Generic Structs (#11)

Reviewed-on: #11
Co-authored-by: cosmonaut <evan@moonside.games>
Co-committed-by: cosmonaut <evan@moonside.games>

.clang-format

@@ -0,0 +1,166 @@
+---
+Language:        Cpp
+# BasedOnStyle:  LLVM
+AccessModifierOffset: -2
+AlignAfterOpenBracket: AlwaysBreak
+AlignConsecutiveMacros: None
+AlignConsecutiveAssignments: None
+AlignConsecutiveBitFields: None
+AlignConsecutiveDeclarations: None
+AlignEscapedNewlines: Right
+AlignOperands:   Align
+AlignTrailingComments: true
+AllowAllArgumentsOnNextLine: false
+AllowAllConstructorInitializersOnNextLine: true
+AllowAllParametersOfDeclarationOnNextLine: false
+AllowShortEnumsOnASingleLine: true
+AllowShortBlocksOnASingleLine: Never
+AllowShortCaseLabelsOnASingleLine: false
+AllowShortFunctionsOnASingleLine: All
+AllowShortLambdasOnASingleLine: All
+AllowShortIfStatementsOnASingleLine: Never
+AllowShortLoopsOnASingleLine: false
+AlwaysBreakAfterDefinitionReturnType: None
+AlwaysBreakAfterReturnType: None
+AlwaysBreakBeforeMultilineStrings: false
+AlwaysBreakTemplateDeclarations: MultiLine
+AttributeMacros:
+  - __capability
+BinPackArguments: false
+BinPackParameters: false
+BraceWrapping:
+  AfterCaseLabel:  false
+  AfterClass:      false
+  AfterControlStatement: Never
+  AfterEnum:       false
+  AfterFunction:   false
+  AfterNamespace:  false
+  AfterObjCDeclaration: false
+  AfterStruct:     false
+  AfterUnion:      false
+  AfterExternBlock: false
+  BeforeCatch:     false
+  BeforeElse:      false
+  BeforeLambdaBody: false
+  BeforeWhile:     false
+  IndentBraces:    false
+  SplitEmptyFunction: true
+  SplitEmptyRecord: true
+  SplitEmptyNamespace: true
+BreakBeforeBinaryOperators: None
+BreakBeforeConceptDeclarations: true
+BreakBeforeBraces: Allman
+BreakBeforeInheritanceComma: false
+BreakInheritanceList: BeforeColon
+BreakBeforeTernaryOperators: true
+BreakConstructorInitializersBeforeComma: false
+BreakConstructorInitializers: BeforeColon
+BreakAfterJavaFieldAnnotations: false
+BreakStringLiterals: true
+ColumnLimit:     80
+CommentPragmas:  '^ IWYU pragma:'
+CompactNamespaces: false
+ConstructorInitializerAllOnOneLineOrOnePerLine: false
+ConstructorInitializerIndentWidth: 4
+ContinuationIndentWidth: 4
+Cpp11BracedListStyle: true
+DeriveLineEnding: true
+DerivePointerAlignment: false
+DisableFormat:   false
+EmptyLineBeforeAccessModifier: LogicalBlock
+ExperimentalAutoDetectBinPacking: false
+FixNamespaceComments: true
+ForEachMacros:
+  - foreach
+  - Q_FOREACH
+  - BOOST_FOREACH
+StatementAttributeLikeMacros:
+  - Q_EMIT
+IncludeBlocks:   Preserve
+IncludeCategories:
+  - Regex:           '^"(llvm|llvm-c|clang|clang-c)/'
+    Priority:        2
+    SortPriority:    0
+    CaseSensitive:   false
+  - Regex:           '^(<|"(gtest|gmock|isl|json)/)'
+    Priority:        3
+    SortPriority:    0
+    CaseSensitive:   false
+  - Regex:           '.*'
+    Priority:        1
+    SortPriority:    0
+    CaseSensitive:   false
+IncludeIsMainRegex: '(Test)?$'
+IncludeIsMainSourceRegex: ''
+IndentCaseLabels: false
+IndentCaseBlocks: false
+IndentGotoLabels: true
+IndentPPDirectives: None
+IndentExternBlock: AfterExternBlock
+IndentRequires:  false
+IndentWidth:     4
+IndentWrappedFunctionNames: false
+InsertTrailingCommas: None
+JavaScriptQuotes: Leave
+JavaScriptWrapImports: true
+KeepEmptyLinesAtTheStartOfBlocks: true
+MacroBlockBegin: ''
+MacroBlockEnd:   ''
+MaxEmptyLinesToKeep: 1
+NamespaceIndentation: None
+ObjCBinPackProtocolList: Auto
+ObjCBlockIndentWidth: 4
+ObjCBreakBeforeNestedBlockParam: true
+ObjCSpaceAfterProperty: false
+ObjCSpaceBeforeProtocolList: true
+PenaltyBreakAssignment: 2
+PenaltyBreakBeforeFirstCallParameter: 19
+PenaltyBreakComment: 300
+PenaltyBreakFirstLessLess: 120
+PenaltyBreakString: 1000
+PenaltyBreakTemplateDeclaration: 10
+PenaltyExcessCharacter: 1000000
+PenaltyReturnTypeOnItsOwnLine: 1000000
+PenaltyIndentedWhitespace: 0
+PointerAlignment: Right
+ReflowComments:  true
+SortIncludes:    true
+SortJavaStaticImport: Before
+SortUsingDeclarations: true
+SpaceAfterCStyleCast: false
+SpaceAfterLogicalNot: false
+SpaceAfterTemplateKeyword: true
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeCaseColon: false
+SpaceBeforeCpp11BracedList: false
+SpaceBeforeCtorInitializerColon: true
+SpaceBeforeInheritanceColon: true
+SpaceBeforeParens: ControlStatements
+SpaceAroundPointerQualifiers: Default
+SpaceBeforeRangeBasedForLoopColon: true
+SpaceInEmptyBlock: false
+SpaceInEmptyParentheses: false
+SpacesBeforeTrailingComments: 1
+SpacesInAngles:  false
+SpacesInConditionalStatement: false
+SpacesInContainerLiterals: true
+SpacesInCStyleCastParentheses: false
+SpacesInParentheses: false
+SpacesInSquareBrackets: false
+SpaceBeforeSquareBrackets: false
+BitFieldColonSpacing: Both
+Standard:        Latest
+StatementMacros:
+  - Q_UNUSED
+  - QT_REQUIRE_VERSION
+TabWidth:        8
+UseCRLF:         false
+UseTab:          Never
+WhitespaceSensitiveMacros:
+  - STRINGIZE
+  - PP_STRINGIZE
+  - BOOST_PP_STRINGIZE
+  - NS_SWIFT_NAME
+  - CF_SWIFT_NAME
+...
+

CMakeLists.txt

@@ -41,12 +41,14 @@ add_executable(
 	# Source
 	src/ast.h
 	src/codegen.h
-	src/identcheck.h
 	src/parser.h
+	src/validation.h
+	src/util.h
     src/ast.c
 	src/codegen.c
-	src/identcheck.c
 	src/parser.c
+	src/validation.c
+	src/util.c
 	src/main.c
 	# Generated code
     ${BISON_Parser_OUTPUTS}

access.w

@@ -0,0 +1,13 @@
+struct G {
+    Foo(t: bool): bool {
+        return t;
+    }
+}
+
+struct Program {
+    static main(): int {
+        g: G = alloc G;
+        g.Foo(true);
+        return 0;
+    }
+}

example.w

@@ -54,6 +54,8 @@ struct Program
             myStruct.myInt = 4;
         }
 
-        return myStruct.myInt;
+        Console.PrintLine("%i", myStruct.myInt);
+
+        return 0;
     }
 }

generators/wraith.lex

@@ -12,6 +12,7 @@
 "double"                    return DOUBLE;
 "string"                    return STRING;
 "bool"                      return BOOL;
+"MemoryAddress"             return MEMORYADDRESS;
 "struct"                    return STRUCT;
 "return"                    return RETURN;
 "static"                    return STATIC;
@@ -40,6 +41,7 @@
 ";"                         return SEMICOLON;
 ":"                         return COLON;
 "?"                         return QUESTION;
+"@"                         return AT;
 "("                         return LEFT_PAREN;
 ")"                         return RIGHT_PAREN;
 "["                         return LEFT_BRACKET;

