From 0f4eb113a97c5f13b7e9ec2e1097408be655721c Mon Sep 17 00:00:00 2001
From: mrexodia <mr.exodia.tpodt@gmail.com>
Date: Tue, 7 Jun 2016 03:17:56 +0200
Subject: [PATCH] basics of AST and parsing

---
 btparser/ast.h                    |  75 +++++++++++++
 btparser/btparser.vcxproj         |   5 +
 btparser/btparser.vcxproj.filters |  11 ++
 btparser/lexer.cpp                |   4 +-
 btparser/lexer.h                  |   2 +-
 btparser/main.cpp                 |  17 ++-
 btparser/parser.cpp               | 178 ++++++++++++++++++++++++++++++
 btparser/parser.h                 |  42 +++++++
 btparser/tests/simple.bt          |   1 +
 9 files changed, 331 insertions(+), 4 deletions(-)
 create mode 100644 btparser/ast.h
 create mode 100644 btparser/parser.cpp
 create mode 100644 btparser/parser.h
 create mode 100644 btparser/tests/simple.bt

diff --git a/btparser/ast.h b/btparser/ast.h
new file mode 100644
index 0000000..5d35d12
--- /dev/null
+++ b/btparser/ast.h
@@ -0,0 +1,75 @@
+#pragma once
+
+#include "lexer.h"
+#include <memory>
+
+namespace AST
+{
+    using namespace std;
+
+    template<class T>
+    using uptr = unique_ptr<T>;
+
+    template <class T, class... Args>
+    static typename enable_if<!is_array<T>::value, unique_ptr<T>>::type make_uptr(Args &&... args)
+    {
+        return uptr<T>(new T(std::forward<Args>(args)...));
+    }
+
+    using Operator = Lexer::Token; //TODO
+    using Type = Lexer::Token; //TODO
+
+    class StatDecl //base class for every node
+    {
+    public:
+        virtual ~StatDecl() {}
+    };
+
+    class Stat : public StatDecl //statement (expressions, control, block)
+    {
+    };
+
+    class Block : public Stat //block
+    {
+        vector<uptr<StatDecl>> mStatDecls;
+    public:
+        explicit Block(vector<uptr<StatDecl>> statDecls)
+            : mStatDecls(move(statDecls)) {}
+    };
+
+    class Expr : public Stat //expression
+    {
+    public:
+        virtual ~Expr() {}
+    };
+
+    class Return : public Stat
+    {
+        uptr<Expr> mExpr;
+    public:
+        explicit Return(uptr<Expr> expr)
+            : mExpr(move(expr)) {}
+    };
+
+    class Decl : public StatDecl //declaration (variables/types)
+    {
+    };
+
+    class Builtin : public Decl //built-in declaration (int x)
+    {
+        Type mType;
+        string mId;
+    public:
+        explicit Builtin(Type type, const string & id)
+            : mType(type), mId(id) {}
+    };
+
+    class Struct : public Decl //struct (can contain code, not just declarations)
+    {
+        string mId;
+        uptr<Block> mBlock;
+    public:
+        explicit Struct(const string & id, uptr<Block> block)
+            : mId(id), mBlock(move(block)) {}
+    };
+};
\ No newline at end of file
diff --git a/btparser/btparser.vcxproj b/btparser/btparser.vcxproj
index e72d0c2..af2e056 100644
--- a/btparser/btparser.vcxproj
+++ b/btparser/btparser.vcxproj
@@ -22,9 +22,11 @@
     <ClCompile Include="filehelper.cpp" />
     <ClCompile Include="lexer.cpp" />
     <ClCompile Include="main.cpp" />
+    <ClCompile Include="parser.cpp" />
     <ClCompile Include="stringutils.cpp" />
   </ItemGroup>
   <ItemGroup>
+    <ClInclude Include="ast.h" />
     <ClInclude Include="dynamicmem.h" />
     <ClInclude Include="filehelper.h" />
     <ClInclude Include="handle.h" />
@@ -34,6 +36,9 @@
     <ClInclude Include="stringutils.h" />
     <ClInclude Include="testfiles.h" />
   </ItemGroup>
+  <ItemGroup>
+    <None Include="parser.h" />
+  </ItemGroup>
   <PropertyGroup Label="Globals">
     <ProjectGuid>{B0411C78-2F06-49E0-8DE9-5C52A466F5DE}</ProjectGuid>
     <RootNamespace>btparser</RootNamespace>
diff --git a/btparser/btparser.vcxproj.filters b/btparser/btparser.vcxproj.filters
index 19271c0..526b5cc 100644
--- a/btparser/btparser.vcxproj.filters
+++ b/btparser/btparser.vcxproj.filters
@@ -27,6 +27,9 @@
     <ClCompile Include="lexer.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
+    <ClCompile Include="parser.cpp">
+      <Filter>Source Files</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <ClInclude Include="filehelper.h">
@@ -53,5 +56,13 @@
     <ClInclude Include="lexer.h">
       <Filter>Header Files</Filter>
     </ClInclude>
+    <ClInclude Include="ast.h">
+      <Filter>Header Files</Filter>
+    </ClInclude>
+  </ItemGroup>
+  <ItemGroup>
+    <None Include="parser.h">
+      <Filter>Header Files</Filter>
+    </None>
   </ItemGroup>
 </Project>
\ No newline at end of file
diff --git a/btparser/lexer.cpp b/btparser/lexer.cpp
index 81dba36..dd7daf5 100644
--- a/btparser/lexer.cpp
+++ b/btparser/lexer.cpp
@@ -51,14 +51,14 @@ bool Lexer::DoLexing(std::vector<TokenState> & tokens, std::string & error)
     {
         auto token = getToken();
         mState.Token = token;
-        if (token == tok_eof)
-            break;
         if (token == tok_error)
         {
             error = StringUtils::sprintf("line %d, col %d: %s", mState.CurLine + 1, mState.LineIndex, mError.c_str());
             return false;
         }
         tokens.push_back(mState);
+        if (token == tok_eof)
+            break;
     }
     return true;
 }
diff --git a/btparser/lexer.h b/btparser/lexer.h
index 3e55583..11c55aa 100644
--- a/btparser/lexer.h
+++ b/btparser/lexer.h
@@ -11,7 +11,7 @@ public:
     enum Token
     {
         //status tokens
-        tok_eof = -10000,
+        tok_eof,
         tok_error,
 
