introduce [*] for unknown length pointers

See #770 Currently it does not have any different behavior than `*` but it is now recommended to use `[*]` for unknown length pointers to be future-proof. Instead of [ * ] being separate tokens as the proposal suggested, this commit implements `[*]` as a single token.
ziglang · Jun 2, 2018 · f06bce5 · f06bce5
1 parent 7b386ea
commit f06bce5
Show file tree

Hide file tree

Showing 9 changed files with 87 additions and 12 deletions.
diff --git a/doc/langref.html.in b/doc/langref.html.in
@@ -6450,7 +6450,7 @@ ContainerInitBody = list(StructLiteralField, ",") | list(Expression, ",")
 
 StructLiteralField = "." Symbol "=" Expression
 
-PrefixOp = "!" | "-" | "~" | ("*" option("align" "(" Expression option(":" Integer ":" Integer) ")" ) option("const") option("volatile")) | "?" | "??" | "-%" | "try" | "await"
+PrefixOp = "!" | "-" | "~" | (("*" | "[*]") option("align" "(" Expression option(":" Integer ":" Integer) ")" ) option("const") option("volatile")) | "?" | "??" | "-%" | "try" | "await"
 
 PrimaryExpression = Integer | Float | String | CharLiteral | KeywordLiteral | GroupedExpression | BlockExpression(BlockOrExpression) | Symbol | ("@" Symbol FnCallExpression) | ArrayType | FnProto | AsmExpression | ContainerDecl | ("continue" option(":" Symbol)) | ErrorSetDecl | PromiseType
 

diff --git a/src/all_types.hpp b/src/all_types.hpp
@@ -625,6 +625,7 @@ struct AstNodePrefixOpExpr {
 };
 
 struct AstNodePointerType {
+    Token *star_token;
     AstNode *align_expr;
     BigInt *bit_offset_start;
     BigInt *bit_offset_end;

diff --git a/src/parser.cpp b/src/parser.cpp
@@ -1174,6 +1174,7 @@ static PrefixOp tok_to_prefix_op(Token *token) {
 
 static AstNode *ast_parse_pointer_type(ParseContext *pc, size_t *token_index, Token *star_tok) {
     AstNode *node = ast_create_node(pc, NodeTypePointerType, star_tok);
+    node->data.pointer_type.star_token = star_tok;
 
     Token *token = &pc->tokens->at(*token_index);
     if (token->id == TokenIdKeywordAlign) {
@@ -1211,11 +1212,11 @@ static AstNode *ast_parse_pointer_type(ParseContext *pc, size_t *token_index, To
 
 /*
 PrefixOpExpression = PrefixOp ErrorSetExpr | SuffixOpExpression
-PrefixOp = "!" | "-" | "~" | ("*" option("align" "(" Expression option(":" Integer ":" Integer) ")" ) option("const") option("volatile")) | "?" | "??" | "-%" | "try" | "await"
+PrefixOp = "!" | "-" | "~" | (("*" | "[*]") option("align" "(" Expression option(":" Integer ":" Integer) ")" ) option("const") option("volatile")) | "?" | "??" | "-%" | "try" | "await"
 */
 static AstNode *ast_parse_prefix_op_expr(ParseContext *pc, size_t *token_index, bool mandatory) {
     Token *token = &pc->tokens->at(*token_index);
-    if (token->id == TokenIdStar) {
+    if (token->id == TokenIdStar || token->id == TokenIdBracketStarBracket) {
         *token_index += 1;
         return ast_parse_pointer_type(pc, token_index, token);
     }

diff --git a/src/tokenizer.cpp b/src/tokenizer.cpp
@@ -219,6 +219,8 @@ enum TokenizeState {
     TokenizeStateSawAtSign,
     TokenizeStateCharCode,
     TokenizeStateError,
+    TokenizeStateLBracket,
+    TokenizeStateLBracketStar,
 };
 