generators/wraith.y

@@ -25,6 +25,7 @@ extern FILE *yyin;
 %token DOUBLE
 %token STRING
 %token BOOL
+%token MEMORYADDRESS
 %token STRUCT
 %token RETURN
 %token STATIC
@@ -54,6 +55,7 @@ extern FILE *yyin;
 %token SEMICOLON
 %token COLON
 %token QUESTION
+%token AT
 %token LEFT_PAREN
 %token RIGHT_PAREN
 %token LEFT_BRACE
@@ -107,9 +109,17 @@ BaseType                : VOID
                         {
                             $$ = MakePrimitiveTypeNode(Bool);
                         }
+                        | MEMORYADDRESS
+                        {
+                            $$ = MakePrimitiveTypeNode(MemoryAddress);
+                        }
+                        | Identifier GenericArgumentClauseNonEmpty
+                        {
+                            $$ = MakeConcreteGenericTypeNode($1, $2);
+                        }
                         | Identifier
                         {
-                            $$ = MakeCustomTypeNode(yytext);
+                            $$ = MakeCustomTypeNode($1);
                         }
                         | REFERENCE LESS_THAN Type GREATER_THAN
                         {
@@ -141,11 +151,40 @@ AccessExpression        : Identifier POINT AccessExpression
                             $$ = $1;
                         }
 
+SystemCallExpression    : AT Identifier
+                        {
+                            $$ = $2;
+                        }
+
 Number                  : NUMBER
                         {
                             $$ = MakeNumberNode(yytext);
                         }
 
+FieldInit               : Identifier COLON Expression
+                        {
+                            $$ = MakeFieldInitNode($1, $3);
+                        }
+
+StructInitFields        : FieldInit
+                        {
+                            $$ = StartStructInitFieldsNode($1);
+                        }
+                        | StructInitFields COMMA FieldInit
+                        {
+                            $$ = AddFieldInitNode($1, $3);
+                        }
+                        |
+                        {
+                            $$ = MakeEmptyFieldInitNode();
+                        }
+                        ;
+
+StructInitExpression    : Type LEFT_BRACE StructInitFields RIGHT_BRACE
+                        {
+                            $$ = MakeStructInitExpressionNode($1, $3);
+                        }
+
 PrimaryExpression       : Number
                         | STRING_LITERAL
                         {
@@ -157,6 +196,7 @@ PrimaryExpression       : Number
                         }
                         | FunctionCallExpression
                         | AccessExpression
+                        | StructInitExpression
                         ;
 
 UnaryExpression         : BANG Expression
@@ -230,9 +270,13 @@ ReturnStatement         : RETURN Expression
                             $$ = MakeReturnVoidStatementNode();
                         }
 
-FunctionCallExpression  : AccessExpression LEFT_PAREN Arguments RIGHT_PAREN
+FunctionCallExpression  : AccessExpression GenericArgumentClause LEFT_PAREN Arguments RIGHT_PAREN
                         {
-                            $$ = MakeFunctionCallExpressionNode($1, $3);
+                            $$ = MakeFunctionCallExpressionNode($1, $4, $2);
+                        }
+                        | SystemCallExpression GenericArgumentClause LEFT_PAREN Arguments RIGHT_PAREN
+                        {
+                            $$ = MakeSystemCallExpressionNode($1, $4, $2);
                         }
 
 PartialStatement        : FunctionCallExpression
@@ -275,11 +319,11 @@ Statements              : Statement
                             $$ = AddStatement($1, $2);
                         }
 
-Arguments               : PrimaryExpression
+Arguments               : Expression
                         {
                             $$ = StartFunctionArgumentSequenceNode($1);
                         }
-                        | Arguments COMMA PrimaryExpression
+                        | Arguments COMMA Expression
                         {
                             $$ = AddFunctionArgumentNode($1, $3);
                         }
@@ -307,14 +351,63 @@ Body                    : LEFT_BRACE Statements RIGHT_BRACE
                             $$ = $2;
                         }
 
-FunctionSignature       : Identifier LEFT_PAREN SignatureArguments RIGHT_PAREN COLON Type
+GenericDeclaration      : Identifier
                         {
-                            $$ = MakeFunctionSignatureNode($1, $6, $3, MakeFunctionModifiersNode(NULL, 0));
+                            $$ = MakeGenericDeclarationNode($1, NULL);
                         }
-                        | STATIC Identifier LEFT_PAREN SignatureArguments RIGHT_PAREN COLON Type
+
+GenericDeclarations     : GenericDeclaration
+                        {
+                            $$ = StartGenericDeclarationsNode($1);
+                        }
+                        | GenericDeclarations COMMA GenericDeclaration
+                        {
+                            $$ = AddGenericDeclaration($1, $3);
+                        }
+
+GenericDeclarationClause    : LESS_THAN GenericDeclarations GREATER_THAN
+                            {
+                                $$ = $2;
+                            }
+                            |
+                            {
+                                $$ = MakeEmptyGenericDeclarationsNode();
+                            }
+
+GenericArgument         : Type
+                        {
+                            $$ = MakeGenericArgumentNode($1);
+                        }
+
+GenericArguments        : GenericArgument
+                        {
+                            $$ = StartGenericArgumentsNode($1);
+                        }
+                        | GenericArguments COMMA GenericArgument
+                        {
+                            $$ = AddGenericArgument($1, $3);
+                        }
+
+GenericArgumentClauseNonEmpty   : LESS_THAN GenericArguments GREATER_THAN
+                                {
+                                    $$ = $2;
+                                }
+                                ;
+
+GenericArgumentClause   : GenericArgumentClauseNonEmpty
+                        |
+                        {
+                            $$ = MakeEmptyGenericArgumentsNode();
+                        }
+
+FunctionSignature       : Identifier GenericDeclarationClause LEFT_PAREN SignatureArguments RIGHT_PAREN COLON Type
+                        {
+                            $$ = MakeFunctionSignatureNode($1, $7, $4, MakeFunctionModifiersNode(NULL, 0), $2);
+                        }
+                        | STATIC Identifier GenericDeclarationClause LEFT_PAREN SignatureArguments RIGHT_PAREN COLON Type
                         {
                             Node *modifier = MakeStaticNode();
-                            $$ = MakeFunctionSignatureNode($2, $7, $4, MakeFunctionModifiersNode(&modifier, 1));
+                            $$ = MakeFunctionSignatureNode($2, $8, $5, MakeFunctionModifiersNode(&modifier, 1), $3);
                         }
 
 FunctionDeclaration     : FunctionSignature Body
@@ -322,9 +415,9 @@ FunctionDeclaration     : FunctionSignature Body
                             $$ = MakeFunctionDeclarationNode($1, $2);
                         }
 
-StructDeclaration       : STRUCT Identifier LEFT_BRACE Declarations RIGHT_BRACE
+StructDeclaration       : STRUCT Identifier GenericDeclarationClause LEFT_BRACE Declarations RIGHT_BRACE
                         {
-                            $$ = MakeStructDeclarationNode($2, $4);
+                            $$ = MakeStructDeclarationNode($2, $5, $3);
                         }
 
