community.general/lib/ansible/executor/module_common.py

928 lines
41 KiB
Python
Raw Normal View History

# (c) 2013-2014, Michael DeHaan <michael.dehaan@gmail.com>
# (c) 2015 Toshio Kuratomi <tkuratomi@ansible.com>
#
# This file is part of Ansible
#
# Ansible is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Ansible is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Ansible. If not, see <http://www.gnu.org/licenses/>.
# Make coding more python3-ish
from __future__ import (absolute_import, division, print_function)
__metaclass__ = type
import ast
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
import base64
import datetime
import imp
import json
import os
import shlex
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
import zipfile
import re
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
from io import BytesIO
from ansible.release import __version__, __author__
from ansible import constants as C
from ansible.errors import AnsibleError
from ansible.executor.powershell import module_manifest as ps_manifest
from ansible.module_utils._text import to_bytes, to_text, to_native
from ansible.plugins.loader import module_utils_loader
# Must import strategy and use write_locks from there
# If we import write_locks directly then we end up binding a
# variable to the object and then it never gets updated.
from ansible.executor import action_write_locks
from ansible.utils.display import Display
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
display = Display()
REPLACER = b"#<<INCLUDE_ANSIBLE_MODULE_COMMON>>"
REPLACER_VERSION = b"\"<<ANSIBLE_VERSION>>\""
REPLACER_COMPLEX = b"\"<<INCLUDE_ANSIBLE_MODULE_COMPLEX_ARGS>>\""
REPLACER_WINDOWS = b"# POWERSHELL_COMMON"
REPLACER_JSONARGS = b"<<INCLUDE_ANSIBLE_MODULE_JSON_ARGS>>"
REPLACER_SELINUX = b"<<SELINUX_SPECIAL_FILESYSTEMS>>"
# We could end up writing out parameters with unicode characters so we need to
# specify an encoding for the python source file
ENCODING_STRING = u'# -*- coding: utf-8 -*-'
b_ENCODING_STRING = b'# -*- coding: utf-8 -*-'
# module_common is relative to module_utils, so fix the path
_MODULE_UTILS_PATH = os.path.join(os.path.dirname(__file__), '..', 'module_utils')
# ******************************************************************************
ANSIBALLZ_TEMPLATE = u'''%(shebang)s
%(coding)s
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
_ANSIBALLZ_WRAPPER = True # For test-module script to tell this is a ANSIBALLZ_WRAPPER
# This code is part of Ansible, but is an independent component.
# The code in this particular templatable string, and this templatable string
# only, is BSD licensed. Modules which end up using this snippet, which is
# dynamically combined together by Ansible still belong to the author of the
# module, and they may assign their own license to the complete work.
#
# Copyright (c), James Cammarata, 2016
# Copyright (c), Toshio Kuratomi, 2016
#
# Redistribution and use in source and binary forms, with or without modification,
# are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
# USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
def _ansiballz_main():
%(rlimit)s
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
import os
import os.path
import sys
import __main__
# For some distros and python versions we pick up this script in the temporary
# directory. This leads to problems when the ansible module masks a python
# library that another import needs. We have not figured out what about the
# specific distros and python versions causes this to behave differently.
#
# Tested distros:
# Fedora23 with python3.4 Works
# Ubuntu15.10 with python2.7 Works
# Ubuntu15.10 with python3.4 Fails without this
# Ubuntu16.04.1 with python3.5 Fails without this
# To test on another platform:
# * use the copy module (since this shadows the stdlib copy module)
# * Turn off pipelining
# * Make sure that the destination file does not exist
# * ansible ubuntu16-test -m copy -a 'src=/etc/motd dest=/var/tmp/m'
# This will traceback in shutil. Looking at the complete traceback will show
# that shutil is importing copy which finds the ansible module instead of the
# stdlib module
scriptdir = None
try:
scriptdir = os.path.dirname(os.path.realpath(__main__.__file__))
except (AttributeError, OSError):
# Some platforms don't set __file__ when reading from stdin
# OSX raises OSError if using abspath() in a directory we don't have
# permission to read (realpath calls abspath)
pass
if scriptdir is not None:
sys.path = [p for p in sys.path if p != scriptdir]
import base64
2018-07-15 01:30:02 +00:00
import imp
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
import shutil
import tempfile
import zipfile
if sys.version_info < (3,):
bytes = str
2018-07-15 01:30:02 +00:00
MOD_DESC = ('.py', 'U', imp.PY_SOURCE)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
PY3 = False
else:
unicode = str
2018-07-15 01:30:02 +00:00
MOD_DESC = ('.py', 'r', imp.PY_SOURCE)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
PY3 = True
ZIPDATA = """%(zipdata)s"""
2018-07-15 01:30:02 +00:00
# Note: temp_path isn't needed once we switch to zipimport
def invoke_module(modlib_path, temp_path, json_params):
