Fix for persistent connection plugin on Python3. Note that fixes are also needed to each terminal plugin. This PR only fixes the ios terminal (as proof that this approach is workable.) Future PRs can address the other terminal types.
* On Python3, pickle needs to work with byte strings, not text strings.
* Set the pickle protocol version to 0 because we're using a pty to feed data to the connection plugin. A pty can't have control characters. So we have to send ascii only. That means
only using protocol=0 for pickling the data.
* ansible-connection isn't being used with py3 in the bug but it needs
several changes to work with python3.
* In python3, closing the pty too early causes no data to be sent. So
leave stdin open until after we finish with the ansible-connection
process.
* Fix typo using traceback.format_exc()
* Cleanup unnecessary StringIO, BytesIO, and to_bytes calls
* Modify the network_cli and terminal plugins for py3 compat. Lots of mixing of text and byte strings that needs to be straightened out to be compatible with python3
* Documentation for the bytes<=>text strategy for terminal plugins
* Update unittests for more bytes-oriented internals
Fixes#24355
Ansible will now automatically retry a connection if SSH returns an error:
mux_client_hello_exchange: write packet: Broken pipe
This is probably a bug in SSH, but because it's safe to retry this
connection there is no need for Ansible to fail because of it.
* Update module_utils.six to latest
We've been held back on the version of six we could use on the module
side to 1.4.x because of python-2.4 compatibility. Now that our minimum
is Python-2.6, we can update to the latest version of six in
module_utils and get rid of the second copy in lib/ansible/compat.
* Move retry logic into _ssh_retry decorator, and apply to exec_command, put_file and fetch_file
* Update tests to reflect change
* Move _ssh_retry to _run, and update tests to reflect
* piped should use exec_command instead of removed _exec_command
* Rework tests to support selectors instead of select.select
* fixes error where eos would close the cli shell
* fixes network_cli connection plugin to check before calling open_shell()
* fixes json commands being sent over eapi
* removes unneeded supports_multiplexing var
* refactors terminal_prompts_re to terminal_stdout_re
* refactors terminal_errors_re to terminal_stderr_re
* updates network_cli unit test cases
The python 'future' module provides a 'builtins' package
to emulate the py3 'builtins' modules. If installed, the
unit tests that reference builtins.__import__ fail because
the future 'builtins' is imported and it is missing __import__.
Use ansible.compat.six.moves.builtins instead of importing
'builtins' or __builtin__
Fixes#14996
- centralized skipping
- also fixed module name broken by previous refactor
- let action modules handle async processing
- moved async into base action class's module exec
- action plugins can now run final action as async
- actually skip copy if base skips
- fixed normal for new paths
- ensure internal stat is never async
- default poll to 10 as per docs
- added hint for callback fix on poll
- restructured late tmp, now a pipeline query
- moving action handler to connection as networking does
- fixed network assumption invocation is always passed
- centralized key cleanup, normalized internal var
- _supress_tmpdir_delete now in _ansible_xxx and gets removed from results
- delay internal key removal till after we use em
- nicer tmp removing, using existing methods
- moved cleanup tmp flag to mking tmp func
At the moment, this change will use EPoll on Linux, KQueue on *BSDs,
etc, so it should alleviate problems with too many open file
descriptors.
* Bundle a copy of selectors2 so that we have the selectors API everywhere.
* Add licensing information to selectors2 file so it's clear what the
licensing terms and conditions are.
* Exclude the bundled copy of selectors2 from our boilerplate code-smell test
* Rewrite ssh_run tests to attempt to work around problem with mocking
select on shippable
Fixes#14143
* Fix a test failure on Python 3.6
tox -e py36 failed with
======================================================================
ERROR: test_action_base__execute_module (units.plugins.action.test_action.TestActionBase)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/mg/src/ansible/test/units/plugins/action/test_action.py", line 507, in test_action_base__execute_module
self.assertEqual(action_base._execute_module(), dict(_ansible_parsed=True, rc=0, stdout="ok", stdout_lines=['ok']))
File "/home/mg/src/ansible/lib/ansible/plugins/action/__init__.py", line 596, in _execute_module
remote_module_path = self._connection._shell.join_path(tmp, remote_module_filename)
File "/home/mg/opt/python36/lib/python3.6/unittest/mock.py", line 939, in __call__
return _mock_self._mock_call(*args, **kwargs)
File "/home/mg/opt/python36/lib/python3.6/unittest/mock.py", line 1005, in _mock_call
ret_val = effect(*args, **kwargs)
File "/home/mg/src/ansible/.tox/py36/lib/python3.6/posixpath.py", line 92, in join
genericpath._check_arg_types('join', a, *p)
File "/home/mg/src/ansible/.tox/py36/lib/python3.6/genericpath.py", line 149, in _check_arg_types
(funcname, s.__class__.__name__)) from None
TypeError: join() argument must be str or bytes, not 'MagicMock'
because os.path.join() now checks argument types since Python 3.6 (due
to pathlib support, I expect).