 Declaration             : FunctionDeclaration

generic.w

@@ -0,0 +1,59 @@
+struct Foo {
+    static Func2<U>(u: U) : U {
+        return u;
+    }
+
+    static Func<T>(t: T): T {
+        foo: T = t;
+        return Foo.Func2(foo);
+    }
+}
+
+struct MemoryBlock<T>
+{
+	start: MemoryAddress;
+	capacity: uint;
+
+	AddressOf(index: uint): MemoryAddress
+    {
+        return start + (index * @sizeof<T>());
+    }
+
+    Get(index: uint): T
+    {
+        return @dereference<T>(AddressOf(index));
+    }
+
+    Set(index: uint, value: T): void
+    {
+        @memcpy(AddressOf(index), @addr(value), @sizeof<T>());
+    }
+
+    Free(): void
+    {
+        @free(start);
+    }
+}
+
+struct Program {
+    static Main(): int {
+        x: int = 4;
+        y: int = Foo.Func(x);
+        block: MemoryBlock<int> = MemoryBlock<int>
+        {
+            capacity: y,
+            start: @malloc(y * @sizeof<int>())
+        };
+        block.Set(0, 5);
+        block.Set(1, 3);
+        block.Set(2, 9);
+        block.Set(3, 100);
+        Console.PrintLine("%p", block.start);
+        Console.PrintLine("%i", block.Get(0));
+        Console.PrintLine("%i", block.Get(1));
+        Console.PrintLine("%i", block.Get(2));
+        Console.PrintLine("%i", block.Get(3));
+        block.Free();
+        return 0;
+    }
+}

ordering.w

@@ -0,0 +1,15 @@
+struct Foo {
+    static Func(): void {
+        Func2();
+    }
+
+    static Func2(): void {
+        Func();
+    }
+}
+
+struct Program {
+    static main(): int {
+        return 0;
+    }
+}

src/ast.h

@@ -2,7 +2,15 @@
 #define WRAITH_AST_H
 
 #include <stdint.h>
-#include "identcheck.h"
+
+/* -Wpedantic nameless union/struct silencing */
+#ifndef WRAITHNAMELESS
+#ifdef __GNUC__
+#define WRAITHNAMELESS __extension__
+#else
+#define WRAITHNAMELESS
+#endif /* __GNUC__ */
+#endif /* WRAITHNAMELESS */
 
 typedef enum
 {
@@ -11,10 +19,11 @@ typedef enum
     Assignment,
     BinaryExpression,
     Comment,
+    ConcreteGenericTypeNode,
     CustomTypeNode,
     Declaration,
     DeclarationSequence,
-    Expression,
+    FieldInit,
     ForLoop,
     FunctionArgumentSequence,
     FunctionCallExpression,
@@ -22,6 +31,11 @@ typedef enum
     FunctionModifiers,
     FunctionSignature,
     FunctionSignatureArguments,
+    GenericArgument,
+    GenericArguments,
+    GenericDeclaration,
+    GenericDeclarations,
+    GenericTypeNode,
     Identifier,
     IfStatement,
     IfElseStatement,
@@ -34,6 +48,9 @@ typedef enum
     StaticModifier,
     StringLiteral,
     StructDeclaration,
+    StructInit,
+    StructInitFields,
+    SystemCall,
     Type,
     UnaryExpression
 } SyntaxKind;
@@ -63,7 +80,8 @@ typedef enum
     UInt,
     Float,
     Double,
-    String
+    String,
+    MemoryAddress
 } PrimitiveType;
 
 typedef union
@@ -72,14 +90,25 @@ typedef union
     BinaryOperator binaryOperator;
 } Operator;
 
-typedef struct TypeTag
+typedef struct TypeTag TypeTag;
+
+typedef struct ConcreteGenericTypeTag
+{
+    char *name;
+    TypeTag **genericArguments;
+    uint32_t genericArgumentCount;
+} ConcreteGenericTypeTag;
+
+struct TypeTag
 {
     enum Type
     {
         Unknown,
         Primitive,
         Reference,
-        Custom
+        Custom,
+        Generic,
+        ConcreteGeneric
     } type;
     union
     {
@@ -89,153 +118,350 @@ typedef struct TypeTag
         struct TypeTag *referenceType;
         /* Valid when type = Custom. */
         char *customType;
+        /* Valid when type = Generic. */
+        char *genericType;
+        /* Valid when type = ConcreteGeneric */
+        ConcreteGenericTypeTag concreteGenericType;
     } value;
-} TypeTag;
+};
 
-typedef struct Node
+typedef struct Node Node;
+
+struct Node
 {
+    Node *parent;
     SyntaxKind syntaxKind;
-    struct Node **children;
-    uint32_t childCount;
-    union
+    WRAITHNAMELESS union
     {
-        UnaryOperator unaryOperator;
-        BinaryOperator binaryOperator;
-    } operator;
-    union
-    {
-        char *string;
-        uint64_t number;
-    } value;
-    PrimitiveType primitiveType;
-    TypeTag *typeTag;
-    IdNode *idLink;
-} Node;
+        struct
+        {
+            Node *accessee;
+            Node *accessor;
+        } accessExpression;
 
-const char* SyntaxKindString(SyntaxKind syntaxKind);
+        struct
+        {
+            Node *type;
+        } allocExpression;
+
+        struct
+        {
+            Node *left;
+            Node *right;
+        } assignmentStatement;
+
+        struct
+        {
+            Node *left;
+            Node *right;
+            BinaryOperator operator;
+        } binaryExpression;
+
+        struct
+        {
+
+        } comment;
+
+        struct
+        {
+            char *name;
+            Node *genericArguments;
+        } concreteGenericType;
+
+        struct
+        {
+            char *name;
+        } customType;
+
+        struct
+        {
+            Node *type;
+            Node *identifier;
+        } declaration;
+
+        struct
+        {
+            Node **sequence;
+            uint32_t count;
+        } declarationSequence;
+
+        struct
+        {
+            Node *identifier;
+            Node *expression;
+        } fieldInit;
+
+        struct
+        {
+            Node *declaration;
+            Node *startNumber;
+            Node *endNumber;
+            Node *statementSequence;
+        } forLoop;
+
+        struct
+        {
+            Node **sequence;
+            uint32_t count;
+        } functionArgumentSequence;
+
+        struct
+        {
+            Node *identifier; /* FIXME: need better name */
+            Node *argumentSequence;
+            Node *genericArguments;
+        } functionCallExpression;
+
+        struct
+        {
+            Node *functionSignature;
+            Node *functionBody;
+        } functionDeclaration;
+
+        struct
+        {
+            Node **sequence;
+            uint32_t count;
+        } functionModifiers;
+
+        struct
+        {
+            Node *identifier;
+            Node *type;
+            Node *arguments;
+            Node *modifiers;
+            Node *genericDeclarations;
+        } functionSignature;
+
+        struct
+        {
+            Node **sequence;
+            uint32_t count;
+        } functionSignatureArguments;
+
+        struct
+        {
+            Node *type;
+        } genericArgument;
+
+        struct
+        {
+            Node **arguments;
+            uint32_t count;
+        } genericArguments;
+
+        struct
+        {
+            Node *identifier;
+            Node *constraint;
+        } genericDeclaration;
+
+        struct
+        {
+            Node **declarations;
+            uint32_t count;
+        } genericDeclarations;
+
+        struct
+        {
+            char *name;
+        } genericType;
+
+        struct
+        {
+            char *name;
+        } identifier;
+
+        struct
+        {
+            Node *expression;
+            Node *statementSequence;
+        } ifStatement;
+
+        struct
+        {
+            Node *ifStatement;
+            Node *elseStatement;
+        } ifElseStatement;
+
+        struct
+        {
+            uint64_t value;
+        } number;
+
+        struct
+        {
+            PrimitiveType type;
+        } primitiveType;
+
+        struct
+        {
+            Node *type;
+        } referenceType;
+
+        struct
+        {
+            Node *expression;
+        } returnStatement;
+
+        struct
+        {
+
+        } returnVoidStatement;
+
+        struct
+        {
+            Node **sequence;
+            uint32_t count;
+        } statementSequence;
+
+        struct
+        {
+
+        } staticModifier; /* FIXME: modifiers should just be an enum */
+
+        struct
+        {
+            char *string;
+        } stringLiteral;
+
+        struct
+        {
+            Node *identifier;
+            Node *declarationSequence;
+            Node *genericDeclarations;
+        } structDeclaration;
+
+        struct
+        {
+            Node *type;
+            Node *initFields;
+        } structInit;
+
+        struct
+        {
+            Node **fieldInits;
+            uint32_t count;
+        } structInitFields;
+
+        struct
+        {
+            Node *identifier;
+            Node *argumentSequence;
+            Node *genericArguments;
+        } systemCall;
+
+        struct
+        {
+            Node *typeNode;
+        } type; /* FIXME: this needs a refactor */
+
+        struct
+        {
+            Node *child;
+            UnaryOperator operator;
+        } unaryExpression;
+    };
+    TypeTag *typeTag;
+};
+
+const char *SyntaxKindString(SyntaxKind syntaxKind);
 
