• Tiger Compiler Labs in C++
  • Tips
  • Contents
  • Overview
  • Difference Between C Labs and C++ Labs
  • Installing Dependencies
  • Compiling and Debugging
  • Testing Your Labs
  • Submitting Your Labs
  • Formatting Your Codes
  • Other Commands
  • Contributing to Tiger Compiler
  • External Documentations

1. 使用docker

    sudo make docker-run
    cd tiger-compiler
    make clean
    make gradelabX

2. 运行单个测试用例

• lab2

  cd build
  ./test_lex /home/stu/tiger-compiler/testdata/lab2/testcases/merge.tig

• lab3

  cd build
  ./test_parse /home/stu/tiger-compiler/testdata/lab3/testcases/merge.tig

• lab4

  cd build
  ./test_semant /home/stu/tiger-compiler/testdata/lab4/testcases/test10.tig

• lab5-part1

  cd build
  ./test_translate /home/stu/tiger-compiler/testdata/lab5or6/testcases/bsearch.tig

• lab5

    cd build
   ./test_codegen /home/stu/tiger-compiler/testdata/lab5or6/testcases/bsearch.tig
   cd ../scripts/lab5_test
   python3 main.py -d ../../testdata/lab5or6/testcases/tfo.tig.s
   python3 main.py ../../testdata/lab5or6/testcases/tfo.tig.s
   cd ../../build
   p label_->Name()

• lab6

   ./tiger-compiler /home/stu/tiger-compiler/testdata/lab5or6/testcases/tfo.tig
   gcc -Wl,--wrap,getchar -m64 /home/stu/tiger-compiler/testdata/lab5or6/testcases/tfo.tig.s /home/stu/tiger-compiler/src/tiger/runtime/runtime.c -o test.out
   ./test.out
   objdump -d test.out > test-fact.asm

• lab7

  cmake -DCMAKE_BUILD_TYPE=DEBUG ..
  ./tiger-compiler /home/stu/tiger-compiler/testdata/lab7/testcases/simple_array.tig
  g++ -g -Wl,--wrap,getchar -m64 /home/stu/tiger-compiler/src/tiger/runtime/runtime.cc /home/stu/tiger-compiler/testdata/lab7/testcases/simple_array.tig.s /home/stu/tiger-compiler/src/tiger/runtime/gc/heap/derived_heap.cc -o test.out

We rewrote the Tiger Compiler labs using the C++ programming language because some features in C++ like inheritance and polymorphism are more suitable for these labs and less error-prone.

We provide you all the codes of all labs at one time. In each lab, you only need to code in some of the directories.

1. Tiger compiler in C++ uses flexc++ and bisonc++ instead of flex and bison because flexc++ and bisonc++ is more flexc++ and bisonc++ are able to generate pure C++ codes instead of C codes wrapped in C++ files.

2. Tiger compiler in C++ uses namespace for modularization and uses inheritance and polymorphism to replace unions used in the old labs.

3. Tiger compiler in C++ uses CMake instead of Makefile to compile and build the target.

We provide you a Docker image that has already installed all the dependencies. You can compile your codes directly in this Docker image.

1. Install Docker.

2. Run a docker container and mount the lab directory on it.

# Run this command in the root directory of the project
docker run -it --privileged -p 2222:22 -v $(pwd):/home/stu/tiger-compiler ipadsse302/tigerlabs_env:latest  # or make docker-run

There are five makeable targets in total, including test_slp, test_lex, test_parse, test_semant, and tiger-compiler.

1. Run container environment and attach to it

# Run container and directly attach to it
docker run -it --privileged -p 2222:22 \
    -v $(pwd):/home/stu/tiger-compiler ipadsse302/tigerlabs_env:latest  # or `make docker-run`
# Or run container in the backend and attach to it later
docker run -dt --privileged -p 2222:22 \
    -v $(pwd):/home/stu/tiger-compiler ipadsse302/tigerlabs_env:latest
docker attach ${YOUR_CONTAINER_ID}

2. Build in the container environment

mkdir build && cd build && cmake .. && make test_xxx  # or `make build`

3. Debug using gdb or any IDEs

gdb test_xxx # e.g. `gdb test_slp`

Note: we will use -DCMAKE_BUILD_TYPE=Release to grade your labs, so make sure your lab passed the released version

Use make

make gradelabx

You can test all the labs by

make gradeall

Push your code to your GitLab repo

git add somefiles
git commit -m "A message"
git push

Note, each experiment has a separate branch, such as lab1. When you finish the lab1, you must submit the code to the lab1 branch. Otherwise, you won't get a full score in your lab.

We provide an LLVM-style .clang-format file in the project directory. You can use it to format your code.

Use clang-format command

find . \( -name "*.h" -o -iname "*.cc" \) | xargs clang-format -i -style=file  # or make format

or config the clang-format file in your IDE and use the built-in format feature in it.

Utility commands can be found in the Makefile. They can be directly run by make xxx in a Unix shell. Windows users cannot use the make command, but the contents of Makefile can still be used as a reference for the available commands.

You can post questions, issues, feedback, or even MR proposals through our main GitLab repository. We are rapidly refactoring the original C tiger compiler implementation into modern C++ style, so any suggestion to make this lab better is welcomed.

You can read external documentations on our course website:

• Lab Assignments
• Environment Configuration of Tiger Compiler Labs