cpython / PCbuild /

Filename Size Date modified Message
13.1 KB
8.3 KB
11.1 KB
10.6 KB
8.8 KB
9.3 KB
8.8 KB
10.7 KB
10.6 KB
8.4 KB
9.3 KB
5.8 KB
488 B
25 B
1.1 KB
253 B
9.4 KB
2.3 KB
11.9 KB
310 B
163 B
974 B
284 B
57.4 KB
5.5 KB
4.5 KB
3.1 KB
2.8 KB
6.8 KB
39.0 KB
921 B
379 B
606 B
833 B
10.0 KB
1.9 KB
11.7 KB
34.1 KB
11.3 KB
14.6 KB
311 B
598 B
1.8 KB
9.0 KB
8.9 KB
374 B
8.6 KB
1.3 KB
10.0 KB
8.6 KB
537 B
Building Python using VC++ 9.0

This directory is used to build Python for Win32 and x64 platforms, e.g. 
Windows 2000, XP, Vista and Windows Server 2008.  In order to build 32-bit
debug and release executables, Microsoft Visual C++ 2008 Express Edition is
required at the very least.  In order to build 64-bit debug and release
executables, Visual Studio 2008 Standard Edition is required at the very
least.  In order to build all of the above, as well as generate release builds
that make use of Profile Guided Optimisation (PG0), Visual Studio 2008
Professional Edition is required at the very least.  The official Python
releases are built with this version of Visual Studio.

For other Windows platforms and compilers, see ../PC/readme.txt.

All you need to do is open the workspace "pcbuild.sln" in Visual Studio,
select the desired combination of configuration and platform and eventually
build the solution. Unless you are going to debug a problem in the core or
you are going to create an optimized build you want to select "Release" as

The PCbuild directory is compatible with all versions of Visual Studio from
VS C++ Express Edition over the standard edition up to the professional
edition. However the express edition does not support features like solution
folders or profile guided optimization (PGO). The missing bits and pieces
won't stop you from building Python.

The solution is configured to build the projects in the correct order. "Build
Solution" or F7 takes care of dependencies except for x64 builds. To make
cross compiling x64 builds on a 32bit OS possible the x64 builds require a 
32bit version of Python.

   You probably don't want to build most of the other subprojects, unless
   you're building an entire Python distribution from scratch, or
   specifically making changes to the subsystems they implement, or are
   running a Python core buildbot test slave; see SUBPROJECTS below)

When using the Debug setting, the output files have a _d added to
their name:  python32_d.dll, python_d.exe, parser_d.pyd, and so on. Both
the build and rt batch files accept a -d option for debug builds.

The 32bit builds end up in the solution folder PCbuild while the x64 builds
land in the amd64 subfolder. The PGI and PGO builds for profile guided
optimization end up in their own folders, too.

Legacy support

You can find build directories for older versions of Visual Studio and 
Visual C++ in the PC directory. The legacy build directories are no longer
actively maintained and may not work out of the box.

    Visual C++ 6.0
    Visual Studio 2003 (7.1)
    Visual Studio 2005 (8.0)


Visual Studio 2008 uses version 9 of the C runtime (MSVCRT9).  The executables
are linked to a CRT "side by side" assembly which must be present on the target
machine.  This is avalible under the VC/Redist folder of your visual studio
distribution. On XP and later operating systems that support
side-by-side assemblies it is not enough to have the msvcrt90.dll present,
it has to be there as a whole assembly, that is, a folder with the .dll
and a .manifest.  Also, a check is made for the correct version.
Therefore, one should distribute this assembly with the dlls, and keep
it in the same directory.  For compatibility with older systems, one should
also set the PATH to this directory so that the dll can be found.
For more info, see the Readme in the VC/Redist folder.

These subprojects should build out of the box.  Subprojects other than the
main ones (pythoncore, python, pythonw) generally build a DLL (renamed to
.pyd) from a specific module so that users don't have to load the code
supporting that module unless they import the module.

