舊博文，搬到 csdn
原文：http://rebootcat.com/2020/08/30/scons/

前言

我是一個 linux c++ 開發者，但是一直對 Makefile 的語法很是頭痛，每次都記不住，所以每次寫 Makefile 都很痛苦，Makefile 里需要你自己編寫依賴和推導規則，這個過程能不能簡單點呢？

對于編譯一個 C++ 工程來說，也許需要的就是頭文件路徑、庫路徑、編譯參數，剩下的東西基本也不重要，這三樣足夠去編譯一個工程了。所以有沒有一個工具能簡單點的去實現 C++ 項目的構建呢？

答案是有的，Scons 就是答案。

Scons

什么是 scons

這里直接引用官網的解釋：

What is SCons?

SCons is an Open Source software construction tool—that is, a next-generation build tool. Think of SCons as an improved, cross-platform substitute for the classic Make utility with integrated functionality similar to autoconf/automake and compiler caches such as ccache. In short, SCons is an easier, more reliable and faster way to build software.

What makes SCons better?

• Configuration files are Python scripts–use the power of a real programming language to solve build problems.
• Reliable, automatic dependency analysis built-in for C, C++ and Fortran–no more “make depend” or “make clean” to get all of the dependencies. Dependency analysis is easily extensible through user-defined dependency Scanners for other languages or file types.
• Built-in support for C, C++, D, Java, Fortran, Yacc, Lex, Qt and SWIG, and building TeX and LaTeX documents. Easily extensible through user-defined Builders for other languages or file types.
• Building from central repositories of source code and/or pre-built targets.
• Built-in support for fetching source files from SCCS, RCS, CVS, BitKeeper and Perforce.
• Built-in support for Microsoft Visual Studio .NET and past Visual Studio versions, including generation of .dsp, .dsw, .sln and .vcproj files.
• Reliable detection of build changes using MD5 signatures; optional, configurable support for traditional timestamps.
• Improved support for parallel builds–like make -j but keeps N jobs running simultaneously regardless of directory hierarchy.
• Integrated Autoconf-like support for finding #include files, libraries, functions and typedefs.
• Global view of all dependencies–no more multiple build passes or reordering targets to build everything.
• Ability to share built files in a cache to speed up multiple builds–like ccache but for any type of target file, not just C/C++ compilation.
• Designed from the ground up for cross-platform builds, and known to work on Linux, other POSIX systems (including AIX, BSD systems, HP/UX, IRIX and Solaris), Windows NT, Mac OS X, and OS/2.

最大特點就是使用 Python 語法來編寫編譯構建腳本，并且支持依賴自動推導，支持編譯 C/C++/D/Java/Fortran等項目，并且是跨平臺的(因為 python 是跨平臺的）。

所以如果你對 python 熟悉的話，而且你和我對 C++ Makefile 有一樣的煩惱，那么這對你將是一個好消息。 你將可以用 python 來編寫構建腳本，而且會很簡單，對于復雜的大型項目也能快速構建好。（也許只要 30 分鐘）

安裝 scons

因為 scons 是基于 python 來構建的，所以毋容置疑，首先是需要準備好 python 環境，然后使用下述命令安裝 scons 工具。

pip install scons

scons 使用語法

注：本文以一個多源文件，多目錄結構的項目 mux 為例，介紹 cmake 的使用，相關源文件以及cmake 腳本可以直接查看源項目。

scons 構建腳本由一個 SConstruct 文件和多個 SConscript 文件構成。

SConstruct 通常位于項目頂層目錄，然后 SConscript 通常位于子目錄（子模塊）。

那么來看一下 SConstruct 腳本長啥樣?