 
@@ -539,8 +541,8 @@ void tokenize(Buf *buf, Tokenization *out) {
                         end_token(&t);
                         break;
                     case '[':
+                        t.state = TokenizeStateLBracket;
                         begin_token(&t, TokenIdLBracket);
-                        end_token(&t);
                         break;
                     case ']':
                         begin_token(&t, TokenIdRBracket);
@@ -852,6 +854,30 @@ void tokenize(Buf *buf, Tokenization *out) {
                         continue;
                 }
                 break;
+            case TokenizeStateLBracket:
+                switch (c) {
+                    case '*':
+                        t.state = TokenizeStateLBracketStar;
+                        set_token_id(&t, t.cur_tok, TokenIdBracketStarBracket);
+                        break;
+                    default:
+                        // reinterpret as just an lbracket
+                        t.pos -= 1;
+                        end_token(&t);
+                        t.state = TokenizeStateStart;
+                        continue;
+                }
+                break;
+            case TokenizeStateLBracketStar:
+                switch (c) {
+                    case ']':
+                        end_token(&t);
+                        t.state = TokenizeStateStart;
+                        break;
+                    default:
+                        invalid_char_error(&t, c);
+                }
+                break;
             case TokenizeStateSawPlusPercent:
                 switch (c) {
                     case '=':
@@ -1467,12 +1493,14 @@ void tokenize(Buf *buf, Tokenization *out) {
         case TokenizeStateLineString:
         case TokenizeStateLineStringEnd:
         case TokenizeStateSawBarBar:
+        case TokenizeStateLBracket:
             end_token(&t);
             break;
         case TokenizeStateSawDotDot:
         case TokenizeStateSawBackslash:
         case TokenizeStateLineStringContinue:
         case TokenizeStateLineStringContinueC:
+        case TokenizeStateLBracketStar:
             tokenize_error(&t, "unexpected EOF");
             break;
         case TokenizeStateLineComment:
@@ -1509,6 +1537,7 @@ const char * token_name(TokenId id) {
         case TokenIdBitShiftRight: return ">>";
         case TokenIdBitShiftRightEq: return ">>=";
         case TokenIdBitXorEq: return "^=";
+        case TokenIdBracketStarBracket: return "[*]";
         case TokenIdCharLiteral: return "CharLiteral";
         case TokenIdCmpEq: return "==";
         case TokenIdCmpGreaterOrEq: return ">=";

diff --git a/src/tokenizer.hpp b/src/tokenizer.hpp
@@ -28,6 +28,7 @@ enum TokenId {
     TokenIdBitShiftRight,
     TokenIdBitShiftRightEq,
     TokenIdBitXorEq,
+    TokenIdBracketStarBracket,
     TokenIdCharLiteral,
     TokenIdCmpEq,
     TokenIdCmpGreaterOrEq,

diff --git a/std/cstr.zig b/std/cstr.zig
@@ -9,13 +9,13 @@ pub const line_sep = switch (builtin.os) {
     else => "\n",
 };
 
-pub fn len(ptr: *const u8) usize {
+pub fn len(ptr: [*]const u8) usize {
     var count: usize = 0;
     while (ptr[count] != 0) : (count += 1) {}
     return count;
 }
 
-pub fn cmp(a: *const u8, b: *const u8) i8 {
+pub fn cmp(a: [*]const u8, b: [*]const u8) i8 {
     var index: usize = 0;
     while (a[index] == b[index] and a[index] != 0) : (index += 1) {}
     if (a[index] > b[index]) {
@@ -27,11 +27,11 @@ pub fn cmp(a: *const u8, b: *const u8) i8 {
     }
 }
 
-pub fn toSliceConst(str: *const u8) []const u8 {
+pub fn toSliceConst(str: [*]const u8) []const u8 {
     return str[0..len(str)];
 }
 