 uint8_t IsPrimitiveType(Node *typeNode);
-Node* MakePrimitiveTypeNode(
-    PrimitiveType type
-);
-Node* MakeCustomTypeNode(
-    char *string
-);
-Node* MakeReferenceTypeNode(
-    Node *typeNode
-);
-Node* MakeTypeNode(
-    Node *typeNode /* can be primitive, custom, or reference */
-);
-Node* MakeIdentifierNode(
-    const char *id
-);
-Node* MakeNumberNode(
-    const char *numberString
-);
-Node* MakeStringNode(
-    const char *string
-);
-Node* MakeStaticNode();
-Node* MakeFunctionModifiersNode(
-    Node **pModifierNodes,
-    uint32_t modifierCount
-);
-Node* MakeUnaryNode(
-    UnaryOperator operator,
-    Node *child
-);
-Node* MakeBinaryNode(
-    BinaryOperator operator,
-    Node *left,
-    Node *right
-);
-Node* MakeDeclarationNode(
-    Node* typeNode,
-    Node* identifierNode
-);
-Node* MakeAssignmentNode(
-    Node *left,
-    Node *right
-);
-Node* StartStatementSequenceNode(
-    Node* statementNode
-);
-Node* AddStatement(
-    Node* statementSequenceNode,
-    Node *statementNode
-);
-Node* MakeReturnStatementNode(
-    Node *expressionNode
-);
-Node* MakeReturnVoidStatementNode();
-Node* StartFunctionSignatureArgumentsNode(
-    Node *argumentNode
-);
-Node* AddFunctionSignatureArgumentNode(
-    Node *argumentsNode,
-    Node *argumentNode
-);
+Node *MakePrimitiveTypeNode(PrimitiveType type);
+Node *MakeCustomTypeNode(Node *identifierNode);
+Node *MakeReferenceTypeNode(Node *typeNode);
+Node *MakeConcreteGenericTypeNode(
+    Node *identifierNode,
+    Node *genericArgumentsNode);
+Node *MakeTypeNode(Node *typeNode);
+Node *MakeIdentifierNode(const char *id);
+Node *MakeNumberNode(const char *numberString);
+Node *MakeStringNode(const char *string);
+Node *MakeStaticNode();
+Node *MakeFunctionModifiersNode(Node **pModifierNodes, uint32_t modifierCount);
+Node *MakeUnaryNode(UnaryOperator operator, Node * child);
+Node *MakeBinaryNode(BinaryOperator operator, Node * left, Node *right);
+Node *MakeDeclarationNode(Node *typeNode, Node *identifierNode);
+Node *MakeAssignmentNode(Node *left, Node *right);
+Node *StartStatementSequenceNode(Node *statementNode);
+Node *AddStatement(Node *statementSequenceNode, Node *statementNode);
+Node *MakeReturnStatementNode(Node *expressionNode);
+Node *MakeReturnVoidStatementNode();
+Node *StartFunctionSignatureArgumentsNode(Node *argumentNode);
+Node *AddFunctionSignatureArgumentNode(Node *argumentsNode, Node *argumentNode);
 Node *MakeEmptyFunctionSignatureArgumentsNode();
-Node* MakeFunctionSignatureNode(
+Node *MakeFunctionSignatureNode(
     Node *identifierNode,
-    Node* typeNode,
-    Node* argumentsNode,
-    Node* modifiersNode
-);
-Node* MakeFunctionDeclarationNode(
-    Node* functionSignatureNode,
-    Node* functionBodyNode
-);
-Node* MakeStructDeclarationNode(
+    Node *typeNode,
+    Node *argumentsNode,
+    Node *modifiersNode,
+    Node *genericArgumentsNode);
+Node *MakeFunctionDeclarationNode(
+    Node *functionSignatureNode,
+    Node *functionBodyNode);
+Node *MakeGenericDeclarationNode(Node *identifierNode, Node *constraintNode);
+Node *MakeEmptyGenericDeclarationsNode();
+Node *StartGenericDeclarationsNode(Node *genericDeclarationNode);
+Node *AddGenericDeclaration(
+    Node *genericDeclarationsNode,
+    Node *genericDeclarationNode);
+Node *MakeGenericArgumentNode(Node *typeNode);
+Node *MakeEmptyGenericArgumentsNode();
+Node *StartGenericArgumentsNode(Node *genericArgumentNode);
+Node *AddGenericArgument(Node *genericArgumentsNode, Node *genericArgumentNode);
+Node *MakeGenericTypeNode(char *name);
+Node *MakeStructDeclarationNode(
     Node *identifierNode,
-    Node *declarationSequenceNode
-);
-Node* StartDeclarationSequenceNode(
-    Node *declarationNode
-);
-Node* AddDeclarationNode(
     Node *declarationSequenceNode,
-    Node *declarationNode
-);
-Node *StartFunctionArgumentSequenceNode(
-    Node *argumentNode
-);
-Node *AddFunctionArgumentNode(
-    Node *argumentSequenceNode,
-    Node *argumentNode
-);
+    Node *genericArgumentsNode);
+Node *StartDeclarationSequenceNode(Node *declarationNode);
+Node *AddDeclarationNode(Node *declarationSequenceNode, Node *declarationNode);
+Node *StartFunctionArgumentSequenceNode(Node *argumentNode);
+Node *AddFunctionArgumentNode(Node *argumentSequenceNode, Node *argumentNode);
 Node *MakeEmptyFunctionArgumentSequenceNode();
-Node* MakeFunctionCallExpressionNode(
+Node *MakeFunctionCallExpressionNode(
     Node *identifierNode,
-    Node *argumentSequenceNode
-);
-Node* MakeAccessExpressionNode(
-    Node *accessee,
-    Node *accessor
-);
-Node* MakeAllocNode(
-    Node *typeNode
-);
-Node* MakeIfNode(
-    Node *expressionNode,
-    Node *statementSequenceNode
-);
-Node* MakeIfElseNode(
+    Node *argumentSequenceNode,
+    Node *genericArgumentsNode);
+Node *MakeSystemCallExpressionNode(
+    Node *identifierNode,
+    Node *argumentSequenceNode,
+    Node *genericArgumentsNode);
+Node *MakeAccessExpressionNode(Node *accessee, Node *accessor);
+Node *MakeAllocNode(Node *typeNode);
+Node *MakeIfNode(Node *expressionNode, Node *statementSequenceNode);
+Node *MakeIfElseNode(
     Node *ifNode,
-    Node *statementSequenceNode
+    Node *elseNode /* can be a conditional or a statement sequence */
 );
-Node* MakeForLoopNode(
+Node *MakeForLoopNode(
     Node *identifierNode,
     Node *startNumberNode,
     Node *endNumberNode,
-    Node *statementSequenceNode
-);
+    Node *statementSequenceNode);
+Node *MakeFieldInitNode(Node *identifierNode, Node *expressionNode);
+Node *StartStructInitFieldsNode(Node *fieldInitNode);
+Node *AddFieldInitNode(Node *structInitFieldsNode, Node *fieldInitNode);
+Node *MakeEmptyFieldInitNode();
+Node *MakeStructInitExpressionNode(Node *typeNode, Node *structInitFieldsNode);
 
-void PrintTree(Node *node, uint32_t tabCount);
-const char* SyntaxKindString(SyntaxKind syntaxKind);
+void PrintNode(Node *node, uint32_t tabCount);
+const char *SyntaxKindString(SyntaxKind syntaxKind);
 
-TypeTag* MakeTypeTag(Node *node);
-char* TypeTagToString(TypeTag *tag);
+/* Helper function for applying a void function generically over the children of
+ * an AST node. Used for functions that need to traverse the entire tree but
+ * only perform operations on a subset of node types. Such functions can match
+ * the syntaxKinds relevant to their purpose and invoke this function in all
+ * other cases. */
+void Recurse(Node *node, void (*func)(Node *));
+
+void LinkParentPointers(Node *node, Node *prev);
+
+TypeTag *MakeTypeTag(Node *node);
+char *TypeTagToString(TypeTag *tag);
+uint8_t TypeTagEqual(TypeTag *typeTagA, TypeTag *typeTagB);
+
+Node *LookupIdNode(Node *current, Node *prev, char *target);
 
