在WAF中指定不同的编译器

问题描述

我想在waf中配置几个构建配置(意味着不同的C编译器).我设法那样做，但是对我来说似乎有点臭.

I want to have several build configurations (meaning different C compilers) configured in waf. I managed to do it that way, but it looks a bit smelly to me.

当前如何操作:我制作了不同的环境，并在加载下一个编译器之前重置c_compiler列表，最后将其重置为所有编译器.

How do I do it currently: I make different envs, and reset the c_compiler list before loading the next compiler, and at the end I reset it to all compilers.

def configure(cnf):
   _os = Utils.unversioned_sys_platform()

   restore_c_compilers = c_compiler[_os]

   # load gcc
   c_compiler[_os] = ['gcc']
   conf.setenv('gcc')
   conf.load('compiler_c')

   # load clang
   conf.setenv('clang')
   c_compiler[_os] = ['clang']
   conf.load('compiler_c')

   c_compiler[_os] = restore_c_compilers  # reset compiler list

有没有一种更好的方法?

Is there a better way to do it?

SO上存在此问题(如何将多个编译器与waf(Python))，但没有合适的答案.

There is this question on SO (How to use multiple compilers with waf (Python)) but with no suitable answer.

推荐答案

好吧，在这种情况下使用waf的方法是变体"(参见waf书第7.2.2节).由于gcc通常是默认的编译器，因此我为每个其他编译器创建了一个变体，以及一组相应的命令和环境.实践中:

Well, the way to go with waf in this case is "variants" (Cf. the waf book §7.2.2). As gcc is usually the default compiler, I create a variant for each other compiler, and a corresponding set of commands and environments. In practise:

def options(opt):
    opt.load('compiler_c')

def configure(conf):

    # here we are in default variant/env
    # we load the default compiler, probably gcc

    conf.load('compiler_c')

    # config for clang variant

    conf.setenv('clang')
    conf.env.CC = ['clang']
    conf.load('compiler_c')

    # config for icc variant

    conf.setenv('icc')
    conf.env.CC = ['icc']
    conf.load('compiler_c')

    # back to default config

    conf.setenv('')

def build(bld):

    bld.program(source = 'main.c', target = 'myexe')

# this create variants commands and build directories

from waflib.Build import BuildContext, \
    CleanContext, InstallContext, UninstallContext

for variant in ['clang', 'icc']:
    for context in [BuildContext, CleanContext, InstallContext, UninstallContext]:
        name = context.__name__.replace('Context','').lower()
        class tmp(context):
            cmd = name + '_' + variant
            variant = variant

获得的内容:额外的命令build_icc，build_clang，clean_icc，clean_clang，...每个变量的目录，即build/icc和build/clang，当然，您的exe将使用相应的编译器.

What you get: extra commands build_icc, build_clang, clean_icc, clean_clang, ... A directory for each variant, namely build/icc and build/clang and of course your exe build with the corresponding compiler.

要测试:

waf build -v        # use gcc, or the default compiler
waf build_icc -v    # use icc
waf build_clang -v  # use clang

您将获得:

build/
├── c4che
│   ├── build.config.py
│   ├── _cache.py
│   ├── clang_cache.py
│   └── icc_cache.py
├── clang
│   ├── main.c.1.o
│   └── myexe
├── icc
│   ├── main.c.1.o
│   └── myexe
├── config.log
├── main.c.1.o
└── myexe

请注意，默认变体位于根构建目录中.其缓存文件为c4che/_cache.py.每个变体都有一个目录和一个名为ater的缓存.

Notice the default variant is in the root build directory. Its cache file is c4che/_cache.py. Each variant has a directory and a cache named ater its name.

这篇关于在WAF中指定不同的编译器的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！