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
# When installed via setuptools (including python setup.py install),
# ansible may be installed with an easy-install.pth file. That file
# may load the system-wide install of ansible rather than the one in
# the module. sitecustomize is the only way to override that setting.
z = zipfile.ZipFile(modlib_path, mode='a')
# py3: modlib_path will be text, py2: it's bytes. Need bytes at the end
sitecustomize = u'import sys\\nsys.path.insert(0,"%%s")\\n' %% modlib_path
sitecustomize = sitecustomize.encode('utf-8')
# Use a ZipInfo to work around zipfile limitation on hosts with
# clocks set to a pre-1980 year (for instance, Raspberry Pi)
zinfo = zipfile.ZipInfo()
zinfo.filename = 'sitecustomize.py'
zinfo.date_time = ( %(year)i, %(month)i, %(day)i, %(hour)i, %(minute)i, %(second)i)
z.writestr(zinfo, sitecustomize)
2018-07-15 01:30:02 +00:00
# Note: Remove the following section when we switch to zipimport
# Write the module to disk for imp.load_module
module = os.path.join(temp_path, '__main__.py')
with open(module, 'wb') as f:
f.write(z.read('__main__.py'))
f.close()
# End pre-zipimport section
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
z.close()
# Put the zipped up module_utils we got from the controller first in the python path so that we
# can monkeypatch the right basic
sys.path.insert(0, modlib_path)
# Monkeypatch the parameters into basic
from ansible.module_utils import basic
basic._ANSIBLE_ARGS = json_params
%(coverage)s
# Run the module! By importing it as '__main__', it thinks it is executing as a script
2018-07-15 01:30:02 +00:00
with open(module, 'rb') as mod:
imp.load_module('__main__', mod, module, MOD_DESC)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
# Ansible modules must exit themselves
print('{"msg": "New-style module did not handle its own exit", "failed": true}')
sys.exit(1)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
def debug(command, zipped_mod, json_params):
# The code here normally doesn't run. It's only used for debugging on the
# remote machine.
#
# The subcommands in this function make it easier to debug ansiballz
# modules. Here's the basic steps:
#
# Run ansible with the environment variable: ANSIBLE_KEEP_REMOTE_FILES=1 and -vvv
# to save the module file remotely::
# $ ANSIBLE_KEEP_REMOTE_FILES=1 ansible host1 -m ping -a 'data=october' -vvv
#
# Part of the verbose output will tell you where on the remote machine the
# module was written to::
# [...]
# <host1> SSH: EXEC ssh -C -q -o ControlMaster=auto -o ControlPersist=60s -o KbdInteractiveAuthentication=no -o
# PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o ConnectTimeout=10 -o
# ControlPath=/home/badger/.ansible/cp/ansible-ssh-%%h-%%p-%%r -tt rhel7 '/bin/sh -c '"'"'LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8
# LC_MESSAGES=en_US.UTF-8 /usr/bin/python /home/badger/.ansible/tmp/ansible-tmp-1461173013.93-9076457629738/ping'"'"''
# [...]
#
# Login to the remote machine and run the module file via from the previous
# step with the explode subcommand to extract the module payload into
# source files::
# $ ssh host1
# $ /usr/bin/python /home/badger/.ansible/tmp/ansible-tmp-1461173013.93-9076457629738/ping explode
# Module expanded into:
# /home/badger/.ansible/tmp/ansible-tmp-1461173408.08-279692652635227/ansible
#
# You can now edit the source files to instrument the code or experiment with
# different parameter values. When you're ready to run the code you've modified
# (instead of the code from the actual zipped module), use the execute subcommand like this::
# $ /usr/bin/python /home/badger/.ansible/tmp/ansible-tmp-1461173013.93-9076457629738/ping execute
# Okay to use __file__ here because we're running from a kept file
basedir = os.path.join(os.path.abspath(os.path.dirname(__file__)), 'debug_dir')
args_path = os.path.join(basedir, 'args')
script_path = os.path.join(basedir, '__main__.py')
if command == 'excommunicate':
print('The excommunicate debug command is deprecated and will be removed in 2.11. Use execute instead.')
command = 'execute'
if command == 'explode':
# transform the ZIPDATA into an exploded directory of code and then
# print the path to the code. This is an easy way for people to look
# at the code on the remote machine for debugging it in that
# environment
z = zipfile.ZipFile(zipped_mod)
for filename in z.namelist():
if filename.startswith('/'):
raise Exception('Something wrong with this module zip file: should not contain absolute paths')
dest_filename = os.path.join(basedir, filename)
if dest_filename.endswith(os.path.sep) and not os.path.exists(dest_filename):
os.makedirs(dest_filename)
else:
directory = os.path.dirname(dest_filename)
if not os.path.exists(directory):
os.makedirs(directory)
f = open(dest_filename, 'wb')
f.write(z.read(filename))
f.close()
# write the args file
f = open(args_path, 'wb')
f.write(json_params)
f.close()
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
print('Module expanded into:')
print('%%s' %% basedir)
exitcode = 0
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
elif command == 'execute':
# Execute the exploded code instead of executing the module from the
# embedded ZIPDATA. This allows people to easily run their modified
# code on the remote machine to see how changes will affect it.
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
# Set pythonpath to the debug dir
sys.path.insert(0, basedir)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
# read in the args file which the user may have modified
with open(args_path, 'rb') as f:
json_params = f.read()
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
# Monkeypatch the parameters into basic
from ansible.module_utils import basic
basic._ANSIBLE_ARGS = json_params
# Run the module! By importing it as '__main__', it thinks it is executing as a script
import imp
with open(script_path, 'r') as f:
importer = imp.load_module('__main__', f, script_path, ('.py', 'r', imp.PY_SOURCE))
# Ansible modules must exit themselves
print('{"msg": "New-style module did not handle its own exit", "failed": true}')
sys.exit(1)
else:
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
print('WARNING: Unknown debug command. Doing nothing.')