 #endif /* WRAITH_AST_H */

src/identcheck.c

@@ -1,328 +0,0 @@
-#include <stdbool.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <string.h>
-
-#include "ast.h"
-#include "identcheck.h"
-
-IdNode* MakeIdNode(NodeType type, char *name, IdNode *parent) {
-    IdNode *node = (IdNode*)malloc(sizeof(IdNode));
-    node->type = type;
-    node->name = strdup(name);
-    node->parent = parent;
-    node->childCount = 0;
-    node->childCapacity = 0;
-    node->children = NULL;
-    node->typeTag = NULL;
-    return node;
-}
-
-void AddChildToNode(IdNode *node, IdNode *child) {
-    if (child == NULL) return;
-
-    if (node->children == NULL) {
-        node->childCapacity = 2;
-        node->children = (IdNode**) malloc(sizeof(IdNode*) * node->childCapacity);
-    } else if (node->childCount == node->childCapacity) {
-        node->childCapacity *= 2;
-        node->children = (IdNode**) realloc(node->children, sizeof(IdNode*) * node->childCapacity);
-    }
-
-    node->children[node->childCount] = child;
-    node->childCount += 1;
-}
-
-IdNode* MakeIdTree(Node *astNode, IdNode *parent) {
-    uint32_t i;
-    IdNode *mainNode;
-    switch (astNode->syntaxKind) {
-        case Assignment: {
-            if (astNode->children[0]->syntaxKind == Declaration) {
-                return MakeIdTree(astNode->children[0], parent);
-            } else {
-                for (i = 0; i < astNode->childCount; i++) {
-                    AddChildToNode(parent, MakeIdTree(astNode->children[i], parent));
-                }
-                return NULL;
-            }
-        }
-
-        case IfStatement: {
-            mainNode = MakeIdNode(OrderedScope, "if", parent);
-            Node *clause = astNode->children[0];
-            Node *stmtSeq = astNode->children[1];
-            for (i = 0; i < clause->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(clause->children[i], mainNode));
-            }
-            for (i = 0; i < stmtSeq->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(stmtSeq->children[i], mainNode));
-            }
-            break;
-        }
-
-        case IfElseStatement: {
-            Node *ifNode = astNode->children[0];
-            Node *elseStmts = astNode->children[1];
-            mainNode = MakeIdNode(OrderedScope, "if-else", parent);
-            IdNode *ifBranch = MakeIdTree(ifNode, mainNode);
-            IdNode *elseBranch = MakeIdNode(OrderedScope, "else", mainNode);
-
-            AddChildToNode(mainNode, ifBranch);
-            for (i = 0; i < elseStmts->childCount; i++) {
-                AddChildToNode(elseBranch, MakeIdTree(elseStmts->children[i], elseBranch));
-            }
-            AddChildToNode(mainNode, elseBranch);
-            break;
-        }
-
-        case ForLoop: {
-            Node *loopDecl = astNode->children[0];
-            Node *loopBody = astNode->children[3];
-            mainNode = MakeIdNode(OrderedScope, "for-loop", parent);
-            AddChildToNode(mainNode, MakeIdTree(loopDecl, mainNode));
-            for (i = 0; i < loopBody->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(loopBody->children[i], mainNode));
-            }
-            break;
-        }
-
-        case Declaration: {
-            mainNode = MakeIdNode(Variable, astNode->children[1]->value.string, parent);
-            mainNode->typeTag = MakeTypeTag(astNode);
-            astNode->children[1]->typeTag = mainNode->typeTag;
-            break;
-        }
-
-        case StructDeclaration: {
-            Node *idNode = astNode->children[0];
-            Node *declsNode = astNode->children[1];
-            mainNode = MakeIdNode(Struct, idNode->value.string, parent);
-            mainNode->typeTag = MakeTypeTag(astNode);
-            for (i = 0; i < declsNode->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(declsNode->children[i], mainNode));
-            }
-            break;
-        }
-
-        case FunctionDeclaration: {
-            Node *sigNode = astNode->children[0];
-            Node *funcNameNode = sigNode->children[0];
-            Node *funcArgsNode = sigNode->children[2];
-            Node *bodyStatementsNode = astNode->children[1];
-            mainNode = MakeIdNode(Function, funcNameNode->value.string, parent);
-            mainNode->typeTag = MakeTypeTag(astNode);
-            astNode->children[0]->children[0]->typeTag = mainNode->typeTag;
-            for (i = 0; i < funcArgsNode->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(funcArgsNode->children[i], mainNode));
-            }
-            for (i = 0; i < bodyStatementsNode->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(bodyStatementsNode->children[i], mainNode));
-            }
-            break;
-        }
-
-        case DeclarationSequence: {
-            mainNode = MakeIdNode(UnorderedScope, "", parent);
-            for (i = 0; i < astNode->childCount; i++) {
-                AddChildToNode(mainNode, MakeIdTree(astNode->children[i], mainNode));
-            }
-            break;
-        }
-
-        case Identifier: {
-            mainNode = MakeIdNode(Placeholder, astNode->value.string, parent);
-            IdNode *lookupNode = LookupId(mainNode, NULL, astNode->value.string);
-            if (lookupNode == NULL) {
-                fprintf(stderr, "wraith: Could not find IdNode for id %s\n", astNode->value.string);
-                TypeTag *tag = (TypeTag*)malloc(sizeof(TypeTag));
-                tag->type = Unknown;
-                astNode->typeTag = tag;
-            } else {
-                astNode->typeTag = lookupNode->typeTag;
-            }
-            break;
-        }
-
-        default: {
-            for (i = 0; i < astNode->childCount; i++) {
-                AddChildToNode(parent, MakeIdTree(astNode->children[i], parent));
-            }
-            return NULL;
-        }
-    }
-
-    astNode->idLink = mainNode;
-    return mainNode;
-}
-
-void PrintIdNode(IdNode *node) {
-    if (node == NULL) {
-        fprintf(stderr, "wraith: Attempted to call PrintIdNode with null value.\n");
-        return;
-    }
-
-    switch(node->type) {
-        case Placeholder:
-            printf("Placeholder (%s)\n", node->name);
-            break;
-        case OrderedScope:
-            printf("OrderedScope (%s)\n", node->name);
-            break;
-        case UnorderedScope:
-            printf("UnorderedScope (%s)\n", node->name);
-            break;
-        case Struct:
-            printf("%s : %s\n", node->name, TypeTagToString(node->typeTag));
-            break;
-        case Function:
-            printf("%s : Function<%s>\n", node->name, TypeTagToString(node->typeTag));
-            break;
-        case Variable:
-            printf("%s : %s\n", node->name, TypeTagToString(node->typeTag));
-            break;
-    }
-}
-
-void PrintIdTree(IdNode *tree, uint32_t tabCount) {
-    if (tree == NULL) {
-        fprintf(stderr, "wraith: Attempted to call PrintIdTree on a null value.\n");
-        return;
-    }
-
-    uint32_t i;
-    for (i = 0; i < tabCount; i++) {
-        printf("| ");
-    }
-
-    PrintIdNode(tree);
-
-    for (i = 0; i < tree->childCount; i++) {
-        PrintIdTree(tree->children[i], tabCount + 1);
-    }
-}
-
-
-int PrintAncestors(IdNode *node) {
-    if (node == NULL) return -1;
-
-    int i;
-    int indent = 1;
-    indent += PrintAncestors(node->parent);
-    for (i = 0; i < indent; i++) {
-        printf(" ");
-    }
-    PrintIdNode(node);
-    return indent;
-}
-
-IdNode* LookdownId(IdNode *root, NodeType targetType, char *targetName) {
-    if (root == NULL) {
-        fprintf(stderr, "wraith: Attempted to call LookdownId on a null value.\n");
-        return NULL;
-    }
-
-    IdNode *result = NULL;
-    IdNode **frontier = (IdNode**)malloc(sizeof(IdNode*));
-    frontier[0] = root;
-    uint32_t frontierCount = 1;
-
-    while (frontierCount > 0) {
-        IdNode *current = frontier[0];
-
-        if (current->type == targetType && strcmp(current->name, targetName) == 0) {
-            result = current;
-            break;
-        }
-
-        uint32_t i;
-        for(i = 1; i < frontierCount; i++) {
-            frontier[i-1] = frontier[i];
-        }
-        size_t newSize = frontierCount + current->childCount - 1;
-        if (frontierCount != newSize) {
-            frontier = (IdNode**) realloc(frontier, sizeof(IdNode*) * newSize);
-        }
-        for (i = 0; i < current->childCount; i++) {
-            frontier[frontierCount + i - 1] = current->children[i];
-        }
-        frontierCount = newSize;
-    }
-
-    free(frontier);
-    return result;
-}
-
-bool ScopeHasOrdering(IdNode *node) {
-    switch (node->type) {
-        case OrderedScope:
-        case Function:
-        case Variable: /* this is only technically true */
-            return true;
-        default:
-            return false;
-    }
-}
-
-IdNode* LookupId(IdNode *node, IdNode *prev, char *target) {
-    if (node == NULL) {
-        return NULL;
-    }
-
-    if (strcmp(node->name, target) == 0 && node->type != Placeholder) {
-        return node;
-    }
-
-    /* If this is the start of our search, we should not attempt to look at child nodes. Only
-     * looking up the scope tree is valid at this point.
-     *
-     * This has the notable side-effect that this function will return NULL if you attempt to look
-     * up a struct's internals starting from the node representing the struct itself. This is
-     * because an IdNode corresponds to the location *where an identifier is first declared.* Thus,
-     * an identifier has no knowledge of identifiers declared "inside" of it.
-     */
-    if (prev == NULL) {
-        return LookupId(node->parent, node, target);
-    }
-
-    /* If the current node forms an ordered scope then we want to prevent ourselves from looking
-     * up identifiers declared after the scope we have just come from.
-     */
-    uint32_t idxLimit;
-    if (ScopeHasOrdering(node)) {
-        uint32_t i;
-        for (i = 0, idxLimit = 0; i < node->childCount; i++, idxLimit++) {
-            if (node->children[i] == prev) {
-                break;
-            }
-        }
-    } else {
-        idxLimit = node->childCount;
-    }
-
-    uint32_t i;
-    for (i = 0; i < idxLimit; i++) {
-        IdNode *child = node->children[i];
-        if (child == prev || child->type == Placeholder) {
-            /* Do not inspect the node we just came from or placeholders. */
-            continue;
-        }
-
-        if (strcmp(child->name, target) == 0) {
-            return child;
-        }
-
-        if (child->type == Struct) {
-            uint32_t j;
-            for (j = 0; j < child->childCount; j++) {
-                IdNode *grandchild = child->children[j];
-                if (strcmp(grandchild->name, target) == 0) {
-                    return grandchild;
-                }
-            }
-        }
-    }
-
-    return LookupId(node->parent, node, target);
-}

