* rhsm_release: Fix the issue that rhsm_release module considers 8, 7Client and 7Workstation as invalid releases.
* Fix the unit test error: The new release_matcher could pass a wider range of patterns but that would not cause extra issue to the whole module.
* Submit the changelog fragment.
* Update changelogs/fragments/2571-rhsm_release-fix-release_matcher.yaml
Co-authored-by: Amin Vakil <info@aminvakil.com>
Co-authored-by: Amin Vakil <info@aminvakil.com>
(cherry picked from commit 593d622438)
Co-authored-by: Tong He <68936428+unnecessary-username@users.noreply.github.com>
* copying from the previous branch
* passing sanity - docs incomplete
* adjusted parameter name
* adjusted unit tests for mode=new
* adjusted integration tests for mode=new
* added 'russoz' to list of maintainers for cpanm
* Update tests/integration/targets/cpanm/tasks/main.yml
* Update tests/integration/targets/cpanm/tasks/main.yml
* ensuring backward compatibility + tests
* added changelog fragment
* version for new parameter and adjusted example
* typo and formatting
* Update plugins/modules/packaging/language/cpanm.py
Co-authored-by: Felix Fontein <felix@fontein.de>
* Update plugins/modules/packaging/language/cpanm.py
Co-authored-by: Felix Fontein <felix@fontein.de>
* Update plugins/modules/packaging/language/cpanm.py
Co-authored-by: Felix Fontein <felix@fontein.de>
* multiple changes
- some fixes from PR
- supporting tests
- integration is no longer unsupported => destructive, should run on
apt- and rpm-based systems only
* only run integration tests in redhat-family > v7 or debian-family
Co-authored-by: Felix Fontein <felix@fontein.de>
* Provide a user friendly message by handling json decode
exception rather than providing a stacktrace
Fixes: #1614
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* Fix typo in redhat_subscription testcase
* Fix macports state=present matching against the wrong package name
Previous implementation returned true if the desired package name occurred anywhere in
the list of all installed packages. For example as a substring of another package name,
or even as a substring of a variant name for a different package.
Instead:
- request macports only list installed packages matching the desired package name, instead of all installed packages.
- Note `port` exits with 0 regardless of whether any packages match the requested name.
- pass `-q` flag: "Do not print the header line. This is useful when parsing the output of port installed in scripts for further processing."
- eliminate `use_unsafe_shell=True` by searching stdout contents natively in python
instead of using `grep`. This has the added benefit of eliminating any potential
misinterpretation of characters in the package name as regex special characters.
If there are zero matching installed packages, `out` is empty.
If there are one or more matches (due to multiple installed versions), the output format
is: ec8a2bc682/src/port/port.tcl (L3320-L3323)
Notably, two leading spaces, the package name, a space, and then other information.
According to blame via github, those lines haven't changed in 11 years.
* Update macports state=active to eliminate use_unsafe_shell
Similar to previous commit (for macports state=present):
- pass `-q` flag: "Do not print the header line. This is useful when parsing the output of port installed in scripts for further processing."
- search stdout contents natively in python instead of using `grep`.
- added parentheses to search string to eliminate false positives if the package name
or variants contain the word `active`. Still could fail if they contain `(active)`, but
that's less likely
If there are zero matching installed packages, `out` is empty.
If there are one or more matches (due to multiple installed versions), the output format
is: ec8a2bc682/src/port/port.tcl (L3320-L3323)
For "state=active", we're looking for a line that contains `(active)` in the output.
* Basic test case of query_port for present and active
* Attempt to fix lint errors in test
* Different mock module creation, changed test cases indentation/spacing
- picked the wrong mock code to cargo-cult. Thanks to felixfontein for this suggestion
- 4 space indentation on continuation line. I thought I had that originally, but it looks like my editor sabotaged me with mixed tabs/spaces
- Remove leading newline on multi-line test cases. I don't think it would make a difference, but I'd read up on how the python syntax works and want to more accurately represent macports output.
fingers crossed this addresses the known build errors
* Add changelog fragment
* Update tests/unit/plugins/modules/packaging/os/test_macports.py
Co-authored-by: Felix Fontein <felix@fontein.de>
* Update changelogs/fragments/1307-macports-fix-status-check.yml
Co-authored-by: Felix Fontein <felix@fontein.de>
Co-authored-by: Felix Fontein <felix@fontein.de>
* * pkgin query_package(..) understands now also package name with version (my-package-1.2nb123456). * pkgin query_package(..) will distinct between not-installed and not-found packages. * pkgin install_package(..) fails with proper error if a non-existing package is attempted to be installed.
* fixup ansibot comments
* add changelog fragment
* add example
* use more pythonic condition
Co-authored-by: Andrew Klychkov <aaklychkov@mail.ru>
* remove extra dot between description and link to PR
Co-authored-by: Andrew Klychkov <aaklychkov@mail.ru>
* fix file extension of changelog fragment
* add pkgin unit tests for query_package function
* fix pep8 issues
* use enum Flag instead mix of strings, bools and None as return value
* use IntEnum instead Flag to support Python versions >= 2.6
* fix pep8 and pylint errors
* use regular class instead enum to avoid issues with older Python releases
* fix comment
* remove the combined package state since without an IntFlag or Flag the usage of it would require to use logical operators
Co-authored-by: Erik van Nooijen <eric.vannooijen@tomtom.com>
Co-authored-by: Andrew Klychkov <aaklychkov@mail.ru>
