This only really impacts pip installs of Salt and the windows onedir
since the linux and macos onedir build every package dependency from
source, not from pre-existing wheels.
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* fixessaltstack/salt#62372 unable to use random shuffle and sample functions as Jinja filters
* move random_shuffle and random_sample logic to utils
* static seed in tests seems to have shifted
* static seed in tests require hash module
* Change Tiamat to onedir in release notes
* Reinstate known issues
* Update release notes with onedir package support policy
* need to check the version of Netmiko python library and then import the exceptions from different locations depending on the result.
* Adding changelog.
* swap out if...else for double try...except.
* Remove extra fix we don't need anymore
* [Docs] include onedir system python note
* Update all platforms to use pycparser 2.21 or greater for Py 3.9 or higher, fixes fips fault with openssl v3.x
* Remove the PyObjC dependency
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Add "<tiamat> python" subcommand to allow execution or arbitrary scripts via bundled Python runtime
* Document usage of bundled Python runtime for Client API
* Use explicit locals for custom script execution, handle exception in similar fashion as Python
* Remove old __file__ replacement
* Apply suggestions from code review
Co-authored-by: Pedro Algarvio <pedro@algarvio.me>
Co-authored-by: nicholasmhughes <nicholasmhughes@gmail.com>
Co-authored-by: Alyssa Rock <alyssa.rock@gmail.com>
Co-authored-by: Gareth J. Greenaway <gareth@saltstack.com>
Co-authored-by: Twangboy <leesh@vmware.com>
Co-authored-by: David Murphy < dmurphy@saltstack.com>
Co-authored-by: Pedro Algarvio <palgarvio@vmware.com>
Co-authored-by: Lukas Raska <lukas@raska.me>
Co-authored-by: Pedro Algarvio <pedro@algarvio.me>
* Check only ssh-rsa encyption for set_known_host
* Windows test fix
* Fix pre-commit
* add CentOS Stream to _OS_FAMILY_MAP, fix#59161
* added changelog and test
* fix syntax
* Use centosstream 8 for testing
* Use ? for matching spaces
Technically this isn't *quite* right as 'CentOSyStream' would also
match, but it's pretty reasonable:
- OS grains shouldn't ever be that kind of close
- This test is only swapping out spaces, and only for the os grain. That
would mean there would have to be two OSes with grains that only
differ by one having a space where another one has any other
character.
- This test really isn't even about matching grains, we're just using
compound matching and that's a reasonable one to use.
* Add centos stream when detecting package manager name
* Fix pre-commit
* Remove tests for fedora 32/33 EOL
* Remove tests for fedora 32/33 EOL
* Remove tests for fedora 33 EOL
* Use centosstream 8 for testing
* Use ? for matching spaces
Technically this isn't *quite* right as 'CentOSyStream' would also
match, but it's pretty reasonable:
- OS grains shouldn't ever be that kind of close
- This test is only swapping out spaces, and only for the os grain. That
would mean there would have to be two OSes with grains that only
differ by one having a space where another one has any other
character.
- This test really isn't even about matching grains, we're just using
compound matching and that's a reasonable one to use.
* 3002.9: Fix pre-commit
* 3003.5 Fix pre-commit
* [3002.9] Replace use of 'sl' with 'paper' for Arch tests, due to 'sl' having key issues
* Remove mojave testing
* Remove mojave and high sierra testing
* Remove mojave testing
* [3002.9] Fix cloud vultr size issue
* Update package name to aspnetcore-runtime-6.0 for redhat 8 pkg tests
* Update package name to aspnetcore-runtime-6.0 for redhat 8 pkg tests
* change amazon linux AMI
* Migrate `unit.modules.test_gpg` to PyTest
* Don't leave any `gpg-agent`'s running behind
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Start a background process to generate entropy.
Some tests have failed because of not enough entropy which then makes
the test timeout.
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* A different approach at generating entropy
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Turn entropy generation into a helper
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* change amazon linux AMI
* change amazon linux AMI
* [3004.2] Fix cloud vultr size issue
* Fix cloud requirements
* Skip pam tests on windows
* Update ami to try to get the tests running
* Update amis to try to get the tests running
* Fixing test_publish_to_pubserv_ipc_tcp, moving the call to socket.socket into the while loop.