exitcode = 0
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
return exitcode
#
# See comments in the debug() method for information on debugging
#
ANSIBALLZ_PARAMS = %(params)s
if PY3:
ANSIBALLZ_PARAMS = ANSIBALLZ_PARAMS.encode('utf-8')
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
try:
# There's a race condition with the controller removing the
# remote_tmpdir and this module executing under async. So we cannot
# store this in remote_tmpdir (use system tempdir instead)
2018-07-15 01:30:02 +00:00
# Only need to use [ansible_module]_payload_ in the temp_path until we move to zipimport
# (this helps ansible-test produce coverage stats)
temp_path = tempfile.mkdtemp(prefix='ansible_%(ansible_module)s_payload_')
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
zipped_mod = os.path.join(temp_path, 'ansible_%(ansible_module)s_payload.zip')
with open(zipped_mod, 'wb') as modlib:
modlib.write(base64.b64decode(ZIPDATA))
if len(sys.argv) == 2:
exitcode = debug(sys.argv[1], zipped_mod, ANSIBALLZ_PARAMS)
else:
2018-07-15 01:30:02 +00:00
# Note: temp_path isn't needed once we switch to zipimport
invoke_module(zipped_mod, temp_path, ANSIBALLZ_PARAMS)
finally:
try:
shutil.rmtree(temp_path)
except (NameError, OSError):
# tempdir creation probably failed
pass
sys.exit(exitcode)
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
if __name__ == '__main__':
_ansiballz_main()
'''
ANSIBALLZ_COVERAGE_TEMPLATE = '''
# Access to the working directory is required by coverage.
# Some platforms, such as macOS, may not allow querying the working directory when using become to drop privileges.
try:
os.getcwd()
except OSError:
os.chdir('/')
os.environ['COVERAGE_FILE'] = '%(coverage_output)s'
import atexit
import coverage
cov = coverage.Coverage(config_file='%(coverage_config)s')
def atexit_coverage():
cov.stop()
cov.save()
atexit.register(atexit_coverage)
cov.start()
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
'''
ANSIBALLZ_RLIMIT_TEMPLATE = '''
import resource
existing_soft, existing_hard = resource.getrlimit(resource.RLIMIT_NOFILE)
# adjust soft limit subject to existing hard limit
requested_soft = min(existing_hard, %(rlimit_nofile)d)
if requested_soft != existing_soft:
try:
resource.setrlimit(resource.RLIMIT_NOFILE, (requested_soft, existing_hard))
except ValueError:
# some platforms (eg macOS) lie about their hard limit
pass
'''
def _strip_comments(source):
# Strip comments and blank lines from the wrapper
buf = []
for line in source.splitlines():
l = line.strip()
if not l or l.startswith(u'#'):
continue
buf.append(line)
return u'\n'.join(buf)
if C.DEFAULT_KEEP_REMOTE_FILES:
# Keep comments when KEEP_REMOTE_FILES is set. That way users will see
# the comments with some nice usage instructions
ACTIVE_ANSIBALLZ_TEMPLATE = ANSIBALLZ_TEMPLATE
else:
# ANSIBALLZ_TEMPLATE stripped of comments for smaller over the wire size
ACTIVE_ANSIBALLZ_TEMPLATE = _strip_comments(ANSIBALLZ_TEMPLATE)
class ModuleDepFinder(ast.NodeVisitor):
# Caveats:
# This code currently does not handle:
# * relative imports from py2.6+ from . import urls
IMPORT_PREFIX_SIZE = len('ansible.module_utils.')
def __init__(self, *args, **kwargs):
"""
Walk the ast tree for the python module.
Save submodule[.submoduleN][.identifier] into self.submodules
self.submodules will end up with tuples like:
- ('basic',)
- ('urls', 'fetch_url')
- ('database', 'postgres')
- ('database', 'postgres', 'quote')
It's up to calling code to determine whether the final element of the
dotted strings are module names or something else (function, class, or
variable names)
"""
super(ModuleDepFinder, self).__init__(*args, **kwargs)
self.submodules = set()
def visit_Import(self, node):
# import ansible.module_utils.MODLIB[.MODLIBn] [as asname]
for alias in (a for a in node.names if a.name.startswith('ansible.module_utils.')):
py_mod = alias.name[self.IMPORT_PREFIX_SIZE:]
py_mod = tuple(py_mod.split('.'))
self.submodules.add(py_mod)
self.generic_visit(node)
def visit_ImportFrom(self, node):
# Specialcase: six is a special case because of its
# import logic
if node.names[0].name == '_six':
self.submodules.add(('_six',))
elif node.module.startswith('ansible.module_utils'):
where_from = node.module[self.IMPORT_PREFIX_SIZE:]
if where_from:
# from ansible.module_utils.MODULE1[.MODULEn] import IDENTIFIER [as asname]
# from ansible.module_utils.MODULE1[.MODULEn] import MODULEn+1 [as asname]
# from ansible.module_utils.MODULE1[.MODULEn] import MODULEn+1 [,IDENTIFIER] [as asname]
py_mod = tuple(where_from.split('.'))
for alias in node.names:
self.submodules.add(py_mod + (alias.name,))
else:
# from ansible.module_utils import MODLIB [,MODLIB2] [as asname]
for alias in node.names:
self.submodules.add((alias.name,))
self.generic_visit(node)
def _slurp(path):
if not os.path.exists(path):
raise AnsibleError("imported module support code does not exist at %s" % os.path.abspath(path))
fd = open(path, 'rb')
data = fd.read()
fd.close()
return data
def _get_shebang(interpreter, task_vars, templar, args=tuple()):
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
"""
Note not stellar API:
Returns None instead of always returning a shebang line. Doing it this
way allows the caller to decide to use the shebang it read from the
file rather than trust that we reformatted what they already have
correctly.