-pub fn toSlice(str: *u8) []u8 {
+pub fn toSlice(str: [*]u8) []u8 {
     return str[0..len(str)];
 }
 

diff --git a/std/zig/parse.zig b/std/zig/parse.zig
@@ -3292,7 +3292,7 @@ fn tokenIdToPrefixOp(id: @TagType(Token.Id)) ?ast.Node.PrefixOp.Op {
         Token.Id.Minus => ast.Node.PrefixOp.Op{ .Negation = void{} },
         Token.Id.MinusPercent => ast.Node.PrefixOp.Op{ .NegationWrap = void{} },
         Token.Id.Ampersand => ast.Node.PrefixOp.Op{ .AddressOf = void{} },
-        Token.Id.Asterisk, Token.Id.AsteriskAsterisk => ast.Node.PrefixOp.Op{
+        Token.Id.Asterisk, Token.Id.AsteriskAsterisk, Token.Id.BracketStarBracket => ast.Node.PrefixOp.Op{
             .PtrType = ast.Node.PrefixOp.PtrInfo{
                 .align_info = null,
                 .const_token = null,

diff --git a/std/zig/parser_test.zig b/std/zig/parser_test.zig
@@ -1,3 +1,10 @@
+test "zig fmt: pointer of unknown length" {
+    try testCanonical(
+        \\fn foo(ptr: [*]u8) void {}
+        \\
+    );
+}
+
 test "zig fmt: spaces around slice operator" {
     try testCanonical(
         \\var a = b[c..d];

diff --git a/std/zig/tokenizer.zig b/std/zig/tokenizer.zig
@@ -143,6 +143,7 @@ pub const Token = struct {
         FloatLiteral,
         LineComment,
         DocComment,
+        BracketStarBracket,
         Keyword_align,
         Keyword_and,
         Keyword_asm,
@@ -263,6 +264,8 @@ pub const Tokenizer = struct {
         Period,
         Period2,
         SawAtSign,
+        LBracket,
+        LBracketStar,
     };
 
     pub fn next(self: *Tokenizer) Token {
@@ -325,9 +328,7 @@ pub const Tokenizer = struct {
                         break;
                     },
                     '[' => {
-                        result.id = Token.Id.LBracket;
-                        self.index += 1;
-                        break;
+                        state = State.LBracket;
                     },
                     ']' => {
                         result.id = Token.Id.RBracket;
@@ -429,6 +430,28 @@ pub const Tokenizer = struct {
                     },
                 },
 
+                State.LBracket => switch (c) {
+                    '*' => {
+                        state = State.LBracketStar;
+                    },
+                    else => {
+                        result.id = Token.Id.LBracket;
+                        break;
+                    },
+                },
+
+                State.LBracketStar => switch (c) {
+                    ']' => {
+                        result.id = Token.Id.BracketStarBracket;
+                        self.index += 1;
+                        break;
+                    },
+                    else => {
+                        result.id = Token.Id.Invalid;
+                        break;
+                    },
+                },
+
                 State.Ampersand => switch (c) {
                     '=' => {
                         result.id = Token.Id.AmpersandEqual;
@@ -1008,6 +1031,7 @@ pub const Tokenizer = struct {
                 State.CharLiteralEscape2,
                 State.CharLiteralEnd,
                 State.StringLiteralBackslash,
+                State.LBracketStar,
                 => {
                     result.id = Token.Id.Invalid;
                 },
@@ -1024,6 +1048,9 @@ pub const Tokenizer = struct {
                 State.Slash => {
                     result.id = Token.Id.Slash;
                 },
+                State.LBracket => {
+                    result.id = Token.Id.LBracket;
+                },
                 State.Zero => {
                     result.id = Token.Id.IntegerLiteral;
                 },
@@ -1142,6 +1169,15 @@ test "tokenizer" {
     testTokenize("test", []Token.Id{Token.Id.Keyword_test});
 }
 
+test "tokenizer - unknown length pointer" {
+    testTokenize(
+        \\[*]u8
+    , []Token.Id{
+        Token.Id.BracketStarBracket,
+        Token.Id.Identifier,
+    });
+}
+
 test "tokenizer - char literal with hex escape" {
     testTokenize(
         \\'\x1b'