4. Creating a Source Distribution
As shown in section A Simple Example, you use the sdist command
to create a source distribution. In the simplest case,
(assuming you haven’t specified any sdist options in the setup script
or config file), sdist creates the archive of the default format for
the current platform. The default format is a gzip’ed tar file
(.tar.gz) on Unix, and ZIP file on Windows.
You can specify as many formats as you like using the --formats
option, for example:
python setup.py sdist --formats=gztar,zip
to create a gzipped tarball and a zip file. The available formats are:
| Format |
Description |
Notes |
zip |
zip file (.zip) |
(1),(3) |
gztar |
gzip’ed tar file
(.tar.gz) |
(2) |
bztar |
bzip2’ed tar file
(.tar.bz2) |
|
ztar |
compressed tar file
(.tar.Z) |
(4) |
tar |
tar file (.tar) |
|
Notes:
- default on Windows
- default on Unix
- requires either external zip utility or
zipfile module (part
of the standard Python library since Python 1.6)
- requires the compress program.
When using any tar format (gztar, bztar, ztar or
tar) under Unix, you can specify the owner and group names
that will be set for each member of the archive.
For example, if you want all files of the archive to be owned by root:
python setup.py sdist --owner=root --group=root
4.1. Specifying the files to distribute
If you don’t supply an explicit list of files (or instructions on how to
generate one), the sdist command puts a minimal default set into the
source distribution:
- all Python source files implied by the
py_modules and
packages options
- all C source files mentioned in the
ext_modules or
libraries options
- scripts identified by the
scripts option
See Installing Scripts.
- anything that looks like a test script:
test/test*.py (currently, the
Distutils don’t do anything with test scripts except include them in source
distributions, but in the future there will be a standard for testing Python
module distributions)
README.txt (or README), setup.py (or whatever you
called your setup script), and setup.cfg
- all files that matches the
package_data metadata.
See Installing Package Data.
- all files that matches the
data_files metadata.
See Installing Additional Files.
Sometimes this is enough, but usually you will want to specify additional files
to distribute. The typical way to do this is to write a manifest template,
called MANIFEST.in by default. The manifest template is just a list of
instructions for how to generate your manifest file, MANIFEST, which is
the exact list of files to include in your source distribution. The
sdist command processes this template and generates a manifest based
on its instructions and what it finds in the filesystem.
If you prefer to roll your own manifest file, the format is simple: one filename
per line, regular files (or symlinks to them) only. If you do supply your own
MANIFEST, you must specify everything: the default set of files
described above does not apply in this case.
Changed in version 2.7: An existing generated MANIFEST will be regenerated without
sdist comparing its modification time to the one of
MANIFEST.in or setup.py.
Changed in version 2.7.1: MANIFEST files start with a comment indicating they are generated.
Files without this comment are not overwritten or removed.
Changed in version 2.7.3: sdist will read a MANIFEST file if no MANIFEST.in
exists, like it did before 2.7.
See The MANIFEST.in template section for a syntax reference.
4.3. The MANIFEST.in template
A MANIFEST.in file can be added in a project to define the list of
files to include in the distribution built by the sdist command.
When sdist is run, it will look for the MANIFEST.in file
and interpret it to generate the MANIFEST file that contains the
list of files that will be included in the package.
This mechanism can be used when the default list of files is not enough.
(See Specifying the files to distribute).
4.3.1. Principle
The manifest template has one command per line, where each command specifies a
set of files to include or exclude from the source distribution. For an
example, let’s look at the Distutils’ own manifest template:
include *.txt
recursive-include examples *.txt *.py
prune examples/sample?/build
The meanings should be fairly clear: include all files in the distribution root
matching *.txt, all files anywhere under the examples directory
matching *.txt or *.py, and exclude all directories matching
examples/sample?/build. All of this is done after the standard
include set, so you can exclude files from the standard set with explicit
instructions in the manifest template. (Or, you can use the
--no-defaults option to disable the standard set entirely.)
The order of commands in the manifest template matters: initially, we have the
list of default files as described above, and each command in the template adds
to or removes from that list of files. Once we have fully processed the
manifest template, we remove files that should not be included in the source
distribution:
- all files in the Distutils “build” tree (default
build/)
- all files in directories named
RCS, CVS, .svn,
.hg, .git, .bzr or _darcs
Now we have our complete list of files, which is written to the manifest for
future reference, and then used to build the source distribution archive(s).
You can disable the default set of included files with the
--no-defaults option, and you can disable the standard exclude set
with --no-prune.
Following the Distutils’ own manifest template, let’s trace how the
sdist command builds the list of files to include in the Distutils
source distribution:
- include all Python source files in the
distutils and
distutils/command subdirectories (because packages corresponding to
those two directories were mentioned in the packages option in the
setup script—see section Writing the Setup Script)
- include
README.txt, setup.py, and setup.cfg (standard
files)
- include
test/test*.py (standard files)
- include
*.txt in the distribution root (this will find
README.txt a second time, but such redundancies are weeded out later)
- include anything matching
*.txt or *.py in the sub-tree
under examples,
- exclude all files in the sub-trees starting at directories matching
examples/sample?/build—this may exclude files included by the
previous two steps, so it’s important that the prune command in the manifest
template comes after the recursive-include command
- exclude the entire
build tree, and any RCS, CVS,
.svn, .hg, .git, .bzr and _darcs
directories
Just like in the setup script, file and directory names in the manifest template
should always be slash-separated; the Distutils will take care of converting
them to the standard representation on your platform. That way, the manifest
template is portable across operating systems.
4.3.2. Commands
The manifest template commands are:
| Command |
Description |
| include pat1 pat2 ... |
include all files matching any of the listed
patterns |
| exclude pat1 pat2 ... |
exclude all files matching any of the listed
patterns |
| recursive-include dir pat1 pat2
... |
include all files under dir matching any of
the listed patterns |
| recursive-exclude dir pat1 pat2
... |
exclude all files under dir matching any of
the listed patterns |
| global-include pat1 pat2 ... |
include all files anywhere in the source tree
matching — & any of the listed patterns |
| global-exclude pat1 pat2 ... |
exclude all files anywhere in the source tree
matching — & any of the listed patterns |
| prune dir |
exclude all files under dir |
| graft dir |
include all files under dir |
The patterns here are Unix-style “glob” patterns: * matches any sequence of
regular filename characters, ? matches any single regular filename
character, and [range] matches any of the characters in range (e.g.,
a-z, a-zA-Z, a-f0-9_.). The definition of “regular filename
character” is platform-specific: on Unix it is anything except slash; on Windows
anything except backslash or colon.