installed to build extensions. This patch makes distutils emit a warning when
the compiler should use an SDK but that SDK is not installed, hopefully reducing
some confusion.
universal build of python on OSX 10.3 to ensure that those flags can be used
to compile code (the universal build uses compiler flags that aren't supported
on 10.3). This patches gives the same treatment to CFLAGS, PY_CFLAGS and
BLDSHARED.
This patchs makes it possible to create a universal build on OSX 10.4 and use
the result to build extensions on 10.3. It also makes it possible to override
the '-arch' and '-isysroot' compiler arguments for specific extensions.
configure time. The current check is too strict and doesn't allow building
extensions that can only run on newer versions of the OS than the version
python was build for, that is python build for 10.3 or later and an extension
for 10.4. This patch relaxes this check.
This turned out to be a reimplementation of patch 1193190.
/path/to/uninstalled/python setup.py build_ext
now failed with pyconfig.h not found. Prior to r45232
the above command did not look for pyconfig.h, but the
bug is really in the look-up code: expecting to find it
in os.curdir is a rather fragile idea.
Bug #1160802: Can't build Zope on Windows w/ 2.4.1c1.
MSVCCompiler.initialize(): set self.initialized to True, as suggested
by AMK. Else we keep growing the PATH endlessly, with each new C
extension built, until putenv() complains.
No change to NEWS because the patch that created this bug is also new
for 2.5a1 (so there's no change here to any code yet released from HEAD).
to make using "-undefined dynamic_lookup" for linking extensions more
automatic on 10.3 and later. So if we're on that platform and
MACOSX_DEPLOYMENT_TARGET is not set we now set it to the current OSX
version during configure. Additionally, distutils will pick up the
configure-time value by default.
Will backport.
scripts without modifying either the distutils installation or the
setup.py scripts of packages with which the new commands will be used.
Specifically, an option is added to distutils that allows additional
packages to be searched for command implementations in addition to
distutils.command. The additional packages can be specified on the
command line or via the installation or personal configuration files
already loaded by distutils.
For discussion, see the thread starting with:
http://mail.python.org/pipermail/distutils-sig/2004-August/004112.html
This closes SF patch #102241.
Patch from Mark Hammond:
bdist_wininst attempts to use the correct MSVC runtime for the current
version of Python. This doesn't work correctly when --target-version
is set. In that case, bdist_wininst still uses the *current*
sys.version (ie, 2.4) rather than the version specified as
--target-version. Thus, the msvc7 runtime based executable stub is
*always* used.
This patch "hard-codes" knowledge of earlier Python versions,
providing the correct result when Python 2.4 is used to build Python
2.3 and earlier distributions.
Remove the short variant (-v) of the --target-version command line
options, it conflicts with the --verbose/-v standard distutils switch.
This is basically the support for package data from Phillip Eby's
setuptools package. I've changed it only to fit it into the core
implementation rather than to live in subclasses, and added
documentation.
of hard linking against the framework).
If $MACOSX_DEPLOYMENT_TARGET is set, and >= 10.3, during configure we
setup extensions to link with dynamic lookup. We also record the
value in the Makefile.
Distutils checks whether a value for MACOSX_DEPLOYMENT_TARGET was
recorded in the Makefile, and if it was insists that the current
value matches.
This is only a partial fix because it only applies to 2.4, and the
"two python problem" exists with Python 2.3 shipped with MacOSX 10.3,
which we have no influence over.
requires and provides. requires is a sequence of strings, of the
form 'packagename-version'. The dependency checking so far merely
does an '__import__(packagename)' and checks for packagename.__version__
You can also leave off the version, and any version of the package
will be installed.
There's a special case for the package 'python' - sys.version_info
is used, so
requires= ( 'python-2.3', )
just works.
Provides is of the same format as requires - but if it's not supplied,
a provides is generated by adding the version to each entry in packages,
or modules if packages isn't there.
Provides is currently only used in the PKG-INFO file. Shortly, PyPI
will grow the ability to accept these lines, and register will be
updated to send them.
There's a new command 'checkdep' command that runs these checks.
For this version, only greater-than-or-equal checking is done. We'll
add the ability to specify an optional operator later.
msvccompiler.get_build_version().
Distributions without a pre-install-script didn't work any longer, we
must at least provide the terminating NUL character.
included in Python distributions for systems other than Windows.
Windows installers can be build on non-Windows systems as long as they
only include pure python module distributions.
