* Add TTY check and argument to disable it (#50603)
* Fix formatting
* add changelog
* rename flag and updated help description
* add tests for tty check
* replace deprecated uses of assertRaisesRegexp to assertRaisesRegex
* fix yaml syntax
* shorten line 79
* Revert "replace deprecated uses of assertRaisesRegexp to assertRaisesRegex"
This reverts commit cea5fe1655.
* change back to assertRaisesRegexp
* fix meaning of parallel in gather_facts
* Update docs with note about parallel not always being faster
* add 'smarter' usage of gahter_timeout for parallel tasks
* restore async when needed, not always
* added typing
* parallelism tests
* Only bypass type validation for null parameters if the default is None. A default is mutually exclusive with required.
* Prevent coercing None to str type. Fail the type check instead.
- Unit tests now report warnings generated during test runs.
- Python 3.12 warnings about `os.fork` usage with threads (due to `pytest-xdist`) are suppressed.
- Added integration tests to verify forked test behavior.
Fixes#73643
* clear_notification method and simplify ifs
* Deduplicate code
* Limit number of Templar creations
* Fix sanity
* Preserve handler callbacks order as they were notified
* Symbolic modes with X or =[ugo] always use original mode (Fixes#80128)
Here's what's happening, by way of this mode example: u=,u=rX
At the first step in the loop, the "u" bits of are set to 0. On the next
step in the loop, the current stat of the filesystem object is used to
determine X, not the "new_mode" in the previous iteration of the loop. So
while most operations kind of operate left to right, "X" is always going
back to the original file to determine whether to set x bit.
The Linux "chmod" (the only one I've tested) doesn't operate this way. In
it, "X" operates on the current state the loop understands it is in,
based on previous operations (and starting with the file permissions).
This is an issue with "X" and any of the "=[ugo]" settings, because
they are lookups. For example, if a file is 755 and you do "ug=rx,o=u",
file module produces 0557 and chmod produces 0555.
This really becomes a problem when you want to recursively change a
directory of files, and the files are currently 755, but you want to
change the directory to 750 and the files to 640. In chmod you can do
"a=,ug=rX,u+w" (or "a=,u=rwX,g=rX"), and have it apply equally to the
directory and the files. I can't come up with a single way in the ansible
file module to deterministically, recursively, set a directory to 750
and the contents to 640 no matter what the current permissions are,
as the code currently is.
The fix is to pass in "new_mode" to _get_octal_mode_from_symbolic_perms
in lib/ansible/module_utils/basic.py inside _symbolic_mode_to_octal. And
then take "new_mode" as an argument and use it instead of the filesystem
object stat.st_mode value.
* Fixing my new unit test, fixing bug in test comments
Most of the bad_parsing tests were no longer running, with several of them no longer being valid.
The invalid tests have been removed and the valid ones rewritten.
Reduce the number of Galaxy API calls made during dependency resolution by fetching remote signatures afterwards, since these are not used in backtracking.
Reduce the verbosity to `-vvvv` (to match other Galaxy API calls) to see this activity.
Co-authored-by: Sviatoslav Sydorenko <webknjaz@redhat.com>
* fix using templated values for include/import role options 'public', 'allow_duplicates', and 'rolespec_validate'
* pass templated values without changing the instance
* Fix templating by setting always_post_validate to True and calling IncludeRole.post_validate() instead
ci_complete
* add changelog
* password lookup, handle ident properly when saved
Currently we format and save ident when present but we didn't account for this when reading the saved file
Also added some more robust error handling.
* ansible-galaxy - support `resolvelib >= 0.5.3, < 1.1.0`
<https://pypi.org/project/resolvelib/1.0.1> released on 2023-03-09:
- <https://github.com/sarugaku/resolvelib/blob/main/CHANGELOG.rst#101-2023-03-09>
- <https://github.com/sarugaku/resolvelib/releases/tag/1.0.1>
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
* Trigger CI by pinning resolvelib with latest version
Since resolvelib is pre-installed in our test containers, we should temporarily pin the latest version allowed to force the tests to run with that version. Once the tests have passed that commit can be reverted.
