Merge pull request #5 from 0xpantera/refactor/compiler-driver-pipeline

Refactor compiler driver pipeline

Author: 0xpantera

CHANGELOG.md

@@ -70,4 +70,22 @@
 - New assembly instructions for binary operations (add, sub, imul, idiv)
 - Proper handling of division and remainder with EAX/EDX registers
 - Sign extension support using cdq instruction
-- Updated instruction fix-up pass for binary operation constraints
+- Updated instruction fix-up pass for binary operation constraints
+
+## 0.4.0.0 -- 2024-12-03
+
+### Added
+- Hierarchical modules for better code organization:
+  - Halcyon.Core
+  - Halcyon.Frontend  
+  - Halcyon.Backend
+  - Halcyon.Driver
+
+### Changed
+- Split Driver.Pipeline into focused modules:
+  - Driver.Stages for pure compilation stages
+  - Driver.External for GCC interactions 
+  - Driver.Output for handling stage results
+  - Driver.Pipeline remains as high-level orchestration
+- Reorganized module exports in cabal file
+- Improved module documentation and organization

README.md

@@ -8,19 +8,34 @@ The compiler currently handles C programs with unary operators and integer const
 
 ```c
 int main(void) {
-    return ~(-42);
+    return -((42 * 10) / (2 + 3));  // Returns -84
 }
 ```
 
 ### Compilation Pipeline
 
-The compiler processes source code through the following stages:
+The compiler is organized into several major subsystems:
 
-1. **Lexical Analysis**: Breaks source code into a sequence of tokens
-2. **Parsing**: Converts tokens into an Abstract Syntax Tree (AST)
-3. **TACKY Generation**: Transforms AST into TACKY intermediate representation
-4. **Code Generation**: Transforms AST into x86_64 assembly
-5. **Code Emission**: Outputs the assembly code to an executable
+- **Frontend** - Parsing and analysis
+  - Lexical analysis (breaks source into tokens)
+  - Parsing (converts tokens to AST)
+  - Token definitions
+
+- **Core** - Core data types and compiler infrastructure
+  - AST, Assembly, and TACKY intermediate representations
+  - Compiler monad and error handling
+
+- **Backend** - Code generation
+  - TACKY to assembly conversion
+  - Register allocation
+  - Assembly output
+
+- **Driver** - Pipeline coordination
+  - Command line interface
+  - Compilation pipeline stages
+  - External tool integration (GCC for preprocessing and linking)
+
+Each subsystem is organized as a hierarchical module that provides a clean interface to its functionality while hiding implementation details.
 
 ### Internal Representations
 
@@ -85,20 +100,24 @@ Programs are represented internally using a series of increasingly lower-level d
   ├── lib/                           # Main library code
   │   ├── Halcyon.hs                 # Library entry point
   │   └── Halcyon/                   # Core modules
+  │       ├── Backend.hs             # Backend subsystem interface
   │       ├── Backend/               # Code generation and emission
   │       │   ├── Codegen.hs         # TACKY to Assembly conversion
   │       │   ├── Emit.hs            # Assembly to text output
   │       │   └── ReplacePseudos.hs  # Register/stack allocation
+  │       ├── Core.hs                # Core subsystem interface
   │       ├── Core/                  # Core data types and utilities
   │       │   ├── Assembly.hs        # Assembly representation
   │       │   ├── Ast.hs             # C language AST
   │       │   ├── Monad.hs           # Compiler monad stack
   │       │   ├── Settings.hs        # Compiler settings and types
   │       │   ├── Tacky.hs           # TACKY IR definition
   │       │   └── TackyGen.hs        # AST to TACKY transformation
+  │       ├── Driver.hs              # Driver subsystem interface
   │       ├── Driver/                # Compiler driver
   │       │   ├── Cli.hs             # Command line interface
   │       │   └── Pipeline.hs        # Compilation pipeline
+  │       ├── Frontend.hs          # Frontend subsystem interface
   │       └── Frontend/              # Parsing and analysis
   │           ├── Lexer.hs           # Lexical analysis
   │           ├── Parse.hs           # Parsing