SConstruct

#!/usr/bin/env python
#-*- coding:utf-8 -*-import sys
import os
import platform
import reenv = Environment()
abs_path = os.getcwd()
print('workspace path:{0}'.format(abs_path))sbuild_dir = 'sbuild'headers = ['.', 'third-party/include']
libs = ['./third-party/lib']abs_headers = []
abs_libs = []for item in headers:abs_item = os.path.join(abs_path, item)abs_headers.append(abs_item)for item in libs:abs_item = os.path.join(abs_path, item)abs_libs.append(abs_item)build_dir = os.path.join(abs_path, sbuild_dir)
abs_libs.append(os.path.join(build_dir, 'lib'))CCFLAGS = '-ggdb -std=c++11'print('\nheaders path:')
print(abs_headers)
print('\n')print('libs path:')
print(abs_libs)
print('\n')print("begin load SConscript")env["headers"] = abs_headers
env["libs"]    = abs_libs
env["MUX_DIR"] = abs_path
env['ccflags'] = CCFLAGS
env['build_dir'] = build_dirExport('env')SConscript(['./mbase/SConscript'])
SConscript(['./message_handle/SConscript'])
SConscript(['./epoll/SConscript'])
SConscript(['./transport/SConscript'])
SConscript(['./demo/bench/SConscript'])
SConscript(['./demo/echo/SConscript'])print("\n All Done, Please Check {0}".format(env['build_dir']))

來分析一下這個文件，源文件可以直接在 我的github下載。

SConstruct 文件主要做了兩件事：

• env 環境變量的構造，主要是頭文件路徑，庫路徑，編譯參數，自定義的一些變量等
• 使用 SConscript 函數解析執行子模塊的 SConscript 文件

需要注意的是 SConstruct 和 SConscript 共享變量使用的就是 env 這個變量，你可以看到上面有一句：

Export('env')

這句很重要。

SConscript

那么位于子模塊或者子目錄的 SConscript 文件長啥樣呢？

#!/usr/bin/env python
#-*- coding:utf-8 -*-import os
import sysImport('env')
project_dir  = env['MUX_DIR']epoll_lib  = 'epoll'epoll_src_path = os.path.join(project_dir, 'epoll/src')
epoll_sources = []
for item in os.listdir(epoll_src_path):if item.endswith('.cc') or item.endswith('.cpp') or item.endswith('.cxx'):abs_item = os.path.join(epoll_src_path, item)epoll_sources.append(abs_item)print('\nbuild target:lib{0}.a'.format(epoll_lib))
print(epoll_sources)lib_dir = os.path.join(env['build_dir'], 'lib')link_libraries = ['mbase']
for lib_name in link_libraries:lib_name = "{0}{1}{2}".format(env['LIBPREFIX'], lib_name, env['LIBSUFFIX'])abs_lib_name = os.path.join(lib_dir, lib_name)epoll_sources.append(abs_lib_name)env.StaticLibrary(target = os.path.join(lib_dir, epoll_lib),source  = epoll_sources,CPPPATH = env['headers'], # includeLIBPATH = env['libs'],    # lib pathLIBS    = ['pthread'],    # link libCCFLAGS = env['ccflags'])

來分析一下這個文件，源文件可以直接在 我的github下載。

SConscript 主要做了兩件事：

• 構造一個源文件列表（用來構建 target 所需要使用的源文件）
• 根據需要構建 static_lib/dynamic_lib/binary

當然，還有一點很重要，上面其實提到了，SConscript 和 SConstruct 用來共享變量使用的是 env 這個變量，所以你可以看到一句很重要的：

Import('env')

構造源文件列表，對于 Python 來說，簡直是小菜一碟，太簡單了；

然后如何生成目標文件呢？

1 生成二進制文件

env.Program(target = os.path.join(bin_dir, echo_server_bin),source  = echo_server_sources,CPPPATH = env['headers'],LIBPATH = env['libs'],LIBS    = ['transport','msghandler','epoll', 'mbase', 'pthread'],CCFLAGS = env['ccflags'])

2 生成靜態庫

env.StaticLibrary(target = os.path.join(lib_dir, epoll_lib),source  = epoll_sources,CPPPATH = env['headers'], # includeLIBPATH = env['libs'],    # lib pathLIBS    = ['pthread'],    # link libCCFLAGS = env['ccflags'])

3 生成動態庫

env.SharedLibrary(target = os.path.join(lib_dir, epoll_lib),source  = epoll_sources,CPPPATH = env['headers'], # includeLIBPATH = env['libs'],    # lib pathLIBS    = ['pthread'],    # link libCCFLAGS = env['ccflags'])

