Bazel:你如何获得生成文件的路径？

Question

在 Bazel 中，给定构建目标，脚本(在 Bazel 之外运行)如何获取生成文件的路径？

场景:我使用 Bazel 进行构建，完成后，我想将结果复制到服务器。我只需要知道要复制哪些文件。我可以对文件列表进行硬编码，但我不想这样做。

一个简单的例子:这个 Bazel 脚本:

genrule(
    name = "main",
    srcs = ["main.in"],
    outs = ["main.out"],
    cmd = "cp $< $@",
)

如果你创建一个名为 main.in 的文件然后运行 ​​ bazel build :main ，巴泽尔报告:

INFO: Found 1 target...
Target //:main up-to-date:
  bazel-genfiles/main.out
INFO: Elapsed time: 6.427s, Critical Path: 0.40s

所以有: bazel-genfiles/main.out .但是我可以使用什么机器可读技术来获得该路径？ (我可以解析 bazel build 的输出，但我们不鼓励这样做。)

我发现的最接近的是使用 bazel query --output=xml :main , 转储有关 :main 的信息以 XML 格式。输出包括这一行:

<rule-output name="//:main.out"/>

这与我想要的非常接近。但是 name是 Bazel 的标签格式；我不知道如何将其作为路径。

我可以在 name 上做某种字符串替换字段，把它变成 bazel-genfiles/main.out ;但即使这样也不可靠。如果我的 genrule已包含 output_to_bindir = 1 ，那么输出将是 bazel-bin/main.out反而。

此外，并非所有规则都有 <rule-output> XML 输出中的字段。例如，如果我的 BUILD文件有这个代码来制作一个 C 库:

cc_library(
    name = "mylib",
    srcs = glob(["*.c"])
)

bazel query --output=xml :mylib的输出不包含 <rule-output>或其他任何有用的东西:

<?xml version="1.1" encoding="UTF-8" standalone="no"?>
<query version="2">
  <rule class="cc_library" location="/Users/mikemorearty/src/bazel/test1/BUILD:8:1" name="//:mylib">
    <string name="name" value="mylib"/>
    <list name="srcs">
      <label value="//:foo.c"/>
    </list>
    <rule-input name="//:foo.c"/>
    <rule-input name="//tools/defaults:crosstool"/>
    <rule-input name="@bazel_tools//tools/cpp:stl"/>
  </rule>
</query>

Answer 1

您可以使用 bazel aquery 获取此信息。查询
Action 图。

这是一个稍微丰富的示例，其中包含来自单个的两个输出文件
规则:

$ ls
BUILD  main.in  WORKSPACE
$ cat WORKSPACE
$ cat BUILD
genrule(
    name = "main",
    srcs = ["main.in"],
    outs = ["main.o1", "main.o2"],
    cmd = "cp $< $(location main.o1); cp $< $(location main.o2)",
)
$ cat main.in
hello

使用 bazel aquery //:main --output=textproto用机器可读的输出查询 Action 图(原型(prototype)是 analysis.ActionGraphContainer ):

$ bazel aquery //:main --output=textproto >aquery_result 2>/dev/null
$ cat aquery_result
artifacts {
  id: "0"
  exec_path: "main.in"
}
artifacts {
  id: "1"
  exec_path: "external/bazel_tools/tools/genrule/genrule-setup.sh"
}
artifacts {
  id: "2"
  exec_path: "bazel-out/k8-fastbuild/genfiles/main.o1"
}
artifacts {
  id: "3"
  exec_path: "bazel-out/k8-fastbuild/genfiles/main.o2"
}
actions {
  target_id: "0"
  action_key: "dd7fd759bbecce118a399c6ce7b0c4aa"
  mnemonic: "Genrule"
  configuration_id: "0"
  arguments: "/bin/bash"
  arguments: "-c"
  arguments: "source external/bazel_tools/tools/genrule/genrule-setup.sh; cp main.in bazel-out/k8-fastbuild/genfiles/main.o1; cp main.in bazel-out/k8-fastbuild/genfiles/main.o2"
  input_dep_set_ids: "0"
  output_ids: "2"
  output_ids: "3"
}
targets {
  id: "0"
  label: "//:main"
  rule_class_id: "0"
}
dep_set_of_files {
  id: "0"
  direct_artifact_ids: "0"
  direct_artifact_ids: "1"
}
configuration {
  id: "0"
  mnemonic: "k8-fastbuild"
  platform_name: "k8"
}
rule_classes {
  id: "0"
  name: "genrule"
}

