使用Bazel 6.1.1编译Python应用时遇Workspace循环依赖错误

1. 未提前引入依赖仓库：WORKSPACE文件开头直接加载@io_bazel_rules_docker的规则文件，但该仓库并未通过http_archive等命令被Bazel拉取和定义，导致Bazel无法找到该仓库，触发循环依赖检测错误。
2. 资源定义顺序错误：container_pull定义@python_container的语句放在了使用它的container_image之后，Bazel解析时会因找不到预定义的仓库而报错。

1. 修正WORKSPACE文件 #

在文件最顶部添加http_archive引入io_bazel_rules_docker仓库，并调整语句顺序，确保先定义资源再使用：

# 引入与Bazel 6.1.1兼容的io_bazel_rules_docker版本
http_archive(
    name = "io_bazel_rules_docker",
    sha256 = "a9f7a68a693d56b3e8e5b32b4a88b04e35e5f8c4e16b2438a3d7a418c3b9f9d0",
    strip_prefix = "rules_docker-0.29.0",
    urls = ["https://github.com/bazelbuild/rules_docker/releases/download/v0.29.0/rules_docker-v0.29.0.tar.gz"],
)

# 加载并初始化容器相关仓库
load(
    "@io_bazel_rules_docker//repositories:repositories.bzl",
    container_repositories = "repositories",
)
container_repositories()

# 加载并初始化Python镜像相关仓库
load(
    "@io_bazel_rules_docker//python:image.bzl",
    _py_image_repos = "repositories",
)
_py_image_repos()

# 先定义python_container镜像，再在后续container_image中使用
container_pull(
    name = "python_container",
    registry = "docker.io/library",
    repository = "python",
    tag = "3.9-slim",
)

# 加载容器相关规则
load("@io_bazel_rules_docker//container:container.bzl", "container_layer", "container_image")

container_layer(
    name = "python_symlink",
    symlinks = {
      "/usr/bin/python": "/usr/local/bin/python",
      "/usr/bin/python3": "/usr/local/bin/python",
    },
)

container_image(
    name = "python",
    base = "@python_container//image",
    layers = [":python_symlink"],
    visibility = ["//visibility:public"],
)

2. 优化BUILD文件（可选） #

如果不需要自定义的@python//image镜像，可直接使用@python_container//image作为基础镜像，简化BUILD内容：

load("@io_bazel_rules_docker//python3:image.bzl", "py3_image")

py3_image(
    name = "main",
    main = "main.py",
    srcs = ["main.py"],
    base = "@python_container//image",
    deps = [],
)

3. 重新执行编译 #

运行以下命令重新编译项目：

bazel build //...

内容的提问来源于stack exchange，提问作者DmitrySemenov