         //keywords
diff --git a/btparser/main.cpp b/btparser/main.cpp
index 365e73c..0a42ad1 100644
--- a/btparser/main.cpp
+++ b/btparser/main.cpp
@@ -3,6 +3,7 @@
 #include "testfiles.h"
 #include "lexer.h"
 #include "filehelper.h"
+#include "parser.h"
 
 bool TestLexer(Lexer & lexer, const std::string & filename)
 {
@@ -84,13 +85,27 @@ void DebugLexerTests(bool output = true)
         DebugLexer(lexer, file, output);
 }
 
+bool DebugParser(const std::string & filename)
+{
+    Parser parser;
+    std::string error;
+    if(!parser.ParseFile("tests\\" + filename, error))
+    {
+        printf("ParseFile failed: %s\n", error.c_str());
+        return false;
+    }
+    puts("ParseFile success!");
+    return true;
+}
+
 int main()
 {
     //GenerateExpectedTests();
     auto ticks = GetTickCount();
+    DebugParser("simple.bt");
     //Lexer lexer;
     //DebugLexer(lexer, "AndroidManifestTemplate.bt", false);
-    RunLexerTests();
+    //RunLexerTests();
     printf("finished in %ums\n", GetTickCount() - ticks);
     system("pause");
     return 0;
diff --git a/btparser/parser.cpp b/btparser/parser.cpp
new file mode 100644
index 0000000..9c39115
--- /dev/null
+++ b/btparser/parser.cpp
@@ -0,0 +1,178 @@
+#include "parser.h"
+#include "stringutils.h"
+
+using namespace AST;
+
+Parser::Parser()
+    : CurToken(Lexer::TokenState())
+{
+}
+
+bool Parser::ParseFile(const string & filename, string & error)
+{
+    if (!mLexer.ReadInputFile(filename))
+    {
+        error = "failed to read input file";
+        return false;
+    }
+    if (!mLexer.DoLexing(mTokens, error))
+        return false;
+    CurToken = mTokens[0];
+    mBinaryTemplate = ParseBinaryTemplate();
+    return !!mBinaryTemplate;
+}
+
+void Parser::NextToken()
+{
+    if (mIndex < mTokens.size() - 1)
+    {
+        mIndex++;
+        CurToken = mTokens[mIndex];
+    }
+}
+
+void Parser::ReportError(const std::string & error)
+{
+    mErrors.push_back(Error(error));
+}
+
+uptr<Block> Parser::ParseBinaryTemplate()
+{
+    vector<uptr<StatDecl>> statDecls;
+    while (true)
+    {
+        auto statDecl = ParseStatDecl();
+        if (!statDecl)
+            break;
+        statDecls.push_back(move(statDecl));
+    }
+    return make_uptr<Block>(move(statDecls));
+}
+
+uptr<StatDecl> Parser::ParseStatDecl()
+{
+    auto decl = ParseDecl();
+    if (decl)
+        return move(decl);
+
+    auto stat = ParseStat();
+    if (stat)
+        return move(stat);
+
+    ReportError("failed to parse StatDecl");
+    return nullptr;
+}
+
+uptr<Stat> Parser::ParseStat()
+{
+    auto block = ParseBlock();
+    if (block)
+        return move(block);
+
+    auto expr = ParseExpr();
+    if (expr)
+        return move(expr);
+
+    auto ret = ParseReturn();
+    if (ret)
+        return move(ret);
+
+    ReportError("failed to parse Stat");
+    return nullptr;
+}
+
+uptr<Block> Parser::ParseBlock()
+{
+    if (CurToken.Token != Lexer::tok_bropen) //'{'
+        return nullptr;
+    NextToken();
+
+    vector<uptr<StatDecl>> statDecls;
+
+    if (CurToken.Token == Lexer::tok_brclose) //'}'
+    {
+        NextToken();
+        return make_uptr<Block>(move(statDecls));
+    }
+
+    ReportError("failed to parse Block");
+    return nullptr;
+}
+