Patch #892660 from Mark Hammond, for distutils bdist_wininst command.
install.c: support for a 'pre-install-script', run before anything has
been installed. Provides a 'message_box' module function for use by
either the pre-install or post-install scripts.
bdist_wininst.py: support for pre-install script. Typo (build->built),
fixes so that --target-version can still work, even when the
distribution has extension modules - in this case, we insist on
--skip-build, as we still can't actually build other versions.
install.c: support for a 'pre-install-script', run before anything has
been installed. Provides a 'message_box' module function for use by
either the pre-install or post-install scripts.
bdist_wininst.py: support for pre-install script. Typo (build->built),
fixes so that --target-version can still work, even when the
distribution has extension modules - in this case, we insist on
--skip-build, as we still can't actually build other versions.
Use case: Sometimes 'compiling' source files (with SWIG, for example)
creates additionl files which included by later sources. The win32all
setup script requires this.
There is no SF item for this, but it was discussed on distutils-sig:
http://mail.python.org/pipermail/distutils-sig/2003-November/003514.html
distutils now looks for the compiler version in sys.version, falling
back to MSVC 6 if the version isn't listed (Python 2.2 and lower).
Add helper routines for reading the registry. Refactor many
module functions into methods of the compiler to avoid passing
lots of state as arguments.
On cygwin, the setup.py script uses unixccompiler.py for compiling and linking
C extensions. The unixccompiler.py script assumes that executables do not get
special extensions, which makes sense for Unix. However, on Cygwin,
executables get an .exe extension.
This causes a problem during the configuration step (python setup.py config),
in which some temporary executables may be generated. As unixccompiler.py does
not know about the .exe extension, distutils fails to clean up after itself: it
does not remove _configtest.exe but tries to remove _configtest instead.
The attached patch to unixccompiler.py sets the correct exe_extension for
cygwin by checking if sys.platform is 'cygwin'. With this patch, distutils
cleans up after itself correctly.
Michiel de Hoon
University of Tokyo, Human Genome Center.
After some more reflection (and no negative feedback), I am reverting the
original patch and applying my version, cygwinccompiler.py-shared.diff,
instead.
My reasons are the following:
1. support for older toolchains is retained
2. support for new toolchains (i.e., ld -shared) is added
The goal of my approach is to avoid breaking older toolchains while adding
better support for newer ones.
The cygwinccompiler.get_versions() function only handles versions numbers of
the form "x.y.z". The attached patch enhances get_versions() to handle "x.y"
too (i.e., the ".z" is optional).
This change causes the unnecessary "--entry _DllMain@12" link option to be
suppressed for recent Cygwin and Mingw toolchains. Additionally, it directs
recent Mingw toolchains to use gcc instead of dllwrap during linking.
Currently, the cygwinccompiler.py compiler handling in
distutils is invoking the cygwin and mingw compilers
with the -static option.
Logically, this means that the linker should choose to
link to static libraries instead of shared/dynamically
linked libraries.
Current win32 binutils expect import libraries to have
a .dll.a suffix and static libraries to have .a suffix.
If -static is passed, it will skip the .dll.a
libraries. This is pain if one has a tree with both
static and dynamic libraries using this naming
convention, and wish to use the dynamic libraries.
The -static option being passed in distutils is to get
around a bug in old versions of binutils where it would
get confused when it found the DLLs themselves.
The decision to use static or shared libraries is site
or package specific, and should be left to the setup
script or to command line options.
specified with an absolute path, the object file is also
written to an absolute path. The patch drops the drive and
leading '/' from the source path, so a path like /path/to/foo.c
results in an object file like build/temp.i686linux/path/to/foo.o.
bdist_wininst.py we will see.)
Removed the base64 encoded binary contents, wininst.exe must be in the
same directory as this file now.
wininst.exe must be recompiled and commited each time the sources in
PC/bdist_wininst are changed.
Previously archive_util.py attempted to spawn an
external 'zip' program for the zip action, if this fails, an
attempt to import zipfile.py is made...
This bites folks who have 'old' or non-conforming zip
programs on windows platforms. This change tries the 'zipfile'
module first, falling back to spawning a zip process if
the module isn't available.
If you have source files srcdir1/foo.c and srcdir2/foo.c, the
temporary .o for both files is written to build/temp.<platform>/foo.o.
This patch sets strip_dir to false for both calls to object_filename,
so now the object files are written to temp.<platform>/srcdir1/foo.o
and .../srcdir2/foo.o.
2.2 bugfix candidate
The two long lines have been reflowed differently; hopefully someone on
BeOS can test them. Rev. 1.53 also converted string.atoi() to int(); I've
left that alone.
Fixed by catching all exceptions that are subclasses of DistutilsError,
so only the error message will be printed. You can still get the
whole traceback by enabling the Distutils debugging mode.
always available on Windows NT. When the function cannot be loaded,
get_special_folder_path raises OSError, "function not available".
Compiled the exe, and rebuilt bdist_wininst.py.