halcyon.cabal

@@ -1,6 +1,6 @@
 cabal-version:      3.0
 name:               halcyon
-version:            0.3.0.0
+version:            0.4.0.0
 -- synopsis:
 -- description:
 license:            BSD-3-Clause
@@ -19,22 +19,36 @@ common warnings
 library
     import:           warnings
     exposed-modules:
-          Halcyon
-        , Halcyon.Core.Settings
-        , Halcyon.Core.Ast
+        -- Main module
+        Halcyon
+        -- Hierarchical modules
+        , Halcyon.Core
+        , Halcyon.Frontend
+        , Halcyon.Backend 
+        , Halcyon.Driver
+        -- Core types (needed for qualified imports)
         , Halcyon.Core.Assembly
+        , Halcyon.Core.Ast
+        , Halcyon.Core.Tacky
+    other-modules:
+        -- Core internals
+        Halcyon.Core.Settings
         , Halcyon.Core.Monad
-        , Halcyon.Core.Tacky 
         , Halcyon.Core.TackyGen
+        -- Frontend internals  
         , Halcyon.Frontend.Lexer
         , Halcyon.Frontend.Parse
         , Halcyon.Frontend.Tokens
+        -- Backend internals
         , Halcyon.Backend.Codegen
         , Halcyon.Backend.Emit
         , Halcyon.Backend.ReplacePseudos
+        -- Driver internals
         , Halcyon.Driver.Cli
         , Halcyon.Driver.Pipeline
-    other-modules:
+        , Halcyon.Driver.External
+        , Halcyon.Driver.Output
+        , Halcyon.Driver.Stages
     -- other-extensions:
     build-depends:    base ^>=4.20.0.0
         , bytestring
@@ -71,6 +85,7 @@ test-suite halcyon-test
         , Test.Tacky
         , Test.Assembly
         , Test.Pipeline
+        , Test.Common
     -- other-extensions:
     type:             exitcode-stdio-1.0
     hs-source-dirs:   test

lib/Halcyon.hs

@@ -7,13 +7,10 @@ module Halcyon
   , Target(..)
   , Program(..)
     -- * Compiler Pipeline
-  , module Halcyon.Driver.Pipeline
-    -- * Command Line Interface
-  , module Halcyon.Driver.Cli
+  , module Halcyon.Driver
   ) where
 
 import Halcyon.Core.Settings
 import Halcyon.Core.Ast (Program(..))
-import Halcyon.Driver.Pipeline
-import Halcyon.Driver.Cli
-import Halcyon.Frontend.Parse ( parseTokens )
+import Halcyon.Driver
+import Halcyon.Frontend ( parseTokens )

lib/Halcyon/Backend.hs

@@ -0,0 +1,21 @@
+{-# LANGUAGE OverloadedStrings #-}
+module Halcyon.Backend
+  ( -- * Code Generation
+    gen  -- from Codegen  
+    -- * Assembly Output 
+  , emitProgram -- from Emit
+    -- * Register/Stack Management
+  , replacePseudos
+  , fixupProgram
+  , ReplacePseudosError(..)
+  , PseudoM
+  ) where
+
+import Halcyon.Backend.Codegen (gen)
+import Halcyon.Backend.Emit (emitProgram)
+import Halcyon.Backend.ReplacePseudos 
+  ( replacePseudos
+  , fixupProgram
+  , ReplacePseudosError(..)
+  , PseudoM
+  )

lib/Halcyon/Core.hs

@@ -0,0 +1,12 @@
+{-# LANGUAGE OverloadedStrings #-}
+module Halcyon.Core
+  ( -- * IR Generation
+    module Halcyon.Core.TackyGen
+    -- * Compiler Infrastructure
+  , module Halcyon.Core.Monad
+  , module Halcyon.Core.Settings
+  ) where
+
+import Halcyon.Core.Monad
+import Halcyon.Core.Settings  
+import Halcyon.Core.TackyGen

lib/Halcyon/Driver.hs

@@ -0,0 +1,21 @@
+{-# LANGUAGE OverloadedStrings #-}
+module Halcyon.Driver
+  ( -- * Command Line Interface
+    module Halcyon.Driver.Cli
+    -- * Main Pipeline
+  , processFile
+  , StageResult(..)
+    -- * Pipeline Stages 
+  , module Halcyon.Driver.Stages
+    -- * External Tools
+  , module Halcyon.Driver.External
+    -- * Stage Output
+  , module Halcyon.Driver.Output
+  ) where
+
+import Halcyon.Driver.Cli
+import Halcyon.Driver.External 
+import Halcyon.Driver.Output
+import Halcyon.Driver.Pipeline (processFile)
+import Halcyon.Driver.Stages
+import Halcyon.Core.Settings (StageResult(..))

