【问题标题】:Calling pipenv in CMake command ModuleNotFoundError在 CMake 命令 ModuleNotFoundError 中调用 pipenv
【发布时间】:2023-03-31 00:19:01
【问题描述】:

我正在尝试使用 CMake 通过 Python 脚本生成 C/C++ 源文件,但遇到了一个问题,pipenv 似乎没有按预期工作。

我试图创建一个简化版本的我的真实世界示例on my Github

cmake_minimum_required(VERSION 2.8.5)
project(example)

include_directories(${PROJECT_SOURCE_DIR}/src)

find_package(Python3 COMPONENTS Interpreter REQUIRED)
find_program(PipEnv pipenv REQUIRED)

# Create command to compile the generate command
add_custom_command(
    OUTPUT
        ${CMAKE_SOURCE_DIR}/generate.py
    COMMAND ${PipEnv} install
    COMMENT "Installing pipenv dependencies [${PipEnv} install]"
    DEPENDS
        ${CMAKE_SOURCE_DIR}/generate.py
        ${CMAKE_SOURCE_DIR}/Pipfile
)

add_custom_target(
    generate_version
    COMMAND ${PipEnv} run ${Python3_EXECUTABLE} ${CMAKE_SOURCE_DIR}/generate.py -o src/version.h config.json
    COMMENT "Generating version header [${PipEnv} run ${CMAKE_SOURCE_DIR}/generate.py -o src/version.h config.json]"
    DEPENDS
        ${CMAKE_SOURCE_DIR}/generate.py
        ${CMAKE_SOURCE_DIR}/config.json
)

add_subdirectory(src)

我收到的错误是:

[main] Building folder: Python_example 
[build] Starting build
[proc] Executing command: "C:\Program Files\CMake\bin\cmake.EXE" --build c:/src/cmake-auto-increment-build-number/Python_example/build --config Debug --target ALL_BUILD -- /maxcpucount:10
[build] CMake is re-running because C:/src/cmake-auto-increment-build-number/Python_example/build/CMakeFiles/generate.stamp is out-of-date.
[build]   the file 'C:/src/cmake-auto-increment-build-number/Python_example/CMakeLists.txt'
[build]   is newer than 'C:/src/cmake-auto-increment-build-number/Python_example/build/CMakeFiles/generate.stamp.depend'
[build]   result='-1'
[build] -- Selecting Windows SDK version 10.0.18362.0 to target Windows 10.0.18363.
[build] -- Configuring done
[build] -- Generating done
[build] -- Build files have been written to: C:/src/cmake-auto-increment-build-number/Python_example/build
[build] Microsoft (R) Build Engine version 16.3.2+e481bbf88 for .NET Framework
[build] Copyright (C) Microsoft Corporation. All rights reserved.
[build] 
[build]   Checking Build System
[build]   Generating version header [C:/Program Files/Python38/Scripts/pipenv.exe run C:/src/cmake-auto-increment-build-number/Python_example/generate.py -o src/version.h config.json]
[build]   Traceback (most recent call last):
[build]     File "C:/src/cmake-auto-increment-build-number/Python_example/generate.py", line 14, in <module>
[build]       from docopt import docopt
[build]   ModuleNotFoundError: No module named 'docopt'
[build] C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\Microsoft.CppCommon.targets(230,5): error MSB6006: "cmd.exe" exited with code 1. [C:\src\cmake-auto-increment-build-number\Python_example\build\generate_version.vcxproj]
[build] Build finished with exit code 1

我尝试将 add_custom_target()COMMAND arg 更改为 ${PipEnv} install,但这没有任何效果。我还需要将其他内容传递给 pipenv 或 CMake 命令以使其正常工作吗?

【问题讨论】:

  • 你的 repo 已经包含了 generate.py 所以你在 add_custom_command() 中的pyenv install 不会被运行(输出已经在这里)
  • 另外你真的应该在二进制目录中创建你的 pyenv 恕我直言
  • 如果我将add_custom_commandOUTPUT 更改为${CMAKE_SOURCE_DIR}/src/version.hadd_custom_targetDEPENDS 以包含${CMAKE_SOURCE_DIR}/src/version.h 我可以让自定义命令运行,但是我仍然会遇到 ModuleNotFoundError ,就好像 pipenv 甚至没有被使用一样。

标签: python cmake pipenv


【解决方案1】:

这里有几个错误...