* Use a more realistic module name in test
The behavior now matches GNU diff.
Fixes#14094.
Example of output before this change:
TASK [healthchecks.io : hourly healthchecks.io ping] ***************************
changed: [ranka]
--- before: /etc/cron.hourly/mg-healthchecks-dot-io
+++ after: /tmp/tmpOTvXTw
@@ -1,2 +1,2 @@
#!/bin/sh
-curl -sS https://hchk.io/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx > /dev/null+curl -sS https://hchk.io/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx > /dev/null
after this change:
TASK [healthchecks.io : hourly healthchecks.io ping] ***************************
changed: [ranka]
--- before: /etc/cron.hourly/mg-healthchecks-dot-io
+++ after: /tmp/tmpOTvXTw
@@ -1,2 +1,2 @@
#!/bin/sh
-curl -sS https://hchk.io/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx > /dev/null
\ No newline at end of file
+curl -sS https://hchk.io/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx > /dev/null
The added unit tests contain more examples.
This commit also takes care to avoid "no newline at EOF" warnings when
no_log is in effect, and also when modules return dicts rather than
strings. (It also removes trailing whitespace from using json
serialization when diffing dicts, because I hate trailing whitespace in
Python source files, even if they're test files.)
This plugin can be used with the lpass cli interface for lastpass.
[lastpass-cli](https://github.com/lastpass/lastpass-cli)
Example:
Add a lookup to your playbooks/variables somewhere:
```
some_variable: "{{ lookup('lastpass','Some Lastpass entry name or ID', field='username') }}"
```
Usage:
* start a lpass session prior to using ansible
* run ansible
* logout when finished
```
lpass login user@domain.com
ansible-playbook foo.yml
lpass logout
```
Since we no longer use a post-validated task in _process_pending_results, we
need to be sure to template fields used in original_task as they are raw and
may contain variables.
This patch also moves the handler tracking to be per-uuid, not per-object.
Doing it per-object had implications for the above due to the fact that the
copy of the original task is now being used, so the only sure way is to track
based on the uuid instead.
Fixes#18289
* Fix synchronize retries
The synchronize module munges its task args on every invocation of
run(). This was problematic because the munged data was not fit for use
by a second pass of the synchronize module. Correct this by using a copy
of the task args on every invocation of run() so that the original args
are not affected.
Local testing using this playbook seems to confirm that things work as
expected:
- hosts: all
tasks:
- delay: 2
register: task_result
retries: 1
until: task_result.rc == 0
synchronize:
dest: /tmp/out
mode: pull
src: /tmp/nonexistent/
fixes#18281
* Update synchroncization fixture assertions
When we started operating on a copy of the task args the test assertions
were no longer asserting things about the munged state but of the
pristine state. Convert the copy of task args to a class member so that
it can be compared against later in testing and update the assertions to
check this munged copy.
* Shuffle objects around for cleaner testing
Attach the temporary args dict to the task rather than the action as
this makes updating the existing tests cleaner.
This adds back the change to the network_cli plugin. Ths change adds
the ensure_connect decorator to the open_shell() method to make sure
the connection is valid before trying to open a shell.
The issue was due to the addition of the decorator that will call
_connect() when there is no connection. The _connect() method should
have been mocked in the test case. This commit fixes the test
case as well
Change was originally reverted in c414ded69a
* Replace pipes.quote for shlex_quote
* More migration of pipes.quote to shlex_quote
Note that we cannot yet move module code over. Modules have six-1.4
bundled which does not have shlex_quote. This shouldn't be a problem as
the function is still importable from pipes.quote. It's just that this
has become an implementation detail that makes us want to import from
shlex instead.
Once we get rid of the python2.4 dependency we can update to a newer
version of bundled six module-side and then we're free to use
shlex_quote everywhere.
Previous changes addressed a corner case, which unfortunately introduced
another bug. This patch adds a new flag to the host state (did_rescue) which
is set to true when the rescue portion of a block completes. This flag is
then checked in _check_failed_state() when the fail_state != FAILED_NONE.
This lead to the discovery of another bug - current strategies are not advancing
hosts to ITERATING_COMPLETE after doing a peek at the next task, leaving the
host state in the run_state of the final task. To address this, before gathering
the list of failed hosts in StrategyBase.run(), a final pass through the iterator
for all hosts is done to ensure each host is in its final state. This way, no
strategy derived from StrategyBase has to worry about it and it's handled.
Fixes#17983
- Use assertRaisesRegexp to make sure correct exceptions are raised.
- Set docker_command to avoid docker dependency (skips find_executable).