"""
interpreter_config = u'ansible_%s_interpreter' % os.path.basename(interpreter).strip()
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
if interpreter_config not in task_vars:
return (None, interpreter)
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
interpreter = templar.template(task_vars[interpreter_config].strip())
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
shebang = u'#!' + interpreter
if args:
shebang = shebang + u' ' + u' '.join(args)
return (shebang, interpreter)
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
def recursive_finder(name, data, py_module_names, py_module_cache, zf):
"""
Using ModuleDepFinder, make sure we have all of the module_utils files that
the module its module_utils files needs.
"""
# Parse the module and find the imports of ansible.module_utils
try:
tree = ast.parse(data)
except (SyntaxError, IndentationError) as e:
raise AnsibleError("Unable to import %s due to %s" % (name, e.msg))
finder = ModuleDepFinder()
finder.visit(tree)
#
# Determine what imports that we've found are modules (vs class, function.
# variable names) for packages
#
normalized_modules = set()
# Loop through the imports that we've found to normalize them
# Exclude paths that match with paths we've already processed
# (Have to exclude them a second time once the paths are processed)
module_utils_paths = [p for p in module_utils_loader._get_paths(subdirs=False) if os.path.isdir(p)]
module_utils_paths.append(_MODULE_UTILS_PATH)
for py_module_name in finder.submodules.difference(py_module_names):
module_info = None
if py_module_name[0] == 'six':
# Special case the python six library because it messes up the
# import process in an incompatible way
module_info = imp.find_module('six', module_utils_paths)
py_module_name = ('six',)
idx = 0
elif py_module_name[0] == '_six':
# Special case the python six library because it messes up the
# import process in an incompatible way
module_info = imp.find_module('_six', [os.path.join(p, 'six') for p in module_utils_paths])
py_module_name = ('six', '_six')
idx = 0
else:
# Check whether either the last or the second to last identifier is
# a module name
for idx in (1, 2):
if len(py_module_name) < idx:
break
try:
module_info = imp.find_module(py_module_name[-idx],
[os.path.join(p, *py_module_name[:-idx]) for p in module_utils_paths])
break
except ImportError:
continue
# Could not find the module. Construct a helpful error message.
if module_info is None:
msg = ['Could not find imported module support code for %s. Looked for' % (name,)]
if idx == 2:
msg.append('either %s.py or %s.py' % (py_module_name[-1], py_module_name[-2]))
else:
msg.append(py_module_name[-1])
raise AnsibleError(' '.join(msg))
# Found a byte compiled file rather than source. We cannot send byte
# compiled over the wire as the python version might be different.
# imp.find_module seems to prefer to return source packages so we just
# error out if imp.find_module returns byte compiled files (This is
# fragile as it depends on undocumented imp.find_module behaviour)
if module_info[2][2] not in (imp.PY_SOURCE, imp.PKG_DIRECTORY):
msg = ['Could not find python source for imported module support code for %s. Looked for' % name]
if idx == 2:
msg.append('either %s.py or %s.py' % (py_module_name[-1], py_module_name[-2]))
else:
msg.append(py_module_name[-1])
raise AnsibleError(' '.join(msg))
if idx == 2:
# We've determined that the last portion was an identifier and
# thus, not part of the module name
py_module_name = py_module_name[:-1]
# If not already processed then we've got work to do
# If not in the cache, then read the file into the cache
# We already have a file handle for the module open so it makes
# sense to read it now
if py_module_name not in py_module_cache:
if module_info[2][2] == imp.PKG_DIRECTORY:
# Read the __init__.py instead of the module file as this is
# a python package
normalized_name = py_module_name + ('__init__',)
if normalized_name not in py_module_names:
normalized_path = os.path.join(os.path.join(module_info[1], '__init__.py'))
normalized_data = _slurp(normalized_path)
py_module_cache[normalized_name] = (normalized_data, normalized_path)
normalized_modules.add(normalized_name)
else:
normalized_name = py_module_name
if normalized_name not in py_module_names:
normalized_path = module_info[1]
normalized_data = module_info[0].read()
module_info[0].close()
py_module_cache[normalized_name] = (normalized_data, normalized_path)
normalized_modules.add(normalized_name)
# Make sure that all the packages that this module is a part of
# are also added
for i in range(1, len(py_module_name)):
py_pkg_name = py_module_name[:-i] + ('__init__',)
if py_pkg_name not in py_module_names:
pkg_dir_info = imp.find_module(py_pkg_name[-1],
[os.path.join(p, *py_pkg_name[:-1]) for p in module_utils_paths])
normalized_modules.add(py_pkg_name)
py_module_cache[py_pkg_name] = (_slurp(pkg_dir_info[1]), pkg_dir_info[1])