src/identcheck.h

@@ -1,46 +0,0 @@
-/* Validates identifier usage in an AST. */
-
-#ifndef WRAITH_IDENTCHECK_H
-#define WRAITH_IDENTCHECK_H
-
-#include <stdint.h>
-
-#include "ast.h"
-
-struct TypeTag;
-struct Node;
-
-typedef enum NodeType {
-    Placeholder,
-    UnorderedScope,
-    OrderedScope,
-    Struct,
-    Function,
-    Variable
-} NodeType;
-
-typedef struct IdNode {
-    NodeType type;
-    char *name;
-    struct TypeTag *typeTag;
-    struct IdNode *parent;
-    struct IdNode **children;
-    uint32_t childCount;
-    uint32_t childCapacity;
-} IdNode;
-
-typedef struct IdStatus {
-    enum StatusCode {
-        Valid,
-    } StatusCode;
-} IdStatus;
-
-
-IdNode* MakeIdTree(struct Node *astNode, IdNode *parent);
-void PrintIdNode(IdNode *node);
-void PrintIdTree(IdNode *tree, uint32_t tabCount);
-int PrintAncestors(IdNode *node);
-IdNode* LookdownId(IdNode *root, NodeType targetType, char *targetName);
-IdNode* LookupId(IdNode *node, IdNode *prev, char* target);
-
-#endif /* WRAITH_IDENTCHECK_H */

src/main.c

@@ -1,9 +1,9 @@
-#include <stdlib.h>
 #include "../lib/dropt/dropt.h"
+#include <stdlib.h>
 
-#include "parser.h"
 #include "codegen.h"
-#include "identcheck.h"
+#include "parser.h"
+#include "validation.h"
 
 int main(int argc, char *argv[])
 {
@@ -14,10 +14,27 @@ int main(int argc, char *argv[])
     int exitCode = EXIT_SUCCESS;
 
     dropt_option options[] = {
-        { 'h', "help", "Shows help.", NULL, dropt_handle_bool, &showHelp, dropt_attr_halt },
-        { 'v', "parse-verbose", "Shows verbose parser output.", NULL, dropt_handle_bool, &parseVerbose },
-        { 'O', "optimize", "Sets optimization level of the output IR. Must be a value between 0 and 3.", "number", dropt_handle_uint, &optimizationLevel },
-        { 0 } /* Required sentinel value. */
+        {'h',
+         "help",
+         "Shows help.",
+         NULL,
+         dropt_handle_bool,
+         &showHelp,
+         dropt_attr_halt},
+        {'v',
+         "parse-verbose",
+         "Shows verbose parser output.",
+         NULL,
+         dropt_handle_bool,
+         &parseVerbose},
+        {'O',
+         "optimize",
+         "Sets optimization level of the output IR. Must be a value between 0 "
+         "and 3.",
+         "number",
+         dropt_handle_uint,
+         &optimizationLevel},
+        {0} /* Required sentinel value. */
     };
 
     dropt_context *droptContext = dropt_new_context(options);
@@ -33,10 +50,13 @@ int main(int argc, char *argv[])
     }
     else
     {
-        char** rest = dropt_parse(droptContext, -1, &argv[1]);
+        char **rest = dropt_parse(droptContext, -1, &argv[1]);
         if (dropt_get_error(droptContext) != dropt_error_none)
         {
-            fprintf(stderr, "wraith: %s\n", dropt_get_error_message(droptContext));
+            fprintf(
+                stderr,
+                "wraith: %s\n",
+                dropt_get_error_message(droptContext));
             exitCode = EXIT_FAILURE;
         }
         else if (showHelp)
@@ -65,12 +85,15 @@ int main(int argc, char *argv[])
                 }
                 else
                 {
-                    {
-                        IdNode *idTree = MakeIdTree(rootNode, NULL);
-                        PrintIdTree(idTree, /*tabCount=*/0);
-                        printf("\n");
-                        PrintTree(rootNode, /*tabCount=*/0);
-                    }
+                    LinkParentPointers(rootNode, NULL);
+                    /* FIXME: ValidateIdentifiers should return some sort of
+                              error status object. */
+                    ValidateIdentifiers(rootNode);
+                    TagIdentifierTypes(rootNode);
+                    ConvertCustomsToGenerics(rootNode);
+                    PrintNode(rootNode, 0);
+
+                    printf("Beginning codegen.\n");
                     exitCode = Codegen(rootNode, optimizationLevel);
                 }
             }

src/parser.c

@@ -1,7 +1,7 @@
 #include <stdio.h>
 
-#include "y.tab.h"
 #include "ast.h"
+#include "y.tab.h"
 
 extern FILE *yyin;
 extern int yydebug;
@@ -31,7 +31,7 @@ int Parse(char *inputFilename, Node **pRootNode, uint8_t parseVerbose)
     {
         if (parseVerbose)
         {
-            PrintTree(*pRootNode, 0);
+            PrintNode(*pRootNode, 0);
         }
     }
     else if (result == 1)

src/util.c

@@ -1,16 +1,38 @@
 #include "util.h"
 