    .dll and .lib
    pythonw.exe, a variant of python.exe that doesn't pop up a DOS box
    tests of the Python C API, run via Lib/test/, and
    implemented by module Modules/_testcapimodule.c
    Python wrapper for accelerated XML parsing, which incorporates stable
    code from the Expat project:
    large tables of Unicode data
    play sounds (typically .wav files) under Windows

Python-controlled subprojects that wrap external projects:
    Wraps SQLite 3.6.21, which is currently built by sqlite3.vcproj (see below).
    Wraps the Tk windowing system.  Unlike _sqlite3, there's no
    corresponding tcltk.vcproj-type project that builds Tcl/Tk from vcproj's
    within our pcbuild.sln, which means this module expects to find a
    pre-built Tcl/Tk in either ..\..\tcltk for 32-bit or ..\..\tcltk64 for
    64-bit (relative to this directory).  See below for instructions to build
    Python wrapper for the libbz2 compression library.  Homepage
    Download the source from the copy into the dist

    svn export

    ** NOTE: if you use the Tools\buildbot\external(-amd64).bat approach for
    obtaining external sources then you don't need to manually get the source
    above via subversion. **

    A custom pre-link step in the bz2 project settings should manage to
    build bzip2-1.0.5\libbz2.lib by magic before bz2.pyd (or bz2_d.pyd) is
    linked in PCbuild\.
    However, the bz2 project is not smart enough to remove anything under
    bzip2-1.0.5\ when you do a clean, so if you want to rebuild bzip2.lib
    you need to clean up bzip2-1.0.5\ by hand.

    All of this managed to build libbz2.lib in 
    bzip2-1.0.5\$platform-$configuration\, which the Python project links in.

    Python wrapper for the secure sockets library.

    Get the source code through

    svn export

    ** NOTE: if you use the Tools\buildbot\external(-amd64).bat approach for
    obtaining external sources then you don't need to manually get the source
    above via subversion. **

    Alternatively, get the latest version from
    You can (theoretically) use any version of OpenSSL you like - the
    build process will automatically select the latest version.

    You must install the NASM assembler from
    for x86 builds.  Put nasmw.exe anywhere in your PATH.
    Note: recent releases of nasm only have nasm.exe. Just rename it to 

    You can also install ActivePerl from
    if you like to use the official sources instead of the files from 
    python's subversion repository. The svn version contains pre-build
    makefiles and assembly files.

    The build process makes sure that no patented algorithms are included.
    For now RC5, MDC2 and IDEA are excluded from the build. You may have 
    to manually remove $(OBJ_D)\i_*.obj from ms\nt.mak if the build process
    complains about missing files or forbidden IDEA. Again the files provided
    in the subversion repository are already fixed.

    The MSVC project simply invokes PCBuild/ to perform
    the build.  This Python script locates and builds your OpenSSL
    installation, then invokes a simple makefile to build the final .pyd. attempts to catch the most common errors (such as not
    being able to find OpenSSL sources, or not being able to find a Perl
    that works with OpenSSL) and give a reasonable error message.
    If you have a problem that doesn't seem to be handled correctly
    (eg, you know you have ActivePerl but we can't find it), please take
    a peek at and suggest patches.  Note that
    should be able to be run directly from the command-line. isn't clever enough to clean OpenSSL - you must do
    this by hand.

The subprojects above wrap external projects Python doesn't control, and as
such, a little more work is required in order to download the relevant source 
files for each project before they can be built.  The buildbots do this each
time they're built, so the easiest approach is to run either external.bat or 
external-amd64.bat in the ..\Tools\buildbot directory from ..\, i.e.:

    C:\..\\projects\python\trunk\PCbuild>cd ..

This extracts all the external subprojects from
via Subversion (so you'll need an svn.exe on your PATH) and places them in 
..\.. (relative to this directory).  The external(-amd64).bat scripts will
also build a debug build of Tcl/Tk; there aren't any equivalent batch files
for building release versions of Tcl/Tk lying around in the Tools\buildbot
directory.  If you need to build a release version of Tcl/Tk it isn't hard
though, take a look at the relevant external(-amd64).bat file and find the
two nmake lines, then call each one without the 'DEBUG=1' parameter, i.e.:

The external-amd64.bat file contains this for tcl:
    nmake -f COMPILERFLAGS=-DWINVER=0x0500 DEBUG=1 MACHINE=AMD64 INSTALLDIR=..\..\tcltk64 clean all install

So for a release build, you'd call it as:
    nmake -f COMPILERFLAGS=-DWINVER=0x0500 MACHINE=AMD64 INSTALLDIR=..\..\tcltk64 clean all install

    XXX Should we compile with OPTS=threads?
    XXX Our installer copies a lot of stuff out of the Tcl/Tk install
    XXX directory.  Is all of that really needed for Python use of Tcl/Tk?