# FIXME: Currently the AnsiBallZ wrapper monkeypatches module args into a global
# variable in basic.py. If a module doesn't import basic.py, then the AnsiBallZ wrapper will
# traceback when it tries to monkypatch. So, for now, we have to unconditionally include
# basic.py.
#
# In the future we need to change the wrapper to monkeypatch the args into a global variable in
# their own, separate python module. That way we won't require basic.py. Modules which don't
# want basic.py can import that instead. AnsibleModule will need to change to import the vars
# from the separate python module and mirror the args into its global variable for backwards
# compatibility.
if ('basic',) not in py_module_names:
pkg_dir_info = imp.find_module('basic', module_utils_paths)
normalized_modules.add(('basic',))
py_module_cache[('basic',)] = (_slurp(pkg_dir_info[1]), pkg_dir_info[1])
# End of AnsiballZ hack
#
# iterate through all of the ansible.module_utils* imports that we haven't
# already checked for new imports
#
# set of modules that we haven't added to the zipfile
unprocessed_py_module_names = normalized_modules.difference(py_module_names)
for py_module_name in unprocessed_py_module_names:
py_module_path = os.path.join(*py_module_name)
py_module_file_name = '%s.py' % py_module_path
zf.writestr(os.path.join("ansible/module_utils",
py_module_file_name), py_module_cache[py_module_name][0])
Ansible Config part2 (#27448) * Ansible Config part2 - made dump_me nicer, added note this is not prod - moved internal key removal function to vars - carry tracebacks in errors we can now show tracebacks for plugins on vvv - show inventory plugin tracebacks on vvv - minor fixes to cg groups plugin - draft config from plugin docs - made search path warning 'saner' (top level dirs only) - correctly display config entries and others - removed unneeded code - commented out some conn plugin specific from base.yml - also deprecated sudo/su - updated ssh conn docs - shared get option method for connection plugins - note about needing eval for defaults - tailored yaml ext - updated strategy entry - for connection pliugins, options load on plugin load - allow for long types in definitions - better display in ansible-doc - cleaned up/updated source docs and base.yml - added many descriptions - deprecated include toggles as include is - draft backwards compat get_config - fixes to ansible-config, added --only-changed - some code reoorg - small license headers - show default in doc type - pushed module utils details to 5vs - work w/o config file - PEPE ATE! - moved loader to it's own file - fixed rhn_register test - fixed boto requirement in make tests - I ate Pepe - fixed dynamic eval of defaults - better doc code skip ipaddr filter tests when missing netaddr removed devnull string from config better becoem resolution * killed extra space with extreeme prejudice cause its an affront against all that is holy that 2 spaces touch each other! shippable timing out on some images, but merging as it passes most
2017-08-15 20:38:59 +00:00
display.vvvvv("Using module_utils file %s" % py_module_cache[py_module_name][1])
# Add the names of the files we're scheduling to examine in the loop to
# py_module_names so that we don't re-examine them in the next pass
# through recursive_finder()
py_module_names.update(unprocessed_py_module_names)
for py_module_file in unprocessed_py_module_names:
recursive_finder(py_module_file, py_module_cache[py_module_file][0], py_module_names, py_module_cache, zf)
# Save memory; the file won't have to be read again for this ansible module.
del py_module_cache[py_module_file]
def _is_binary(b_module_data):
textchars = bytearray(set([7, 8, 9, 10, 12, 13, 27]) | set(range(0x20, 0x100)) - set([0x7f]))
start = b_module_data[:1024]
return bool(start.translate(None, textchars))
def _find_module_utils(module_name, b_module_data, module_path, module_args, task_vars, templar, module_compression, async_timeout, become,
become_method, become_user, become_password, become_flags, environment):
"""
Given the source of the module, convert it to a Jinja2 template to insert
module code and return whether it's a new or old style module.
"""
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
module_substyle = module_style = 'old'
# module_style is something important to calling code (ActionBase). It
# determines how arguments are formatted (json vs k=v) and whether
# a separate arguments file needs to be sent over the wire.
# module_substyle is extra information that's useful internally. It tells
# us what we have to look to substitute in the module files and whether
# we're using module replacer or ansiballz to format the module itself.
if _is_binary(b_module_data):
module_substyle = module_style = 'binary'
elif REPLACER in b_module_data:
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
# Do REPLACER before from ansible.module_utils because we need make sure
# we substitute "from ansible.module_utils basic" for REPLACER
module_style = 'new'
module_substyle = 'python'
b_module_data = b_module_data.replace(REPLACER, b'from ansible.module_utils.basic import *')
elif b'from ansible.module_utils.' in b_module_data:
module_style = 'new'
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
module_substyle = 'python'
elif REPLACER_WINDOWS in b_module_data:
module_style = 'new'
module_substyle = 'powershell'
b_module_data = b_module_data.replace(REPLACER_WINDOWS, b'#Requires -Module Ansible.ModuleUtils.Legacy')
2017-11-21 03:08:30 +00:00
elif re.search(b'#Requires -Module', b_module_data, re.IGNORECASE) \
or re.search(b'#Requires -Version', b_module_data, re.IGNORECASE)\
or re.search(b'#AnsibleRequires -OSVersion', b_module_data, re.IGNORECASE) \
or re.search(b'#AnsibleRequires -CSharpUtil', b_module_data, re.IGNORECASE):
Fixes for WinRM/PowerShell support in v2. - Add support for inserting module args into PowerShell modules. Fixes #11661. - Support Windows paths containing spaces. Applies changes from #10727 to v2. Fixes #9999. Should also fix ansible/ansible-modules-core#944 and ansible/ansible-modules-core#1007. - Change how execution policy is set for running remote scripts. Applies changes from #11092 to v2. Also fixes ansible/ansible-modules-core#1776. - Use codepage 65001 (UTF-8) for WinRM connection instead of default (CP437), convert command to UTF-8 and results from UTF-8. Replaces changes from #10024. Fixes #11198. - Close WinRM connection when task completes. - Use win_stat, win_file and win_copy modules instead of stat, file and copy when called from within other action plugins (only when using WinRM+PowerShell). - Unquote Windows path arguments before passing to win_stat, win_file, win_copy and slurp modules (only when using WinRM/PowerShell). - Check for win_ping module to determine if core modules are missing (only when using WinRM/PowerShell). - Add stdout_lines to result from running low level commands (so stdout_lines is available when using raw/script). - Update copy action plugin to use shell functions for joining paths and checking for trailing slash. - Update fetch action plugin to unquote source path when using Windows paths. - Add win_copy and win_template action plugins that inherit from copy and template. - Support running .bat and .cmd scripts using default system encoding instead of UTF-8. - Always send PowerShell commands as base64-encoded blobs to allow for running simple PowerShell commands via raw. - Support running modules on Windows with interpreters other than PowerShell. - Update integration tests to support above changes and test unicode fixes. - Add test for win_user error from ansible/ansible-modules-core#1241 (fixed by ansible/ansible-modules-core#1774). - Add test for additional win_stat output values (implemented by ansible/ansible-modules-core#1473). - Add test for OS architecture and name from setup.ps1 (implemented by ansible/ansible-modules-core#1100). All WinRM integration tests pass for me with these changes.