上面 3 個函數的參數都是類似的：

• target: 指定需要生成的目標文件，通常我自己會寫一個絕對路徑；對于 lib 來說只需要寫名字就行，前綴和后綴不需要寫。(eg. target = ‘/root/scons_repo/sbuild/lib/test’ ，會生成 /root/scons_repo/sbuild/lib/libtest.a)
• source: 編譯目標文件需要的源文件列表
• CPPPATH: 通常就是需要 Include 的頭文件路徑
• LIBPATH: 通常就是需要鏈接的庫路徑
• LIBS: 需要鏈接的庫列表
• CCFLAGS: 編譯參數

attention:

上面有一個坑我自己碰到的，當我構建目標生成一個靜態庫的時候，需要鏈接其他的靜態庫，如果使用 $LIBPATH 和 $LIBS 指定鏈接庫的話，scons 并沒有鏈接這些庫。嘗試了很多方法，搜索了很多，也沒有解決這個問題。

最后是這樣解決的。把需要鏈接的靜態庫添加到 source 參數中，和其他 cc/cpp 源文件一樣放在一起，并且這些庫需要使用絕對路徑。

通常為了跨平臺的方便，需要考慮lib 的前后綴，可以這樣寫：

link_libraries = ['test1', 'test2']
for lib_name in link_libraries:lib_name = "{0}{1}{2}".format(env['LIBPREFIX'], lib_name, env['LIBSUFFIX'])abs_lib_name = os.path.join(lib_dir, lib_name)sources.append(abs_lib_name)

scons 命令

上面詳細講解了如何使用 python 編寫構建腳本，那么寫好之后怎么用呢？

常用的幾個命令：

編譯：

scons

如果需要并行編譯：

scons -j4

清理：

scons -c

然后就會按照你腳本里寫的方式去構建目標了。

這里貼一下 我的項目 編譯的輸出：

$ scons
scons: Reading SConscript files ...
workspace path:/mnt/centos-share/workspace/muxheaders path:
['/mnt/centos-share/workspace/mux/.', '/mnt/centos-share/workspace/mux/third-party/include']libs path:
['/mnt/centos-share/workspace/mux/./third-party/lib', '/mnt/centos-share/workspace/mux/sbuild/lib']begin load SConscriptbuild target:libmbase.a
['/mnt/centos-share/workspace/mux/mbase/src/packet.cc']build target:libmsghandler.a
['/mnt/centos-share/workspace/mux/message_handle/src/message_handler.cc']build target:libepoll.a
['/mnt/centos-share/workspace/mux/epoll/src/epoll_tcp_client.cc', '/mnt/centos-share/workspace/mux/epoll/src/epoll_tcp_server.cc']build target:libtransport.a
['/mnt/centos-share/workspace/mux/transport/src/tcp_transport.cc']build target:bench_server
['bench_server.cc']build target:bench_client
['client.cc']build target:echo_server
['echo_server.cc']build target:echo_client
['client.cc']All Done, Please Check /mnt/centos-share/workspace/mux/sbuild
scons: done reading SConscript files.
scons: Building targets ...
g++ -o demo/bench/bench_server.o -c -ggdb -std=c++11 -I. -Ithird-party/include demo/bench/bench_server.cc
g++ -o demo/bench/client.o -c -ggdb -std=c++11 -I. -Ithird-party/include demo/bench/client.cc
g++ -o demo/echo/client.o -c -ggdb -std=c++11 -I. -Ithird-party/include demo/echo/client.cc
g++ -o demo/echo/echo_server.o -c -ggdb -std=c++11 -I. -Ithird-party/include demo/echo/echo_server.cc
g++ -o epoll/src/epoll_tcp_client.o -c -ggdb -std=c++11 -I. -Ithird-party/include epoll/src/epoll_tcp_client.cc
g++ -o epoll/src/epoll_tcp_server.o -c -ggdb -std=c++11 -I. -Ithird-party/include epoll/src/epoll_tcp_server.cc
g++ -o mbase/src/packet.o -c -ggdb -std=c++11 -I. -Ithird-party/include mbase/src/packet.cc
g++ -o message_handle/src/message_handler.o -c -ggdb -std=c++11 -I. -Ithird-party/include message_handle/src/message_handler.cc
g++ -o transport/src/tcp_transport.o -c -ggdb -std=c++11 -I. -Ithird-party/include transport/src/tcp_transport.cc
ar rc sbuild/lib/libmbase.a mbase/src/packet.o
ranlib sbuild/lib/libmbase.a
ar rc sbuild/lib/libepoll.a epoll/src/epoll_tcp_client.o epoll/src/epoll_tcp_server.o sbuild/lib/libmbase.a
ranlib sbuild/lib/libepoll.a
ar rc sbuild/lib/libtransport.a transport/src/tcp_transport.o sbuild/lib/libepoll.a sbuild/lib/libmbase.a
ranlib sbuild/lib/libtransport.a
ar rc sbuild/lib/libmsghandler.a message_handle/src/message_handler.o sbuild/lib/libmbase.a
ranlib sbuild/lib/libmsghandler.a
g++ -o sbuild/bin/bench_client demo/bench/client.o -Lthird-party/lib -Lsbuild/lib -ltransport -lmsghandler -lepoll -lmbase -lpthread
g++ -o sbuild/bin/bench_server demo/bench/bench_server.o -Lthird-party/lib -Lsbuild/lib -ltransport -lmsghandler -lepoll -lmbase -lpthread
g++ -o sbuild/bin/echo_client demo/echo/client.o -Lthird-party/lib -Lsbuild/lib -ltransport -lmsghandler -lepoll -lmbase -lpthread
g++ -o sbuild/bin/echo_server demo/echo/echo_server.o -Lthird-party/lib -Lsbuild/lib -ltransport -lmsghandler -lepoll -lmbase -lpthread
scons: done building targets.