- Use a fake path for docker_command to make sure mock.patch is working.
- Remove shebangs from:
- ini files
- unit tests
- module_utils
- plugins
- module_docs_fragments
- non-executable Makefiles
- Change non-modules from '/usr/bin/python' to '/usr/bin/env python'.
- Change '/bin/env' to '/usr/bin/env'.
Also removed main functions from unit tests (since they no longer
have a shebang) and fixed a python 3 compatibility issue with
update_bundled.py so it does not need to specify a python 2 shebang.
A script was added to check for unexpected shebangs in files.
This script is run during CI on Shippable.
Nothing seems to use this now.
Was added originally added in2d11cfab92f9d26448461b4bc81f466d1910a15e
but the code that used it was removed in
e02b98274b
If the sftp fails, roll over to scp by default. This saves users
from having to know about the scp_if_ssh method when sftp is broken
on the remote host.
* Improve unit testing of 'password' lookup
The tests showed some UnicodeErrors for the
cases where the 'chars' param include unicode,
causing the 'getattr(string, c, c)' to fail.
So the candidate char generation code try/excepts
UnicodeErrors there now.
Some refactoring of the password.py module to make
it easier to test, and some new tests that cover more
of the password and salt generation.
* More refactoring and fixes.
* manual merge of text enc fixes from pr17475
* moving methods to module scope
* more refactoring
* A few more text encoding fixes/merges
* remove now unused code
* Add test cases and data for _gen_candidate_chars
* more test coverage for password lookup
* wip
* More text encoding fixes and test coverage
* cleanups
* reenable text_type assert
* Remove unneeded conditional in _random_password
* Add docstring for _gen_candidate_chars
* remove redundant to_text and list comphenesion
* Move set of 'chars' default in _random_password
on py2, C.DEFAULT_PASSWORD_CHARS is a regular str
type, so the assert here fails. Move setting the
default into the method and to_text(DEFAULT_PASSWORD_CHARS)
if it's needed.
* combine _random_password and _gen_password
* s/_create_password_file/_create_password_file_dir
* native strings for exception msgs
* move password to_text to _read_password_file
* move to_bytes(content) to _write_password_file
* add more test assertions about genned pw's
* Some cleanups to alikins and abadger's password lookup refactoring:
* Make DEFAULT_PASSWORD_CHARS into a text string in constants.py
- Move this into the nonconfigurable section of constants.
* Make utils.encrypt.do_encrypt() return a text string because all the
hashes in passlib should be returning ascii-only strings and they are
text strings in python3.
* Make the split up of functions more sane:
- Don't split such that conditionals have to occur in two separate functions.
- Don't go overboard: Good to split file system manipulation from parsing
but we don't need to do every file manipulation in a separate
function.
- Don't split so that creation of the password store happens in two
parts.
- Don't split in such a way that no decisions are made in run.
* Organize functions by when it gets called from run().
* Run all potential characters through the gen_candidate_chars function
because it does both normalization and validation.
* docstrings for functions
* Change when we store salt slightly. Store it whenever it was already
present in the file as well as when encrypt is requested. This will
head of potential idempotence bugs where a user has two playbook tasks
using the same password and in one they need it encrypted but in the
other they need it plaintext.
* Reorganize tests to follow the order of the functions so it's easier
to figure out if/where a function has been tested.
* Add tests for the functions that read and write the password file.
* Add tests of run() when the password has already been created.
* Test coverage currently at 100%
We couldn't copy to_unicode, to_bytes, to_str into module_utils because
of licensing. So once created it we had two sets of functions that did
the same things but had different implementations. To remedy that, this
change removes the ansible.utils.unicode versions of those functions.
for `VariableManager._get_magic_variables()`.
This saves a lot of time re-iterating the nearly always constant global
list of groups and their members.
Generate once and cache, and invalidate cache in case `add_host:` or
`group_by:` are used.
* Rm py2.7+ code in docker connection plugin
The docker connection plugin was using subprocess.check_output
which only exists in python 2.7 and later. Connection plugins
need to support python2.6 so this replaces it with Popen/communicate()
* Handle docker ver errors in docker connection
Add unit tests for DockerConnection
Fixes#16971
test/units/plugins/action/test_action.py had code
for handling a bug in python 3.4's mock_open that
causes errors when reading binary data.
Moved to compat/tests/mock.py so other tests can
use it by default.
Instead of immediately returning a failed code (indicating a break in
the play execution), we internally 'or' that failure code with the result
(now an integer flag instead of a boolean) so that we can properly handle
the rescue/always portions of blocks and still remember that the break
condition was hit.
Fixes#16937
* switch cwd to basedir of task
This restores previous behaviour in pre 2.0 and allows for 'local type' plugins
and actions to have a more predictable relative path.
fixes#14489
* removed FIXME since prev commit 'fixes' this
* fix tests, now they need a loader (thanks jimi!)
