Bazel自定义ccppc工具链报错：无法匹配//ccppc_tools:toolchain_type

问题场景 #

在Bazel中实现自定义工具链以使用ccppc编译器，遵循官方文档配置后，执行构建命令时Bazel无法识别//ccppc_tools:toolchain_type，报错找不到匹配的工具链。

执行命令 #

bazel build --platforms=//src:my_platform //src:TEST_OBJECT

报错信息 #

ERROR: C:/tools/dev/test_partition/src/BUILD:11:14: While resolving toolchains for target //src:TEST_OBJECT: No matching toolchains found for types //ccppc_tools:toolchain_type.
To debug, rerun with --toolchain_resolution_debug='//ccppc_tools:toolchain_type'
ERROR: Analysis of target '//src:TEST_OBJECT' failed; build aborted:
INFO: Elapsed time: 0.564s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (2 packages loaded, 3 targets configured)

项目结构 #

Project
  | 
  + ccppc_tools/ -- rules.bzl BUILD
  + src/ -- command.c BUILD
  + BUILD
  + WORKSPACE

相关配置文件 #

ccppc_tools/rules.bzl #

##################################################################### TOOLCHAIN PART

# Declare the information provider
CCPPCInfo = provider(
    doc = "Information about how to invoke the ccppc compiler.",
    # In the real world, compiler_path and system_lib might hold File objects,
    # but for simplicity they are strings for this example. arch_flags is a list
    # of strings.
    fields = ["compiler_path", "system_lib", "arch_flags"],
)

# Rules for the CCPPC
def _ccppc_toolchain_impl(ctx):
    toolchain_info = platform_common.ToolchainInfo(
        ccppcinfo = CCPPCInfo(
            compiler_path = ctx.attr.compiler_path,
            system_lib = ctx.attr.system_lib,
            arch_flags = ctx.attr.arch_flags,
        ),
    )
    return [toolchain_info]

ccppc_toolchain = rule(
    implementation = _ccppc_toolchain_impl,
    attrs = {
        "compiler_path": attr.string(),
        "system_lib": attr.label(
                    mandatory = False,
                    cfg = "target",
                ),
        "arch_flags": attr.string_list(),
    },
)


##################################################################### USER RULES PART

# Rule for creating objects with CCPPC
def _ccppc_objects_impl(ctx):
    info = ctx.toolchains["//ccppc_tools:toolchain_type"].ccppcinfo

    command = "%s %s -c %s" % (
        info.compiler_path,
        info.system_lib,
        " ".join(info.arch_flags),
    )
    print("CALLED CCPPC OBJECTS: ")
    print(command)

ccppc_objects = rule(
    implementation = _ccppc_objects_impl,
    attrs = {
        "srcs": attr.label_list(allow_files = True)
    },
    toolchains = ["//ccppc_tools:toolchain_type"],
)

ccppc_tools/BUILD #

load("//ccppc_tools:rules.bzl","ccppc_toolchain")

toolchain_type(name = "toolchain_type")

# Declare a ccppc toolchain
ccppc_toolchain(
    name = "ccppc_windows",
    arch_flags = [
        "--arch=WINDOWS",
        "--DTEST",
    ],
    compiler_path = "ccppc",
)

# Declare a toolchain from the ccppc_toolchain
toolchain(
    name = "ccppc_windows_toolchain",
    exec_compatible_with = [
        "@platforms//os:windows",
        "@platforms//cpu:x86_64",
    ],
    target_compatible_with = [
        "@platforms//os:linux",
        "@platforms//cpu:ppc",
    ],
    toolchain = ":ccppc_windows",
    toolchain_type = ":toolchain_type",
)

src/BUILD #

load("//:ccppc_tools/rules.bzl","ccppc_objects")

platform(
    name = "my_platform",
    constraint_values = [
        "@platforms//os:linux",
    ],
)

ccppc_objects(
    name = "TEST_OBJECT",
    srcs = ["src/command.c"]
)

WORKSPACE #

register_toolchains(
    "//ccppc_tools:ccppc_windows_toolchain",
)

解决方案 #

1. 修复平台约束不匹配问题 #

自定义平台//src:my_platform仅指定了操作系统约束，缺少工具链要求的CPU约束，导致Bazel无法匹配到工具链。修改src/BUILD中的platform定义，添加CPU约束：

platform(
    name = "my_platform",
    constraint_values = [
        "@platforms//os:linux",
        "@platforms//cpu:ppc",  # 新增该约束，与工具链target_compatible_with中的CPU匹配
    ],
)

2. 处理system_lib属性空值问题 #

当前ccppc_windows规则未赋值system_lib属性，在_ccppc_objects_impl中直接拼接会导致空字符串，可通过两种方式修复：

• 方式一：在ccppc_toolchain规则中给system_lib设置默认空值
  修改ccppc_tools/rules.bzl中ccppc_toolchain的attrs：

  "system_lib": attr.label(
      mandatory = False,
      cfg = "target",
      default = "",  # 添加默认值
  ),
  

• 方式二：在实现函数中判断空值
  修改_ccppc_objects_impl：

  def _ccppc_objects_impl(ctx):
      info = ctx.toolchains["//ccppc_tools:toolchain_type"].ccppcinfo
      system_lib_arg = info.system_lib if info.system_lib else ""
      command = "%s %s -c %s" % (
          info.compiler_path,
          system_lib_arg,
          " ".join(info.arch_flags),
      )
      print("CALLED CCPPC OBJECTS: ")
      print(command)
  

3. 验证工具链解析过程（可选） #

执行调试命令查看工具链匹配的详细日志，确认约束是否正确匹配：

bazel build --platforms=//src:my_platform //src:TEST_OBJECT --toolchain_resolution_debug='//ccppc_tools:toolchain_type'

内容的提问来源于stack exchange，提问作者mtribiere