$ scons -c
scons: Reading SConscript files ...
workspace path:/mnt/centos-share/workspace/muxheaders path:
['/mnt/centos-share/workspace/mux/.', '/mnt/centos-share/workspace/mux/third-party/include']libs path:
['/mnt/centos-share/workspace/mux/./third-party/lib', '/mnt/centos-share/workspace/mux/sbuild/lib']begin load SConscriptbuild target:libmbase.a
['/mnt/centos-share/workspace/mux/mbase/src/packet.cc']build target:libmsghandler.a
['/mnt/centos-share/workspace/mux/message_handle/src/message_handler.cc']build target:libepoll.a
['/mnt/centos-share/workspace/mux/epoll/src/epoll_tcp_client.cc', '/mnt/centos-share/workspace/mux/epoll/src/epoll_tcp_server.cc']build target:libtransport.a
['/mnt/centos-share/workspace/mux/transport/src/tcp_transport.cc']build target:bench_server
['bench_server.cc']build target:bench_client
['client.cc']build target:echo_server
['echo_server.cc']build target:echo_client
['client.cc']All Done, Please Check /mnt/centos-share/workspace/mux/sbuild
scons: done reading SConscript files.
scons: Cleaning targets ...
Removed demo/bench/bench_server.o
Removed demo/bench/client.o
Removed demo/echo/client.o
Removed demo/echo/echo_server.o
Removed epoll/src/epoll_tcp_client.o
Removed epoll/src/epoll_tcp_server.o
Removed mbase/src/packet.o
Removed message_handle/src/message_handler.o
Removed transport/src/tcp_transport.o
Removed sbuild/lib/libmbase.a
Removed sbuild/lib/libepoll.a
Removed sbuild/lib/libtransport.a
Removed sbuild/lib/libmsghandler.a
Removed sbuild/bin/bench_client
Removed sbuild/bin/bench_server
Removed sbuild/bin/echo_client
Removed sbuild/bin/echo_server
scons: done cleaning targets.

寫在最后

scons 使用 python 腳本來構建項目，如果對 python 熟悉的話，那么編寫編譯構建腳本將會大大提高效率，再也不用局限在 Makefile 的蛋疼語法里面了。

當然 scons 的缺點也有，據說在大型項目的時候，可能會很慢。這個我還沒碰到過，因為沒有用到大型項目中。

下一篇，分享下 cmake 構建 C++ 項目的一些語法和步驟。

cmake教程|cmake入門實戰

另外，文中涉及到的項目可以在我的github 找到。

Blog:

• rebootcat.com

• email: linuxcode2niki@gmail.com

2020-08-30 于杭州
By 史矛革

總結

以上是生活随笔為你收集整理的Scons构建C++项目的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。