2015-07-24 16:39:54 +00:00
module_style = 'new'
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
module_substyle = 'powershell'
elif REPLACER_JSONARGS in b_module_data:
module_style = 'new'
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
module_substyle = 'jsonargs'
elif b'WANT_JSON' in b_module_data:
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
module_substyle = module_style = 'non_native_want_json'
shebang = None
# Neither old-style, non_native_want_json nor binary modules should be modified
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
# except for the shebang line (Done by modify_module)
if module_style in ('old', 'non_native_want_json', 'binary'):
return b_module_data, module_style, shebang
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
output = BytesIO()
py_module_names = set()
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
if module_substyle == 'python':
params = dict(ANSIBLE_MODULE_ARGS=module_args,)
python_repred_params = repr(json.dumps(params))
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
try:
compression_method = getattr(zipfile, module_compression)
except AttributeError:
display.warning(u'Bad module compression string specified: %s. Using ZIP_STORED (no compression)' % module_compression)
compression_method = zipfile.ZIP_STORED
lookup_path = os.path.join(C.DEFAULT_LOCAL_TMP, 'ansiballz_cache')
cached_module_filename = os.path.join(lookup_path, "%s-%s" % (module_name, module_compression))
zipdata = None
# Optimization -- don't lock if the module has already been cached
if os.path.exists(cached_module_filename):
display.debug('ANSIBALLZ: using cached module: %s' % cached_module_filename)
with open(cached_module_filename, 'rb') as module_data:
zipdata = module_data.read()
else:
if module_name in action_write_locks.action_write_locks:
display.debug('ANSIBALLZ: Using lock for %s' % module_name)
lock = action_write_locks.action_write_locks[module_name]
else:
# If the action plugin directly invokes the module (instead of
# going through a strategy) then we don't have a cross-process
# Lock specifically for this module. Use the "unexpected
# module" lock instead
display.debug('ANSIBALLZ: Using generic lock for %s' % module_name)
lock = action_write_locks.action_write_locks[None]
display.debug('ANSIBALLZ: Acquiring lock')
with lock:
display.debug('ANSIBALLZ: Lock acquired: %s' % id(lock))
# Check that no other process has created this while we were
# waiting for the lock
if not os.path.exists(cached_module_filename):
display.debug('ANSIBALLZ: Creating module')
# Create the module zip data
zipoutput = BytesIO()
zf = zipfile.ZipFile(zipoutput, mode='w', compression=compression_method)
# Note: If we need to import from release.py first,
# remember to catch all exceptions: https://github.com/ansible/ansible/issues/16523
zf.writestr('ansible/__init__.py',
b'from pkgutil import extend_path\n__path__=extend_path(__path__,__name__)\n__version__="' +
to_bytes(__version__) + b'"\n__author__="' +
to_bytes(__author__) + b'"\n')
zf.writestr('ansible/module_utils/__init__.py', b'from pkgutil import extend_path\n__path__=extend_path(__path__,__name__)\n')
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
zf.writestr('__main__.py', b_module_data)
py_module_cache = {('__init__',): (b'', '[builtin]')}
recursive_finder(module_name, b_module_data, py_module_names, py_module_cache, zf)
zf.close()
zipdata = base64.b64encode(zipoutput.getvalue())
# Write the assembled module to a temp file (write to temp
# so that no one looking for the file reads a partially
# written file)
if not os.path.exists(lookup_path):
# Note -- if we have a global function to setup, that would
# be a better place to run this
os.makedirs(lookup_path)
display.debug('ANSIBALLZ: Writing module')
with open(cached_module_filename + '-part', 'wb') as f:
f.write(zipdata)
# Rename the file into its final position in the cache so
# future users of this module can read it off the
# filesystem instead of constructing from scratch.
display.debug('ANSIBALLZ: Renaming module')
os.rename(cached_module_filename + '-part', cached_module_filename)
display.debug('ANSIBALLZ: Done creating module')
if zipdata is None:
display.debug('ANSIBALLZ: Reading module after lock')
# Another process wrote the file while we were waiting for
# the write lock. Go ahead and read the data from disk
# instead of re-creating it.
try:
with open(cached_module_filename, 'rb') as f:
zipdata = f.read()
except IOError:
raise AnsibleError('A different worker process failed to create module file. '
'Look at traceback for that process for debugging information.')