数据并不完全集中在一处，但请注意:

带有 ID 的工件 2和 3对应我们两个想要的
输出文件，并将这些工件的输出位置列为
磁盘上文件相对于工作区根目录的路径；

artifacts目标 ID 0 的条目与工件相关联
ID 2和 3 ;和

targets ID 为 "0" 的条目与 //:main 相关联
标签。

鉴于这个简单的结构，我们可以轻松地将脚本组合在一起
列出与提供的标签对应的所有输出文件。我找不到
直接依赖于 Bazel 对 analysis.proto 的定义的方式或其
来自外部存储库的语言绑定(bind)，因此您可以修补
以下脚本进入 bazelbuild/bazel存储库本身:

工具/list_outputs/list_outputs.py

# Copyright 2019 The Bazel Authors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
r"""Parse an `aquery` result to list outputs created for a target.

Use this binary in conjunction with `bazel aquery` to determine the
paths on disk to output files of a target.

Example usage: first, query the action graph for the target that you
want to analyze:

    bazel aquery //path/to:target --output=textproto >/tmp/aquery_result

Then, from the Bazel repository:

    bazel run //tools/list_outputs -- \
        --aquery_result /tmp/aquery_result \
        --label //path/to:target \
        ;

This will print a list of zero or more output files emitted by the given
target, like:

    bazel-out/k8-fastbuild/foo.genfile
    bazel-out/k8-fastbuild/bar.genfile

If the provided label does not appear in the output graph, an error will
be raised.
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import sys

from absl import app
from absl import flags
from google.protobuf import text_format
from src.main.protobuf import analysis_pb2


flags.DEFINE_string(
    "aquery_result",
    None,
    "Path to file containing result of `bazel aquery ... --output=textproto`.",
)
flags.DEFINE_string(
    "label",
    None,
    "Label whose outputs to print.",
)


def die(message):
  sys.stderr.write("fatal: %s\n" % (message,))
  sys.exit(1)


def main(unused_argv):
  if flags.FLAGS.aquery_result is None:
    raise app.UsageError("Missing `--aquery_result` argument.")
  if flags.FLAGS.label is None:
    raise app.UsageError("Missing `--label` argument.")

  if flags.FLAGS.aquery_result == "-":
    aquery_result = sys.stdin.read()
  else:
    with open(flags.FLAGS.aquery_result) as infile:
      aquery_result = infile.read()
  label = flags.FLAGS.label

  action_graph_container = analysis_pb2.ActionGraphContainer()
  text_format.Merge(aquery_result, action_graph_container)

  matching_targets = [
      t for t in action_graph_container.targets
      if t.label == label
  ]
  if len(matching_targets) != 1:
    die(
        "expected exactly one target with label %r; found: %s"
        % (label, sorted(t.label for t in matching_targets))
    )
  target = matching_targets[0]

  all_artifact_ids = frozenset(
      artifact_id
      for action in action_graph_container.actions
      if action.target_id == target.id
      for artifact_id in action.output_ids
  )
  for artifact in action_graph_container.artifacts:
    if artifact.id in all_artifact_ids:
      print(artifact.exec_path)


if __name__ == "__main__":
  app.run(main)

工具/列表输出/构建

# Copyright 2019 The Bazel Authors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

package(default_visibility = ["//visibility:public"])

licenses(["notice"])  # Apache 2.0

filegroup(
    name = "srcs",
    srcs = glob(["**"]),
)

py_binary(
    name = "list_outputs",
    srcs = ["list_outputs.py"],
    srcs_version = "PY2AND3",
    deps = [
        "//third_party/py/abseil",
        "//src/main/protobuf:analysis_py_proto",
    ],
)

作为一个 Git 补丁，为了您的方便:
https://gist.github.com/wchargin/5e6a43a203d6c95454aae2886c5b54e4

请注意，此代码尚未经过审查或验证
正确性；我提供它主要作为一个例子。如果有用的话
你，那么也许这个周末我可以为它写一些测试并公关它
反对巴泽尔本身。