-#include <string.h>
+#include <stdlib.h>
 
-char* strdup (const char* s)
+char *strdup(const char *s)
 {
-  size_t slen = strlen(s);
-  char* result = malloc(slen + 1);
-  if(result == NULL)
-  {
-    return NULL;
-  }
+    size_t slen = strlen(s);
+    char *result = (char *)malloc(sizeof(char) * (slen + 1));
+    if (result == NULL)
+    {
+        return NULL;
+    }
 
-  memcpy(result, s, slen+1);
-  return result;
-}
+    memcpy(result, s, slen + 1);
+    return result;
+}
+
+char *w_strcat(char *s, char *s2)
+{
+    size_t slen = strlen(s);
+    size_t slen2 = strlen(s2);
+    s = realloc(s, sizeof(char) * (slen + slen2 + 1));
+    strcat(s, s2);
+    return s;
+}
+
+uint64_t str_hash(char *str)
+{
+    uint64_t hash = 5381;
+    size_t c;
+
+    while ((c = *str++))
+    {
+        hash = ((hash << 5) + hash) + c; /* hash * 33 + c */
+    }
+
+    return hash;
+}

src/util.h

@@ -1,8 +1,11 @@
 #ifndef WRAITH_UTIL_H
 #define WRAITH_UTIL_H
 
+#include <stdint.h>
 #include <string.h>
 
-char* strdup (const char* s);
+char *strdup(const char *s);
+char *w_strcat(char *s, char *s2);
+uint64_t str_hash(char *str);
 