* Lib/distutils/command/bdist_rpm.py
(bdist_rpm.initialize_options): Included verify_script attribute.
(bdist_rpm.finalize_package_data): Ensure that verify_script is a filename.
(bdist_rpm._make_spec_file): Included verify_script in script_options
tuple.
* Misc/NEWS
Mention change.
[#413582] g++ must be called for c++ extensions
[#454030] distutils cannot link C++ code with GCC
topdir = "Lib/distutils"
* bcppcompiler.py
(BCPPCompiler.create_static_lib): Fixed prototype, removing extra_preargs
and extra_postargs parameters. Included target_lang parameter.
(BCPPCompiler.link): Included target_lang parameter.
* msvccompiler.py
(MSVCCompiler.create_static_lib): Fixed prototype, removing extra_preargs
and extra_postargs parameters. Included target_lang parameter.
(MSVCCompiler.link): Included target_lang parameter.
* ccompiler.py
(CCompiler): New language_map and language_order attributes, used by
CCompiler.detect_language().
(CCompiler.detect_language): New method, will return the language of
a given source, or list of sources. Individual source language is
detected using the language_map dict. When mixed sources are used,
language_order will stablish the language precedence.
(CCompiler.create_static_lib, CCompiler.link, CCompiler.link_executable,
CCompiler.link_shared_object, CCompiler.link_shared_lib):
Inlcuded target_lang parameter.
* cygwinccompiler.py
(CygwinCCompiler.link): Included target_lang parameter.
* emxccompiler.py
(EMXCCompiler.link): Included target_lang parameter.
* mwerkscompiler.py
(MWerksCompiler.link): Included target_lang parameter.
* extension.py
(Extension.__init__): New 'language' parameter/attribute, initialized
to None by default. If provided will overlap the automatic detection
made by CCompiler.detect_language(), in build_ext command.
* sysconfig.py
(customize_compiler): Check Makefile for CXX option, and also the
environment variable CXX. Use the resulting value in the 'compiler_cxx'
parameter of compiler.set_executables().
* unixccompiler.py
(UnixCCompiler): Included 'compiler_cxx' in executables dict, defaulting
to 'cc'.
(UnixCCompiler.create_static_lib): Included target_lang parameter.
(UnixCCompiler.link): Included target_lang parameter, and made
linker command use compiler_cxx, if target_lang is 'c++'.
* command/build_ext.py
(build_ext.build_extension): Pass new ext.language attribute
to compiler.link_shared_object()'s target_lang parameter. If
ext.language is not provided, detect language using
compiler.detect_language(sources) instead.
* command/config.py
(config._link): Pass already available lang parameter as target_lang
parameter of compiler.link_executable().
customize_compiler() now looks at various environment variables and uses
their values to override the configured C compiler/preprocessor/linker
binary and flags.
whether the Distutils being used supports a particularly capability.
(This idea was originally suggested by Juergen Hermann as a method
on the Distribution class. I think it makes more sense as a
function in core.py, and that's what this patch implements.)
arguments, triggering a warning instead of raising an exception. (In
1.5.2/2.0, it will print to stderr.)
Bugfix candidate for all previous versions. This changes behaviour,
but the old behaviour wasn't very useful. If Distutils version X+1
adds a new keyword argument, using the new keyword means your setup.py
file won't work with Distutils version X any more.
Strangely, two out of three patches there seem already committed; but
the essential one (get rid of the assert in object_filenames in
ccompiler.py) was not yet applied.
This makes the build procedure for Twisted work again.
This is *not* a backport candidate despite the fact that identical
code appears to exist in 2.2.2; Twisted builds fine there, so there
must have been a change elsewhere.
The default implementation calls _compile() to compile individual
files. This method must be implemented by the subclass. This change
factors out most of the remaining common code in all the compilers
except mwerks.
Two new tests are needed:
Don't skip building an extension if any of the depends files are newer
than the target.
Pass ext.depends to compiler.compile() so that it can track individual
files.
Always use _setup_compile() to do the grunt work of processing
arguments, figuring out which files to compile, and emitting debug
messages for files that are up-to-date.
Use _get_cc_args() when possible.
This change is not backwards compatible. If a compiler subclass
exists outside the distutils package, it may get called with the
unexpected keyword arg. It's easy to extend that compiler by having
it ignore the argument, and not much harder to do the right thing. If
this ends up being burdensome, we can change it before 2.3 final to
work harder at compatibility.
Also add _setup_compile() and _get_cc_args() helper functions that
factor out much of the boilerplate for each concrete compiler class.
Remove __init__ that just called base class __init__ with same args.
Fold long argument lists into fewer, shorter lines.
Remove parens in tuple unpacks.
Don't put multiple statements on one line with a semicolon.