zipdata = to_text(zipdata, errors='surrogate_or_strict')
shebang, interpreter = _get_shebang(u'/usr/bin/python', task_vars, templar)
if shebang is None:
shebang = u'#!/usr/bin/python'
# Enclose the parts of the interpreter in quotes because we're
# substituting it into the template as a Python string
interpreter_parts = interpreter.split(u' ')
interpreter = u"'{0}'".format(u"', '".join(interpreter_parts))
# FUTURE: the module cache entry should be invalidated if we got this value from a host-dependent source
rlimit_nofile = C.config.get_config_value('PYTHON_MODULE_RLIMIT_NOFILE', variables=task_vars)
if not isinstance(rlimit_nofile, int):
rlimit_nofile = int(templar.template(rlimit_nofile))
if rlimit_nofile:
rlimit = ANSIBALLZ_RLIMIT_TEMPLATE % dict(
rlimit_nofile=rlimit_nofile,
)
else:
rlimit = ''
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
coverage_config = os.environ.get('_ANSIBLE_COVERAGE_CONFIG')
if coverage_config:
# Enable code coverage analysis of the module.
# This feature is for internal testing and may change without notice.
coverage = ANSIBALLZ_COVERAGE_TEMPLATE % dict(
coverage_config=coverage_config,
coverage_output=os.environ['_ANSIBLE_COVERAGE_OUTPUT']
)
else:
coverage = ''
now = datetime.datetime.utcnow()
output.write(to_bytes(ACTIVE_ANSIBALLZ_TEMPLATE % dict(
zipdata=zipdata,
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
ansible_module=module_name,
params=python_repred_params,
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
shebang=shebang,
interpreter=interpreter,
coding=ENCODING_STRING,
year=now.year,
month=now.month,
day=now.day,
hour=now.hour,
minute=now.minute,
second=now.second,
AnsiballZ improvements Now that we don't need to worry about python-2.4 and 2.5, we can make some improvements to the way AnsiballZ handles modules. * Change AnsiballZ wrapper to use import to invoke the module We need the module to think of itself as a script because it could be coded as: main() or as: if __name__ == '__main__': main() Or even as: if __name__ == '__main__': random_function_name() A script will invoke all of those. Prior to this change, we invoked a second Python interpreter on the module so that it really was a script. However, this means that we have to run python twice (once for the AnsiballZ wrapper and once for the module). This change makes the module think that it is a script (because __name__ in the module == '__main__') but it's actually being invoked by us importing the module code. There's three ways we've come up to do this. * The most elegant is to use zipimporter and tell the import mechanism that the module being loaded is __main__: * https://github.com/abadger/ansible/blob/5959f11c9ddb7b6eaa9c3214560bd85e631d4055/lib/ansible/executor/module_common.py#L175 * zipimporter is nice because we do not have to extract the module from the zip file and save it to the disk when we do that. The import machinery does it all for us. * The drawback is that modules do not have a __file__ which points to a real file when they do this. Modules could be using __file__ to for a variety of reasons, most of those probably have replacements (the most common one is to find a writable directory for temporary files. AnsibleModule.tmpdir should be used instead) We can monkeypatch __file__ in fom AnsibleModule initialization but that's kind of gross. There's no way I can see to do this from the wrapper. * Next, there's imp.load_module(): * https://github.com/abadger/ansible/blob/340edf7489/lib/ansible/executor/module_common.py#L151 * imp has the nice property of allowing us to set __name__ to __main__ without changing the name of the file itself * We also don't have to do anything special to set __file__ for backwards compatibility (although the reason for that is the drawback): * Its drawback is that it requires the file to exist on disk so we have to explicitly extract it from the zipfile and save it to a temporary file * The last choice is to use exec to execute the module: * https://github.com/abadger/ansible/blob/f47a4ccc76/lib/ansible/executor/module_common.py#L175 * The code we would have to maintain for this looks pretty clean. In the wrapper we create a ModuleType, set __file__ on it, read the module's contents in from the zip file and then exec it. * Drawbacks: We still have to explicitly extract the file's contents from the zip archive instead of letting python's import mechanism handle it. * Exec also has hidden performance issues and breaks certain assumptions that modules could be making about their own code: http://lucumr.pocoo.org/2011/2/1/exec-in-python/ Our plan is to use imp.load_module() for now, deprecate the use of __file__ in modules, and switch to zipimport once the deprecation period for __file__ is over (without monkeypatching a fake __file__ in via AnsibleModule). * Rename the name of the AnsiBallZ wrapped module This makes it obvious that the wrapped module isn't the module file that we distribute. It's part of trying to mitigate the fact that the module is now named __main)).py in tracebacks. * Shield all wrapper symbols inside of a function With the new import code, all symbols in the wrapper become visible in the module. To mitigate the chance of collisions, move most symbols into a toplevel function. The only symbols left in the global namespace are now _ANSIBALLZ_WRAPPER and _ansiballz_main. revised porting guide entry Integrate code coverage collection into AnsiballZ. ci_coverage ci_complete
2018-06-20 18:23:59 +00:00
coverage=coverage,
rlimit=rlimit,
)))
b_module_data = output.getvalue()
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
elif module_substyle == 'powershell':
# Powershell/winrm don't actually make use of shebang so we can
# safely set this here. If we let the fallback code handle this
# it can fail in the presence of the UTF8 BOM commonly added by
# Windows text editors
shebang = u'#!powershell'
# create the common exec wrapper payload and set that as the module_data
# bytes
b_module_data = ps_manifest._create_powershell_wrapper(
b_module_data, module_args, environment, async_timeout, become,
become_method, become_user, become_password, become_flags,
module_substyle
)
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
elif module_substyle == 'jsonargs':
module_args_json = to_bytes(json.dumps(module_args))
