than ()) {
- if (parsing_generic_declaration && token () != Token.DOT) {
+ if (parsing_generic_declaration && (parsing_generic_declaration_doc || token () != Token.DOT)) {
d = Token.OP_GENERICS_LT_DECL;
} else {
d = Token.OP_GENERICS_LT;
xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
}
- //
- // Updates current comment location.
- //
- private void update_comment_location ()
- {
- if (current_comment_location.IsNull) {
- // "-2" is for heading "//" or "/*"
- current_comment_location =
- new Location (ref_line, hidden ? -1 : col - 2);
- }
- }
-
//
// Checks if there was incorrect doc comments and raise
// warnings.
public void check_incorrect_doc_comment ()
{
if (xml_comment_buffer.Length > 0)
- warn_incorrect_doc_comment ();
- }
-
- //
- // Raises a warning when tokenizer found incorrect doccomment
- // markup.
- //
- private void warn_incorrect_doc_comment ()
- {
- if (doc_state != XmlCommentState.Error) {
- doc_state = XmlCommentState.Error;
- // in csc, it is 'XML comment is not placed on
- // a valid language element'. But that does not
- // make sense.
- Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
- }
+ WarningMisplacedComment (Location);
}
//
return null;
}
+ Report Report {
+ get { return context.Report; }
+ }
+
void reset_doc_comment ()
{
xml_comment_buffer.Length = 0;
- current_comment_location = Location.Null;
}
public void cleanup ()
{
if (ifstack != null && ifstack.Count >= 1) {
- int state = (int) ifstack.Pop ();
+ int state = ifstack.Pop ();
if ((state & REGION) != 0)
Report.Error (1038, Location, "#endregion directive expected");
else
projects / mono.git / blobdiff