In find_library_file() compute the library_filename() upfront.
This is a conservative version of SF patch 504889. It uses the log
module instead of calling print in various places, and it ignores the
verbose argument passed to many functions and set as an attribute on
some objects. Instead, it uses the verbosity set on the logger via
the command line.
The log module is now preferred over announce() and warn() methods
that exist only for backwards compatibility.
XXX This checkin changes a lot of modules that have no test suite and
aren't exercised by the Python build process. It will need
substantial testing.
# XXX this isn't used anywhere, and worse, it has the same name as a method
# in Command with subtly different semantics. (This one just has one
# source -> one dest; that one has many sources -> one dest.) Nuke it?
Yes. Nuke it.
modules, distutils does not understand that the build version of the
source tree is needed.
This patch fixes distutils.sysconfig to understand that the running
Python is part of the build tree and needs to use the appropriate
"shape" of the tree. This does not assume anything about the current
directory, so can be used to build 3rd-party modules using Python's
build tree as well.
This is useful since it allows us to use a non-installed debug-mode
Python with 3rd-party modules for testing. It as the side-effect that
set_python_build() is no longer needed (the hack which was added to
allow distutils to be used to build the "standard" extension modules).
This closes SF patch #547734.
--install-script ... command line option to bdist_wininst) at the end
of the installation and at the start of deinstallation. Output
(stdout, stderr) of the script (if any) is displayed in the last
screen at installation, or in a simple message box at deinstallation.
sys.argv[1] for the script will contain '-install' at installation
time or '-remove' at deinstallation time.
The installation script runs in an environment (embedded by the
bdist_wininst runtime) where an additional function is available as
builtin:
create_shortcut(path, description, filename,
[arguments[, workdir[, iconpath, iconindex]]])
Recreated this file after source changes.
installations are present, by always unlinking the destination file
before copying to it. Without the unlink(), the copied file remains
owned by its previous UID, causing the subsequent chmod() to fail.
Bugfix candidate, though it may cause changes on platforms where
file ownership behaves differently.
contain the type of the file (regular file, socket, link, &c.).
This means that install_scripts will now print
"changing mode of <file> to 775" instead of "... to 100775".
2.2 bugfix candidate, I suppose, though this isn't actually fixing a bug.
present - at least the swigged file should be named <name>_wrap.c as
this is also SWIG's default. (Even better would be to generate the
wrapped sources in a different location, but I'll leave this for
later).
Newer versions of SWIG don't accept the -dnone flag any more.
Since virtually nobody uses SWIG with distutils, this should do no
harm.
Suggested be Martin Bless on c.l.p.
crashes.
If no external zip-utility is found, the archive is created by the
zipfile module, which behaves different now than in 2.1: if the
zip-file is created in the root directory if the distribution, it will
contain an (empty) version of itself.
This triggered the above bug - so it's better to create the zip-file
far away in the TMP directory.
>
> When using 'distutils' (shipped with Python 2.1) I've found that my
> Python scripts installed with a first line of:
>
> #!/usr/bin/python2.1None
>
> This is caused by distutils trying to patch the first line of the python
> script to use the current interpreter.
distutils for the library modules built as shared objects. A better solution
appears possible, but with the threat that the distutils becomes more
magical ("complex").
This closes SF bug #458343.
bdist_wininst doesn't use the NT SCHEME any more, instead
a custom SCHEME is used, which is exchanged at installation
time, depending on the python version used.
Avoid a bogus warning frpom install_lib about installing
into a directory not on sys.path.
modules and extensions on Windows is now $PREFIX/Lib/site-packages.
Includes backwards compatibility code for pre-2.2 Pythons. Contributed
by Paul Moore.
libraries. This is done by adding a .get_source_files() method,
contributed by Rene Liebscher and slightly modified.
Remove an unused local variable spotted by PyChecker
though 'licence' is still supported for backward-compatibility
(Should I add a warning to get_licence(), or not bother?)
Also fixes an UnboundLocalError noticed by PyChecker
Phillip J. Eby
Martin v. Löwis
Tim Peters
Ronald Oussoren
Andrew M. Kuchling
Marc-André Lemburg
Bob Ippolito
Georg Brandl
Trent Mick
Armin Rigo
Anthony Baxter
Guido van Rossum
Brett Cannon
Walter Dörwald
Fred Drake
Thomas Heller
Jack Jansen
Fredrik Lundh
Mark Hammond
Sean Reifschneider
Sjoerd Mullender
Hye-Shik Chang
Andrew MacIntyre
Jeremy Hylton
Gustavo Niemeyer
Skip Montanaro
Jason Tishler
Just van Rossum
Neal Norwitz
Michael W. Hudson
Barry Warsaw
Neil Schemenauer
Greg Ward