Please make those changes without force pushing, so that we keep the reference to the passing CI run. We can squash the commits when merging the PR so the temporary commits won't be in the final commit merged to the devel branch.
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
* https://github.com/ansible/ansible/pull/80196#discussion_r1136003637
Also test resolvelib with multiple supported versions.
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
* Revert "Trigger CI by pinning resolvelib with latest version"
This reverts commit 5518e5dbca.
---------
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
* Pin setuptools to lowest supported @ PEP 517 test
This allows catching the behavior of builds under old setuptools.
* Stop invoking `setup.py install` in tests
This is not the part we care about since it involves dealing with the
external runtime dependencies rather than building our source
distribution.
This patch modifies the in-tree build backend to build sdists that swap
out pointers to it in the `pyproject.toml`'s `[build-system]` section.
The effect of this is that the first build from source (for example,
from a Git checkout) uses our PEP 517 in-tree build backend. But the
produced tarball has `build-backend` set to `setuptools.build_meta`
which is the native build backend of `setuptools`. So any following
builds from that sdist will skip using the in-tree build backend,
calling the setuptools' one.
The good news is that if the first build generated the manpages, they
will be included and won't go anywhere even though, a different build
system is in place.
Combined with #80253, this will make sure not to modify the current
source checkout on that first build.
Co-authored-by: Matt Clay <matt@mystile.com>
* 🧪 Switch macOS 13.2 to 12.0 in CI
The former revealed unexpected flakiness while the latter is the
previous value that was used to be stable. This is a temporary revert.
* Skip lookup_url integration test under macOS 12.0
* ✨ Add macOS 13.2 to `ansible-test`
* 🧪 Replace macOS 12.0 with 13.2 in the CI matrix
* Skip `lookup_url` under macOS 13.2
This is due to https://wefearchange.org/2018/11/forkmacos.rst.html
that manifests itself as follows:
TASK [lookup_url : Test that retrieving a url works] ***************************
objc[15394]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called.
objc[15394]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in t
he fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
ERROR! A worker was found in a dead state
* 📝 Extend ansible-test change note w/ macOS 13.2
* 🐛 Make integration tests compatible w/ modern Git
This patch makes use of the `init.defaultBranch` setting to unify
the test across new and old Git versions since one defaults to
`master` and the other uses `main` for the default branch.
Where possible, it uses the `HEAD` committish to avoid having to
normalize the branch name.
The change fixes the following integration tests:
* `ansible-galaxy`
* `ansible-galaxy-collection-scm` (recursive collection)
* `git`
* 🐛Replace `git-symbolic-ref` with a repo template
This custom Git repository template emulates the `init.defaultBranch` setting
on Git versions below 2.28. Ref: https://superuser.com/a/1559582.
Other workarounds mentioned there, like invoking
`git symbolic-ref HEAD refs/heads/main` after each `git init` turned
out to have mysterious side effects that break the tests in surprising ways.
* 🎨 Make Git integration test non-destructive
This patch makes use of the `$HOME` environment variable to trick Git
into using a user-global config generated in the temporary directory.
* unarchive - properly handle relative path for dest
* Add integration test
* Return output of underlying commands with increased verbosity
* Revert "Return output of underlying commands with increased verbosity"
This reverts commit a2790c8275cdc5697b65670a0beffdc74b741bf6.
* Warn when a relative destination path was provided
* Create a queue per WorkerProcess to receive intra-task updates
* Update `pause` action to use the worker queue
* Deprecate ConnectionBase()._new_stdin
* Add new `Display` convenience method `prompt_until` to manage both controller- and worker-sourced prompting without cross-fork stdin sharing, in-worker mechanism to handle request-response over new worker queue.
This provides the same test coverage as the previous tests, without the dependency on git.
It also includes many more specific test cases with assertions, instead of simply relying on the code to not raise an exception.
Martin Krizek