This will be cleaned up in the future; ideally Tcl/Tk will be brought into our
pcbuild.sln as custom .vcproj files, just as we've recently done with the
sqlite3.vcproj file, which will remove the need for Tcl/Tk to be built
separately via a batch file.

XXX trent.nelson 02-Apr-08:
    Having the external subprojects in ..\.. relative to this directory is a
    bit of a nuisance when you're working on py3k and trunk in parallel and
    your directory layout mimics that of Python's subversion layout, e.g.:


    I'd like to change things so that external subprojects are fetched from
    ..\external instead of ..\.., then provide some helper scripts or batch
    files that would set up a new ..\external directory with svn checkouts of
    the relevant branches in, or
    alternatively, use junctions to link ..\external with a pre-existing
    externals directory being used by another branch.  i.e. if I'm usually
    working on trunk (and have previously created trunk\external via the
    provided batch file), and want to do some work on py3k, I'd set up a
    junction as follows (using the directory structure above as an example):

        C:\..\python\trunk\external <- already exists and has built versions
                                       of the external subprojects 

        C:\..\python\branches\py3k>linkd.exe external ..\..\trunk\external
        Link created at: external

    Only a slight tweak would be needed to the buildbots such that bots
    building trunk and py3k could make use of the same facility.  (2.5.x
    builds need to be kept separate as they're using Visual Studio 7.1.)
/XXX trent.nelson 02-Apr-08

Building for Itanium

Official support for Itanium builds have been dropped from the build. Please
contact us and provide patches if you are interested in Itanium builds.

The project files support a ReleaseItanium configuration which creates
Win64/Itanium binaries. For this to work, you need to install the Platform
SDK, in particular the 64-bit support. This includes an Itanium compiler
(future releases of the SDK likely include an AMD64 compiler as well).
In addition, you need the Visual Studio plugin for external C compilers,
from The plugin will wrap cl.exe, to
locate the proper target compiler, and convert compiler options
accordingly. The project files require atleast version 0.9.

Building for AMD64

The build process for AMD64 / x64 is very similar to standard builds. You just
have to set x64 as platform. In addition, the HOST_PYTHON environment variable
must point to a Python interpreter (at least 2.4), to support cross-compilation.

Building Python Using the free MS Toolkit Compiler

Microsoft has withdrawn the free MS Toolkit Compiler, so this can no longer
be considered a supported option. Instead you can use the free VS C++ Express

Profile Guided Optimization

The solution has two configurations for PGO. The PGInstrument
configuration must be build first. The PGInstrument binaries are
lniked against a profiling library and contain extra debug
information. The PGUpdate configuration takes the profiling data and
generates optimized binaries.

The build_pgo.bat script automates the creation of optimized binaries. It
creates the PGI files, runs the unit test suite or PyBench with the PGI
python and finally creates the optimized files.

Static library

The solution has no configuration for static libraries. However it is easy
it build a static library instead of a DLL. You simply have to set the 
"Configuration Type" to "Static Library (.lib)" and alter the preprocessor
macro "Py_ENABLE_SHARED" to "Py_NO_ENABLE_SHARED". You may also have to
change the "Runtime Library" from "Multi-threaded DLL (/MD)" to 
"Multi-threaded (/MT)".

Visual Studio properties

The PCbuild solution makes heavy use of Visual Studio property files 
(*.vsprops). The properties can be viewed and altered in the Property
Manager (View -> Other Windows -> Property Manager).

 * debug (debug macro: _DEBUG)
 * pginstrument (PGO)
 * pgupdate (PGO)
    +-- pginstrument
 * pyd (python extension, release build)
    +-- release
    +-- pyproject
 * pyd_d (python extension, debug build)
    +-- debug
    +-- pyproject
 * pyproject (base settings for all projects, user macros like PyDllName)
 * release (release macro: NDEBUG)
 * x64 (AMD64 / x64 platform specific settings)

The pyproject propertyfile defines _WIN32 and x64 defines _WIN64 and _M_X64
although the macros are set by the compiler, too. The GUI doesn't always know
about the macros and confuse the user with false information.


If you want to create your own extension module DLL, there's an example
with easy-to-follow instructions in ../PC/example/; read the file
readme.txt there first.