 #endif /* WRAITH_UTIL_H */

src/validation.c

@@ -0,0 +1,493 @@
+#include "validation.h"
+
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+Node *GetIdFromStruct(Node *structDecl)
+{
+    if (structDecl->syntaxKind != StructDeclaration)
+    {
+        fprintf(
+            stderr,
+            "wraith: Attempted to call GetIdFromStruct on node with kind: "
+            "%s.\n",
+            SyntaxKindString(structDecl->syntaxKind));
+        return NULL;
+    }
+
+    return structDecl->structDeclaration.identifier;
+}
+
+Node *GetIdFromFunction(Node *funcDecl)
+{
+    if (funcDecl->syntaxKind != FunctionDeclaration)
+    {
+        fprintf(
+            stderr,
+            "wraith: Attempted to call GetIdFromFunction on node with kind: "
+            "%s.\n",
+            SyntaxKindString(funcDecl->syntaxKind));
+        return NULL;
+    }
+
+    Node *sig = funcDecl->functionDeclaration.functionSignature;
+    return sig->functionSignature.identifier;
+}
+
+Node *GetIdFromDeclaration(Node *decl)
+{
+    if (decl->syntaxKind != Declaration)
+    {
+        fprintf(
+            stderr,
+            "wraith: Attempted to call GetIdFromDeclaration on node with kind: "
+            "%s.\n",
+            SyntaxKindString(decl->syntaxKind));
+    }
+
+    return decl->declaration.identifier;
+}
+
+bool AssignmentHasDeclaration(Node *assign)
+{
+    return (
+        assign->syntaxKind == Assignment &&
+        assign->assignmentStatement.left->syntaxKind == Declaration);
+}
+
+Node *GetIdFromAssignment(Node *assign)
+{
+    if (assign->syntaxKind != Assignment)
+    {
+        fprintf(
+            stderr,
+            "wraith: Attempted to call GetIdFromAssignment on node with kind: "
+            "%s.\n",
+            SyntaxKindString(assign->syntaxKind));
+    }
+
+    if (AssignmentHasDeclaration(assign))
+    {
+        return GetIdFromDeclaration(assign->assignmentStatement.left);
+    }
+
+    return NULL;
+}
+
+bool NodeMayHaveId(Node *node)
+{
+    switch (node->syntaxKind)
+    {
+    case StructDeclaration:
+    case FunctionDeclaration:
+    case Declaration:
+    case Assignment:
+        return true;
+    default:
+        return false;
+    }
+}
+
+Node *TryGetId(Node *node)
+{
+    switch (node->syntaxKind)
+    {
+    case Assignment:
+        return GetIdFromAssignment(node);
+    case Declaration:
+        return GetIdFromDeclaration(node);
+    case FunctionDeclaration:
+        return GetIdFromFunction(node);
+    case StructDeclaration:
+        return GetIdFromStruct(node);
+    default:
+        return NULL;
+    }
+}
+
+Node *LookupFunctionArgId(Node *funcDecl, char *target)
+{
+    Node *args = funcDecl->functionDeclaration.functionSignature
+                     ->functionSignature.arguments;
+
+    uint32_t i;
+    for (i = 0; i < args->functionArgumentSequence.count; i += 1)
+    {
+        Node *arg = args->functionArgumentSequence.sequence[i];
+        if (arg->syntaxKind != Declaration)
+        {
+            fprintf(
+                stderr,
+                "wraith: Encountered %s node in function signature args "
+                "list.\n",
+                SyntaxKindString(arg->syntaxKind));
+            continue;
+        }
+
+        Node *argId = GetIdFromDeclaration(arg);
+        if (argId != NULL && strcmp(target, argId->identifier.name) == 0)
+            return argId;
+    }
+
+    return NULL;
+}
+
+Node *LookupStructInternalId(Node *structDecl, char *target)
+{
+    Node *decls = structDecl->structDeclaration.declarationSequence;
+
+    uint32_t i;
+    for (i = 0; i < decls->declarationSequence.count; i += 1)
+    {
+        Node *match = TryGetId(decls->declarationSequence.sequence[i]);
+        if (match != NULL && strcmp(target, match->identifier.name) == 0)
+            return match;
+    }
+
+    return NULL;
+}
+
+Node *InspectNode(Node *node, char *target)
+{
+    /* If this node may have an identifier declaration inside it, attempt to
+     * look up the identifier
+     * node itself, returning it if it matches the given target name. */
+    if (NodeMayHaveId(node))
+    {
+        Node *candidateId = TryGetId(node);
+        if (candidateId != NULL &&
+            strcmp(target, candidateId->identifier.name) == 0)
+            return candidateId;
+    }
+
+    /* If the candidate node was not the one we wanted, but the node node is a
+     * function declaration, it's possible that the identifier we want is one of
+     * the function's parameters rather than the function's name itself. */
+    if (node->syntaxKind == FunctionDeclaration)
+    {
+        Node *match = LookupFunctionArgId(node, target);
+        if (match != NULL)
+            return match;
+    }
+
+    /* Likewise if the node node is a struct declaration, inspect the struct's
+     * internals
+     * to see if a top-level definition is the one we're looking for. */
+    if (node->syntaxKind == StructDeclaration)
+    {
+        Node *match = LookupStructInternalId(node, target);
+        if (match != NULL)
+            return match;
+    }
+
+    return NULL;
+}
+
+/* FIXME: Handle staged lookups for AccessExpressions. */
+/* FIXME: Similar to above, disallow inspection of struct internals outside of
+ *        AccessExpressions. */
+Node *LookupId(Node *current, Node *prev, char *target)
+{
+    if (current == NULL)
+        return NULL;
+
+    Node *match;
+
+    /* First inspect the current node to see if it contains the target
+     * identifier. */
+    match = InspectNode(current, target);
+    if (match != NULL)
+        return match;
+
+    /* If this is the start of our search, we should not attempt to look at
+     * child nodes. Only looking up the AST is valid at this point.
+     *
+     * This has the notable side-effect that this function will return NULL if
+     * you attempt to look up a struct's internals starting from the node
+     * representing the struct itself. The same is true for functions. */
+    if (prev == NULL)
+        return LookupId(current->parent, current, target);
+
+    uint32_t i;
+    uint32_t idxLimit;
+    switch (current->syntaxKind)
+    {
+    case DeclarationSequence:
+        for (i = 0; i < current->declarationSequence.count; i += 1)
+        {
+            Node *decl = current->declarationSequence.sequence[i];
+            match = InspectNode(decl, target);
+            if (match != NULL)
+                return match;
+        }
+        break;
+    case StatementSequence:
+        idxLimit = current->statementSequence.count;
+        for (i = 0; i < current->statementSequence.count; i += 1)
+        {
+            if (current->statementSequence.sequence[i] == prev)
+            {
+                idxLimit = i;
+                break;
+            }
+        }
+
+        for (i = 0; i < idxLimit; i += 1)
+        {
+            Node *stmt = current->statementSequence.sequence[i];
+            if (stmt == prev)
+                break;
+
+            match = InspectNode(stmt, target);
+            if (match != NULL)
+                return match;
+        }
+        break;
+    }
+
+    return LookupId(current->parent, current, target);
+}
+
+/* FIXME: This function should be extended to handle multi-stage ID lookups for
+ *        AccessExpression nodes. */
+/* FIXME: Make this function return an error status object of some kind.
+ *        A non-OK status should halt compilation. */
+void ValidateIdentifiers(Node *node)
+{
+    if (node == NULL)
+        return;
+
+    /* Skip over generic arguments. They contain Identifiers but are not
+     * actually identifiers, they declare types. */
+    if (node->syntaxKind == GenericDeclarations)
+        return;
+
+    if (node->syntaxKind != Identifier)
+    {
+        Recurse(node, *ValidateIdentifiers);
+        return;
+    }
+
+    char *name = node->identifier.name;
+    Node *decl = LookupId(node, NULL, name);
+    if (decl == NULL)
+    {
+        /* FIXME: Express this case as an error with AST information, see the
+         *        FIXME comment above. */
+        fprintf(
+            stderr,
+            "wraith: Could not find definition of identifier %s.\n",
+            name);
+    }
+}
+
+/* FIXME: This function should be extended to handle multi-stage ID lookups for
+ *        AccessExpression nodes. */
+void TagIdentifierTypes(Node *node)
+{
+    if (node == NULL)
+        return;
+
+    switch (node->syntaxKind)
+    {
+    case AllocExpression:
+        node->typeTag = MakeTypeTag(node);
+        break;
+
+    case Declaration:
+        node->declaration.identifier->typeTag = MakeTypeTag(node);
+        break;
+
+    case FunctionDeclaration:
+        node->functionDeclaration.functionSignature->functionSignature
+            .identifier->typeTag = MakeTypeTag(node);
+        break;
+
+    case StructDeclaration:
+        node->structDeclaration.identifier->typeTag = MakeTypeTag(node);
+        break;
+
+    case GenericDeclaration:
+        node->genericDeclaration.identifier->typeTag = MakeTypeTag(node);
+        break;
+
+    case Type:
+        node->typeTag = MakeTypeTag(node);
+        break;
+
+    case Identifier:
+    {
+        if (node->typeTag != NULL)
+            return;
+
+        char *name = node->identifier.name;
+        Node *declaration = LookupId(node, NULL, name);
+        /* FIXME: Remove this case once ValidateIdentifiers returns error status
+         * info and halts compilation. See ValidateIdentifiers FIXME. */
+        if (declaration == NULL)
+        {
+            TypeTag *tag = (TypeTag *)malloc(sizeof(TypeTag));
+            tag->type = Unknown;
+            node->typeTag = tag;
+        }
+        else
+        {
+            node->typeTag = declaration->typeTag;
+        }
+        break;
+    }
+    }
+
+    Recurse(node, *TagIdentifierTypes);
+}
+
+Node *LookupType(Node *current, char *target)
+{
+    if (current == NULL)
+        return NULL;
+
+    switch (current->syntaxKind)
+    {
+    /* If we've encountered a function declaration, check to see if it's generic
+     * and, if so, if one of its type parameters is the target. */
+    case FunctionDeclaration:
+    {
+        Node *typeArgs = current->functionDeclaration.functionSignature
+                             ->functionSignature.genericDeclarations;
+        uint32_t i;
+        for (i = 0; i < typeArgs->genericDeclarations.count; i += 1)
+        {
+            Node *arg = typeArgs->genericDeclarations.declarations[i];
+            Node *argId = arg->genericDeclaration.identifier;
+            char *argName = argId->identifier.name;
+            /* note: return the GenericDeclaration, not the Identifier, so that
+             * the caller can differentiate between generics and customs. */
+            if (strcmp(target, argName) == 0)
+            {
+                return arg;
+            }
+        }
+
+        return LookupType(current->parent, target);
+    }
+
+    case StructDeclaration:
+    {
+        uint32_t i;
+        Node *typeArgs = current->structDeclaration.genericDeclarations;
+        for (i = 0; i < typeArgs->genericDeclarations.count; i += 1)
+        {
+            Node *arg = typeArgs->genericDeclarations.declarations[i];
+            Node *argId = arg->genericDeclaration.identifier;
+            char *argName = argId->identifier.name;
+            /* note: return the GenericDeclaration, not the Identifier, so that
+             * the caller can differentiate between generics and customs. */
+            if (strcmp(target, argName) == 0)
+            {
+                return arg;
+            }
+        }
+
+        Node *structId = GetIdFromStruct(current);
+        if (strcmp(target, structId->identifier.name) == 0)
+        {
+            return structId;
+        }
+
+        return LookupType(current->parent, target);
+    }
+
+    /* If we encounter a declaration sequence, search each of its children for
+     * struct definitions in case one of them is the target. */
+    case DeclarationSequence:
+    {
+        uint32_t i;
+        for (i = 0; i < current->declarationSequence.count; i += 1)
+        {
+            Node *decl = current->declarationSequence.sequence[i];
+            if (decl->syntaxKind == StructDeclaration)
+            {
+                Node *structId = GetIdFromStruct(decl);
+                if (strcmp(target, structId->identifier.name) == 0)
+                    return structId;
+            }
+        }
+
+        return LookupType(current->parent, target);
+    }
+
+    default:
+        return LookupType(current->parent, target);
+    }
+}
+
+/* FIXME: This function should be modified to handle type parameters over
+ * structs. */
+void ConvertCustomsToGenerics(Node *node)
+{
+    if (node == NULL)
+        return;
+
+    switch (node->syntaxKind)
+    {
+    case Declaration:
+    {
+        Node *id = node->declaration.identifier;
+        Node *type = node->declaration.type->type.typeNode;
+        if (type->syntaxKind == CustomTypeNode)
+        {
+            char *target = id->typeTag->value.customType;
+            Node *typeLookup = LookupType(node, target);
+            if (typeLookup != NULL &&
+                typeLookup->syntaxKind == GenericDeclaration)
+            {
+                id->typeTag->type = Generic;
+                free(node->declaration.type);
+                node->declaration.type =
+                    MakeGenericTypeNode(id->typeTag->value.genericType);
+            }
+        }
+        break;
+    }
+
+    case FunctionSignature:
+    {
+        Node *id = node->functionSignature.identifier;
+        Node *type = node->functionSignature.type->type.typeNode;
+        if (type->syntaxKind == CustomTypeNode)
+        {
+            char *target = id->typeTag->value.customType;
+            Node *typeLookup = LookupType(node, target);
+            if (typeLookup != NULL &&
+                typeLookup->syntaxKind == GenericDeclaration)
+            {
+                id->typeTag->type = Generic;
+                free(node->functionSignature.type);
+                node->functionSignature.type =
+                    MakeGenericTypeNode(id->typeTag->value.genericType);
+            }
+        }
+        break;
+    }
+
+    case GenericArgument:
+    {
+        Node *typeNode = node->genericArgument.type;
+        if (typeNode->typeTag->type == Custom)
+        {
+            char *target = typeNode->typeTag->value.customType;
+            Node *typeLookup = LookupType(node, target);
+            if (typeLookup != NULL &&
+                typeLookup->syntaxKind == GenericDeclaration)
+            {
+                typeNode->typeTag->type = Generic;
+            }
+        }
+        break;
+    }
+    }
+
+    Recurse(node, *ConvertCustomsToGenerics);
+}

src/validation.h

@@ -0,0 +1,10 @@
+#ifndef WRAITH_VALIDATION_H
+#define WRAITH_VALIDATION_H
+
+#include "ast.h"
+
+void ValidateIdentifiers(Node *node);
+void TagIdentifierTypes(Node *node);
+void ConvertCustomsToGenerics(Node *node);
+
+#endif /* WRAITH_VALIDATION_H */