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
# these strings could be included in a third-party module but
# officially they were included in the 'basic' snippet for new-style
# python modules (which has been replaced with something else in
# ansiballz) If we remove them from jsonargs-style module replacer
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
# then we can remove them everywhere.
python_repred_args = to_bytes(repr(module_args_json))
b_module_data = b_module_data.replace(REPLACER_VERSION, to_bytes(repr(__version__)))
b_module_data = b_module_data.replace(REPLACER_COMPLEX, python_repred_args)
b_module_data = b_module_data.replace(REPLACER_SELINUX, to_bytes(','.join(C.DEFAULT_SELINUX_SPECIAL_FS)))
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
# The main event -- substitute the JSON args string into the module
b_module_data = b_module_data.replace(REPLACER_JSONARGS, module_args_json)
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
facility = b'syslog.' + to_bytes(task_vars.get('ansible_syslog_facility', C.DEFAULT_SYSLOG_FACILITY), errors='surrogate_or_strict')
b_module_data = b_module_data.replace(b'syslog.LOG_USER', facility)
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
return (b_module_data, module_style, shebang)
def modify_module(module_name, module_path, module_args, templar, task_vars=None, module_compression='ZIP_STORED', async_timeout=0, become=False,
become_method=None, become_user=None, become_password=None, become_flags=None, environment=None):
"""
Used to insert chunks of code into modules before transfer rather than
doing regular python imports. This allows for more efficient transfer in
a non-bootstrapping scenario by not moving extra files over the wire and
also takes care of embedding arguments in the transferred modules.
This version is done in such a way that local imports can still be
used in the module code, so IDEs don't have to be aware of what is going on.
Example:
from ansible.module_utils.basic import *
... will result in the insertion of basic.py into the module
from the module_utils/ directory in the source tree.
For powershell, this code effectively no-ops, as the exec wrapper requires access to a number of
properties not available here.
"""
2017-09-12 07:11:13 +00:00
task_vars = {} if task_vars is None else task_vars
environment = {} if environment is None else environment
with open(module_path, 'rb') as f:
# read in the module source
b_module_data = f.read()
(b_module_data, module_style, shebang) = _find_module_utils(module_name, b_module_data, module_path, module_args, task_vars, templar, module_compression,
async_timeout=async_timeout, become=become, become_method=become_method,
become_user=become_user, become_password=become_password, become_flags=become_flags,
environment=environment)
if module_style == 'binary':
return (b_module_data, module_style, to_text(shebang, nonstring='passthru'))
elif shebang is None:
b_lines = b_module_data.split(b"\n", 1)
if b_lines[0].startswith(b"#!"):
b_shebang = b_lines[0].strip()
# shlex.split on python-2.6 needs bytes. On python-3.x it needs text
args = shlex.split(to_native(b_shebang[2:], errors='surrogate_or_strict'))
# _get_shebang() takes text strings
args = [to_text(a, errors='surrogate_or_strict') for a in args]
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
interpreter = args[0]
b_new_shebang = to_bytes(_get_shebang(interpreter, task_vars, templar, args[1:])[0],
errors='surrogate_or_strict', nonstring='passthru')
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
if b_new_shebang:
b_lines[0] = b_shebang = b_new_shebang
if os.path.basename(interpreter).startswith(u'python'):
b_lines.insert(1, b_ENCODING_STRING)
shebang = to_text(b_shebang, nonstring='passthru', errors='surrogate_or_strict')
Ziploader * Ziploader proof of concept (jimi-c) * Cleanups to proof of concept ziploader branch: * python3 compatible base64 encoding * zipfile compression (still need to enable toggling this off for systems without zlib support in python) * Allow non-wildcard imports (still need to make this recusrsive so that we can have module_utils code that imports other module_utils code.) * Better tracebacks: module filename is kept and module_utils directory is kept so that tracebacks show the real filenames that the errors appear in. * Make sure we import modules that are used into the module_utils files that they are used in. * Set ansible version in a more pythonic way for ziploader than we were doing in module replacer * Make it possible to set the module compression as an inventory var This may be necessary on systems where python has been compiled without zlib compression. * Refactoring of module_common code: * module replacer only replaces values that make sense for that type of file (example: don't attempt to replace python imports if we're in a powershell module). * Implement configurable shebang support for ziploader wrapper * Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG) via environment variable. * Remove strip_comments param as we're never going to use it (ruins line numbering) * Don't repeat ourselves about detecting REPLACER * Add an easy way to debug * Port test-module to the ziploader-aware modify_module() * strip comments and blank lines from the wrapper so we send less over the wire. * Comments cleanup * Remember to output write the module line itself in powershell modules * for line in lines strips the newlines so we have to add them back in
2016-04-05 18:06:17 +00:00
else:
# No shebang, assume a binary module?
pass
b_module_data = b"\n".join(b_lines)
return (b_module_data, module_style, shebang)