Due to the fact that roles may be instantiated with different sets of
params (multiple inclusions of the same role or via role dependencies),
simply tracking notified handlers by name does not work. This patch
changes the way we track handler notifications by using the handler
object itself instead of just the name, allowing for multiple internal
instances. Normally this would be bad, but we also modify the way we
search for handlers by first looking at the notifying tasks dependency
chain (ensuring that roles find their own handlers first) and then at
the main list of handlers, using the first match it finds.
This patch also modifies the way we setup the internal list of handlers,
which should allow us to correctly identify if a notified handler exists
more easily.
Fixes#15084
This allows the PlaybookExecutor to receive more information regarding
what happened internal to the TaskQueueManager and strategy, to determine
things like whether or not the play iteration should stop.
Fixes#15523
* Port urls.py to python3
Fixes (largely normalizing byte vs text strings) for python3
* Rework what we do with attributes that aren't set already.
* Comments
* Make ziploader's ansible and ansible.module_utils libraries into
namespace packages.
* Move __version__ and __author__ from ansible/__init__ to
ansible/release.py. This is because namespace packages only load one
__init__.py. If that is not the __init__.py with the author and
version info then those won't be available.
* In ziplaoder, move the version ito ANSIBLE_CONSTANTS.
* Change PluginLoader to properly construct the path to the plugins even
when namespace packages are present.
This makes our recursive, ast.parse performance measures as fast as
pre-ziploader baseline.
Since this unittest isn't testing that the returned module data is
correct we don't need to worry about os.rename not having any module
data. Should devise a separate test for the module and caching code
* 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
Before this patch, if config was ['/some/path'] then it would enter the
else block and config would be set to [].
The regression this patch fixes was introduced by 700db154.
this was taken out in an effort to default to the user's shell but creates issues as this is not known ahead of time
and its painful to set executable and shell_type for all servers, it should only be needed for those that restrict the user
to specific shells and when /bin/sh is not available. raw and command may still bypass this by explicitly passing None.
fixes#13882
still conditional
commit 24efa310b58c431b4d888a6315d1285da918f670
Author: James Cammarata <jimi@sngx.net>
Date: Tue Dec 29 11:23:52 2015 -0500
Adding an additional test for copy exclusion
Adds a negative test for the situation when an exclusion doesn't
exist in the target to be copied.
commit 643ba054877cf042177d65e6e2958178bdd2fe88
Merge: e6ee59f66a8f7e
Author: James Cammarata <jimi@sngx.net>
Date: Tue Dec 29 10:59:18 2015 -0500
Merge branch 'speedup' of https://github.com/chrismeyersfsu/ansible into chrismeyersfsu-speedup
commit 66a8f7e873
Author: Chris Meyers <chris.meyers.fsu@gmail.com>
Date: Mon Dec 28 09:47:00 2015 -0500
better api and tests added
* _copy_results = deepcopy for better performance
* _copy_results_exclude to deepcopy but exclude certain fields. Pop
fields that do not need to be deep copied. Re-assign popped fields
after deep copy so we don't modify the original, to be copied, object.
* _copy_results_exclude unit tests
commit 93490960ff
Author: Chris Meyers <chris.meyers.fsu@gmail.com>
Date: Fri Dec 25 23:17:26 2015 -0600
remove uneeded deepcopy fields
* Fix the task_vars parameter to not default to a mutable type (dict)
* Implement invocation in the base class's run() method have each action
module call the run() method's implemention in the base class.
* Return values from the action plugins' run() method takes the return
value from the base class run() method into account so that invocation
makes its way to the output.
Fixes#12869
There doesn't appear to be anything that actually uses tmp_path in the
connection plugins so we don't need to pass that in to exec_command.
That change also means that we don't need to pass tmp_path around in
many places in the action plugins any more. there may be more cleanup
that can be done there as well (the action plugin's public run() method
takes tmp as a keyword arg but that may not be necessary).
As a sideeffect of this patch, some potential problems with chmod and
the patch, assemble, copy, and template modules has been fixed (those
modules called _remote_chmod() with the wrong order for their
parameters. Removing the tmp parameter fixed them.)
This adds a parse_address(pattern) utility function that returns
(host,port), and uses it wherever where we accept IPv4 and IPv6
addresses and hostnames (or host patterns): the inventory parser
the the add_host action plugin.
It also introduces a more extensive set of unit tests that supersedes
the old add_host unit tests (which didn't actually test add_host, but
only the parsing function).
Also making PlayContext a child class of the Playbook Base class,
which gives it access to all of the FieldAttribute code to ensure
field values are correctly typed after post_validation
Fixes#11381
Toshio Kuratomi