  1. 您的自定义命令输出和依赖于同一个文件,generate.py
    注意:您可以在跟踪中看到一个 make 错误

  2. 默认情况下,add_custom_command 将当前源目录作为工作直接,而自定义目标在当前构建目录中运行。
    即两者都在不同的目录中运行 -> 使用了两个不同的 pipenv...

  3. 你应该使用pipenv run python 而不是pipenv run ${Python3_EXECUTABLE}...

  4. 在这里我的 Dockerfile 运行一些测试

# Create a virtual environment with all tools installed
# ref: https://hub.docker.com/_/ubuntu
FROM ubuntu:rolling AS base
LABEL maintainer="mizux.dev@gmail.com"
# Install system build dependencies
ENV PATH=/usr/local/bin:$PATH
RUN apt-get update -qq \
&& apt-get install -yq git wget build-essential \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

# Install CMake 3.18.1
RUN wget "https://cmake.org/files/v3.18/cmake-3.18.1-Linux-x86_64.sh" \
&& chmod a+x cmake-3.18.1-Linux-x86_64.sh \
&& ./cmake-3.18.1-Linux-x86_64.sh --prefix=/usr/local/ --skip-license \
&& rm cmake-3.18.1-Linux-x86_64.sh
CMD [ "/usr/bin/bash" ]

# Install Python 3.8
RUN apt-get update -qq \
&& apt-get install -yq python3-dev python3-pip \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# Install pipenv
RUN python3 -m pip install pipenv

FROM base AS devel
WORKDIR /home/sample
COPY . .

FROM devel AS build
RUN cmake -S. -Bbuild
RUN cmake --build build --target all -v

构建:

docker build --target=build --tag so .

播放(即停止一个开发阶段然后生成一个容器来播放):

docker build --target=devel --tag so .
docker run --rm -it so

所以你修改了 CMakeLists.txt

cmake_minimum_required(VERSION 2.8.5)
project(example)

include_directories(${PROJECT_SOURCE_DIR}/src)

find_package(Python3 COMPONENTS Interpreter REQUIRED)
find_program(PipEnv pipenv REQUIRED)

# Create command to compile the generate command
add_custom_command(
    OUTPUT Pipfile.lock
    COMMAND pwd
    COMMAND ${PipEnv} install
    COMMENT "Installing pipenv dependencies [${PipEnv} install]"
    DEPENDS
        Pipfile
    WORKING_DIRECTORY
        ${CMAKE_CURRENT_SOURCE_DIR}
)

add_custom_target(
    generate_version
    COMMAND pwd
    COMMAND ${PipEnv} run python generate.py -o src/version.h config.json
    COMMENT "Generating version header [${PipEnv} run ${CMAKE_SOURCE_DIR}/generate.py -o src/version.h config.json]"
    DEPENDS
        Pipfile.lock
        generate.py
        config.json
    WORKING_DIRECTORY
        ${CMAKE_CURRENT_SOURCE_DIR}
)

add_subdirectory(src)

【讨论】:

  • 谢谢!这样可行!我没有意识到不同的工作目录。顺便说一句,您不需要(也不应该)在您的add_custom_command 中包含COMMAND ${PipEnv} run ${Python3_EXECUTABLE} -m pip install docopt,因为这是使用 pipenv 安装依赖项的全部意义所在。
  • 其实这里不工作^^;,看来你需要使用pipenv run python而不是pipenv run ${Python3_EXECUTABLE}
  • 但这不是在你的系统PATH上使用python(可以是python 2),而使用${Python3_EXECUTABLE}将使用CMake找到并保证的任何python版本是python版本3+?
  • 坦率地说,我不知道,但由于您的 Pipfile 明确包含 python 3.8...,BTW pipenv install 也有一个选项--python 但同样因为我不使用它,文档是稀疏所以我不会进一步挖掘,顺便说一句,venv/pyenv 在很大程度上足以满足我的用例;)另外我不知道 pipenv 的位置以及它的站点包......
猜你喜欢
  • 1970-01-01
  • 1970-01-01
  • 1970-01-01
  • 1970-01-01
  • 2018-03-05
  • 1970-01-01
  • 1970-01-01
  • 2017-11-14
  • 2023-01-26
相关资源
最近更新 更多