+uptr<Expr> Parser::ParseExpr()
+{
+    return nullptr;
+}
+
+uptr<Return> Parser::ParseReturn()
+{
+    if (CurToken.Token == Lexer::tok_return)
+    {
+        NextToken();
+        auto expr = ParseExpr();
+        if (!expr)
+        {
+            ReportError("failed to parse Return (ParseExpr failed)");
+            return nullptr;
+        }
+        return make_uptr<Return>(move(expr));
+    }
+    return nullptr;
+}
+
+uptr<Decl> Parser::ParseDecl()
+{
+    auto builtin = ParseBuiltin();
+    if (builtin)
+        return move(builtin);
+    auto stru = ParseStruct();
+    if (stru)
+        return move(stru);
+    return nullptr;
+}
+
+uptr<Builtin> Parser::ParseBuiltin()
+{
+    if (CurToken.Token == Lexer::tok_uint) //TODO: properly handle types
+    {
+        auto type = CurToken.Token;
+        NextToken();
+        if (CurToken.Token != Lexer::tok_identifier)
+        {
+            ReportError("failed to parse Builtin (no identifier)");
+            return nullptr;
+        }
+        auto id = CurToken.IdentifierStr;
+        NextToken();
+        if (CurToken.Token != Lexer::tok_semic)
+        {
+            ReportError("failed to parse Builtin (no semicolon)");
+            return nullptr;
+        }
+        NextToken();
+        return make_uptr<Builtin>(type, id);
+    }
+    return nullptr;
+}
+
+uptr<Struct> Parser::ParseStruct()
+{
+    if (CurToken.Token == Lexer::tok_struct)
+    {
+        NextToken();
+        string id;
+        if (CurToken.Token == Lexer::tok_identifier)
+        {
+            id = CurToken.IdentifierStr;
+            NextToken();
+        }
+        auto block = ParseBlock();
+        if (!block)
+        {
+            ReportError("failed to parse Struct (ParseBlock)");
+            return nullptr;
+        }
+        return make_uptr<Struct>(id, move(block));
+    }
+    return nullptr;
+}
diff --git a/btparser/parser.h b/btparser/parser.h
new file mode 100644
index 0000000..90271a9
--- /dev/null
+++ b/btparser/parser.h
@@ -0,0 +1,42 @@
+#pragma once
+
+#include "lexer.h"
+#include "ast.h"
+
+class Parser
+{
+public:
+    struct Error
+    {
+        explicit Error(const std::string & text)
+            : text(text) {}
+
+        std::string text;
+    };
+
+    explicit Parser();
+    bool ParseFile(const std::string & filename, std::string & error);
+
+private:
+    Lexer mLexer;
+    std::vector<Lexer::TokenState> mTokens;
+    size_t mIndex = 0;
+    AST::uptr<AST::Block> mBinaryTemplate = nullptr;
+    std::vector<Error> mErrors;
+
+    Lexer::TokenState CurToken;
+    void NextToken();
+    void ReportError(const std::string & error);
+
+    AST::uptr<AST::Block> ParseBinaryTemplate();
+    AST::uptr<AST::StatDecl> ParseStatDecl();
+
+    AST::uptr<AST::Stat> ParseStat();
+    AST::uptr<AST::Block> ParseBlock();
+    AST::uptr<AST::Expr> ParseExpr();
+    AST::uptr<AST::Return> ParseReturn();
+
+    AST::uptr<AST::Decl> ParseDecl();
+    AST::uptr<AST::Builtin> ParseBuiltin();
+    AST::uptr<AST::Struct> ParseStruct();
+};
\ No newline at end of file
diff --git a/btparser/tests/simple.bt b/btparser/tests/simple.bt
new file mode 100644
index 0000000..f3c4d04
--- /dev/null
+++ b/btparser/tests/simple.bt
@@ -0,0 +1 @@
+uint x;
\ No newline at end of file