5 Commits (ca635fac4a3f577f8fb5fcd33ca954aa30bdf2e5)

Author	SHA1	Message	Date
Matt Martz	9e622ddb67	[stable-2.17] Unsafe persistence (#82779) ... * Ensure that unsafe is more difficult to lose [stable-2.16] (#82293) * Ensure that unsafe is more difficult to lose * Add Task.untemplated_args, and switch assert over to use it * Don't use re in first_found, switch to using native string methods * If nested templating results in unsafe, just error, don't continue * ci_complete (cherry picked from commit 270b39f6ff) * Fix various issues in unsafe_proxy (#82326) - Use str/bytes directly instead of text_type/binary_type - Fix AnsibleUnsafeBytes.__str__ implementation - Fix AnsibleUnsafeBytes.__format__ return type - Remove invalid methods from AnsibleUnsafeBytes (casefold, format, format_map) - Use `chars` instead of `bytes` to match stdlib naming - Remove commented out code (cherry picked from commit 59aa0145d2) * Additional Unsafe fixes (#82376) * Allow older pickle protocols to pickle unsafe classes. Fixes #82356 * Address issues when iterating or getting single index from AnsibleUnsafeBytes. Fixes #82375 * clog frag (cherry picked from commit afe3fc184f) * [stable-2.16] Enable directly using `AnsibleUnsafeText` with Python `pathlib` (#82510) * Enable directly using `AnsibleUnsafeText` with Python `pathlib`. Fixes #82414 (cherry picked from commit c6a652c081) * Prevent failures due to unsafe plugin name (#82759) (cherry picked from commit 56f31126ad) * Address issues from merge conflicts --------- Co-authored-by: Matt Clay <matt@mystile.com> Co-authored-by: Martin Krizek <martin.krizek@gmail.com>	2 years ago
maulikd	9558f53a58	Fix git submodules path issue (#75655) ... * Fix git path used for submodules * Fix git submodules integration test	4 years ago
Sam Doran	18feeb51a8	Change integration tests in order to pass on Fedora 30 (#58081) ... * Use different package for DNF tests Ninja caused errors in Fedora 30. This works in both Fedora 29 and 30. * Fix git integration tests Git >= 2.21.0 has either a bug or change in behavior where it errors when fetching a repository containing submodules that are behind the upstream submodule commits. It's weird and I don't fully understand it. Get around this my checking out specific commits from a repository rather than switch the origin URL. * Fix PostgreSQL tests The error message is slightly different	7 years ago
Abhijeet Kasurde	a047fe0e4c	Correct usage for shutil.rmtree (#31541) ... * Correct usage for shutil.rmtree Fix adds correct usage of shutil.rmtree in git module Fixes: #31225 Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com> * Include archive tests so they get run * Use new include syntax * Cleanup syntax on git tests - use multi-line YAML - remove unneeded {{ }} around vars in conditionals - remove unneeded quotes - add task file name to task names for easier troubleshooting when things fail * Make archive tests work for RHEL/CentOS 6 The older versions of Jinja2 in RHEL/CentOS 6 required assertion tasks using the map filter to be skipped. The older version of git required gzip compression to be skipped on RHEL/CentOS 6. * Account for ansible_distribution_major_version missing	8 years ago
Robin Roth	cd02d0ca1d	Cleanup git tests (#19737) ... * Cleanup git tests * Split git tests in seperate files * Remove use of repo_depth_url * Use native yaml * Remove unnecessary remote/local clones * Fix newlines for yamllint * If the hash is valid (full-length) but doesn't exist, git returns 128 instead of 1. * Ensure git doesn't use hardlinks for shallow clones	9 years ago