Separate runtime / development container using DooD

[21kazu13]

• Separate into three containers:
  1. development
  2. python3 runtime
  3. pypy3 runtime
• Apply private repository clone after docker compose up
• Automatically run aclogin function after start container

[21kazu13]

test dood here: https://github.com/21kazu13/judgeenv/commit/e8aacfe4c62438f493b5d9b40ac44494da583304

Complete follows:

• basic dood (run another container from inside the container using host socket)
• share a volume between them
• consider and show example command for runner

[21kazu13]

Consider how to prepare to use

1. Clone this repository
2. 初回設定（docker build / runner側）
3. initializeCommand (ssh-agent起動確認 : 参考）
4. reopen in container
5. postcompose command (source clone / aclogin)

※gitは引き継がれる想定 https://zenn.dev/coloname33/articles/11590555881668

docker内部必要なコマンド定義

• oj test with runner
• oj test with error
• acc new
• acc s (acc s main.py -- -l 5055/50xx -y)
• oj-template (generate.py creation)
• oj generate-input
• oj generate-output

[21kazu13]

These can do it

• oj test with runner
• oj test with error

// memo: 下記で全体マウントして今のディレクトリでコマンド走らせられる $ docker run --rm -it --mount type=volume,src=atcoder,dst=/shared -w /shared/$(basename $(pwd)) ubuntu:22.10 bash // memo: ojはこれで走る（マウントは検証用のため適当） $ docker run --rm -t -v ./:/home/oj/run -w /home/oj/run runner_pypy oj t -c pypy\ $(ls .py) $ docker run --rm -t -v ./:/home/oj/run -w /home/oj/run runner_pypy oj t -c pypy\ $(ls .py) -e 1e-3 // -t: 色付き // pypy\ $(ls *.py): 良しなにfilename指定