lib/Halcyon/Driver/External.hs

@@ -0,0 +1,52 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE RecordWildCards #-}
+module Halcyon.Driver.External
+  ( -- * GCC Interface
+    preprocess
+  , compileToExecutable
+  ) where
+
+import Control.Monad (unless)
+import Control.Monad.IO.Class (liftIO)
+import Data.Text (Text)
+import Data.Text qualified as T
+import Data.Text.IO qualified as TIO
+import System.Exit (ExitCode(..))
+import System.FilePath (replaceExtension, takeDirectory, takeFileName, dropExtension, (</>))
+import System.Directory (makeAbsolute)
+import System.Process.Typed (proc, readProcess)
+
+import Halcyon.Core.Monad (MonadCompiler, CompilerError(..), throwError)
+import Halcyon.Driver.Cli (AppOptions(..))
+
+-- | Preprocess source file using GCC
+preprocess :: MonadCompiler m => AppOptions -> m FilePath
+preprocess AppOptions{..} = do
+  absFile <- liftIO $ makeAbsolute file
+  let fileDir = takeDirectory absFile
+      fileName = takeFileName absFile
+      outputFileName = replaceExtension fileName ".i"
+      output = fileDir </> outputFileName
+      processConfig = proc "gcc" ["-E", "-P", absFile, "-o", output]
+  
+  (exitCode, _, stderr) <- liftIO $ readProcess processConfig
+  if exitCode == ExitSuccess
+    then return output
+    else throwError $ SystemError $ "GCC preprocessing failed: " <> T.pack (show stderr)
+
+-- | Compile assembly to executable using GCC
+compileToExecutable :: MonadCompiler m => AppOptions -> Text -> m ()
+compileToExecutable AppOptions{..} assembly = do
+  inputFileAbs <- liftIO $ makeAbsolute file
+  let inputDir = takeDirectory inputFileAbs
+      baseName = dropExtension $ takeFileName inputFileAbs
+      asmFilePath = inputDir </> (baseName ++ ".s")
+      outputFilePath = inputDir </> baseName
+
+  liftIO $ TIO.writeFile asmFilePath assembly
+  
+  let processConfig = proc "gcc" [asmFilePath, "-o", outputFilePath]
+  (exitCode, _, stderr) <- liftIO $ readProcess processConfig
+  
+  unless (exitCode == ExitSuccess) $
+    throwError $ SystemError $ "GCC compilation failed: " <> T.pack (show stderr)

lib/Halcyon/Driver/Output.hs

@@ -0,0 +1,28 @@
+{-# LANGUAGE OverloadedStrings #-}
+module Halcyon.Driver.Output
+  ( handleStageResult
+  ) where
+
+import Control.Monad.IO.Class (liftIO)
+import qualified Data.Text.IO as TIO
+
+import Halcyon.Core.Monad (MonadCompiler)
+import Halcyon.Core.Settings (StageResult(..))
+import Halcyon.Driver.Cli (AppOptions(..))
+import Halcyon.Driver.External (compileToExecutable)
+
+-- | Handle the result of a compilation stage
+handleStageResult :: MonadCompiler m => AppOptions -> StageResult -> m ()
+handleStageResult opts@AppOptions{} = \case
+  StageResultTokens tokens -> 
+    liftIO $ print tokens
+  StageResultAST ast ->
+    liftIO $ print ast
+  StageResultTacky tacky ->
+    liftIO $ print tacky
+  StageResultAsm asm ->
+    liftIO $ print asm
+  StageResultAssembly assembly ->
+    liftIO $ TIO.putStrLn assembly
+  StageResultExecutable assembly -> 
+    compileToExecutable opts assembly