* Add static requirements for 3.8 and 3.9 on Windows
* Fix requirements
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* The whole CI process is already slower than GH Actions, no caches.
* Pre-commit must not run with ``PIP_EXTRA_INDEX_URL`` set.
* Lint fixes
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Compile cloud requirements
* Run add requirements files for 3.8 and 3.9
* Fix docs and cloud requirements
* [3003.5] Fix cloud vultr size issue
* Windows test fix
* Skip test if docker not running
* [3003.5] Fix pre-commit
* Update Markup and contextfunction imports for jinja versions >=3.1.
* update bootstrap to 2022.03.15
* update bootstrap to 2022.03.15
* skipping tests/pytests/integration/modules/test_virt.py on 3002.x and 3003.x branches.
* Windows test fix
* Skip PAM tests on Windows
Windows has no ctypes with the PAM bits, so we should go ahead and skip
on Windows.
* Skip PAM auth tests on Windows
Windows lacks the correct bits, so...
* Fix pre-commit
* Skipping tests since they're also skipped on the master branch
Fixes#403
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Skip test that only runs because the patch binary is now available.
The feature though, was only added in 3004.
Fixes#404
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Skip test which is only supposed to run in Linux
Fixes https://github.com/saltstack/salt-priv/issues/405
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* GPG tests do not work on windows yet
* Fix tests
* Fix pre-commit
* skip tests.integration.modules.test_mac_brew_pkg.BrewModuleTest.test_list_upgrades and tests.integration.modules.test_state.StateModuleTest.test_get_file_from_env_in_top_match on Mac OS.
* skip tests.integration.modules.test_mac_brew_pkg.BrewModuleTest.test_list_upgrades and tests.integration.modules.test_state.StateModuleTest.test_get_file_from_env_in_top_match on Mac OS.
* Removing skip, moving it to different PR.
* Skipping tests on 3002.9.
* test fix
* Do not run patch tests on 3003.5. Feature not added till 3004
* skipping tests/pytests/integration/modules/test_virt.py on 3002.x and 3003.x branches.
* Fix pre-commit
* [3004.2] Update freebsd ami
* Bump the git version for freebsd CI tests
* removing versions that are no longer available from the tests.pytests.scenarios.compat.test_with_versions tests.
* Skip tests on windows when NOT using static requirements
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* removing versions that are no longer available from the tests.pytests.scenarios.compat.test_with_versions tests.
* test_issue_36469_tcp causes a fatal python error when run on Mac OS, so skipping.
* Fix tests
* Fix pre-commit
* Do not run patch tests on 3003.5. Feature not added till 3004
* Skip archive tar tests on windows
* [3002.9] Skip archive tar tests on windows
* GPG tests do not work on windows yet
* Skip test which is only supposed to run in Linux
Fixes https://github.com/saltstack/salt-priv/issues/405
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Skip test that only runs because the patch binary is now available.
The feature though, was only added in 3004.
Fixes#404
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Skipping tests since they're also skipped on the master branch
Fixes#403
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Fix pre-commit
* Fix pre-commit
* Fix pre-commit
* Fix pre-commit
* retry sdb.get if it returns None
None is an entirely valid return - see EtcdClient.get in
salt/utils/etcd_util.py
* drop py2/six
* fix etcd sdb.set as well
* Fix etdcd-sdb test failure
If docker container is up and running, but etcd isn't responding yet
it's possible that we get some failing tests. This should wait a
reasonable amount of time for things to come up. Or just skip the test.
* Fix etdcd-sdb test failure
If docker container is up and running, but etcd isn't responding yet
it's possible that we get some failing tests. This should wait a
reasonable amount of time for things to come up. Or just skip the test.
* Skip the tests from unit/transport/test_zero.py that are hanging on Mac.
* skip tests in tests/pytests/unit/states/test_archive.py for 3002.9
* 3002.9 Skipping CA permissions tests on Windows, similar to 3003.5 and 3004.2
* change skipif to skip
* Rollback Windows AMIs to use Python 3.7
* Rollback AMI's to Python 3.7... fix tests
* Fix failing test_archive tests
* Build using pyenv
* Add symlinks to openssl and rpath
* Add shasum for zeromq 4.3.4
* Fix docs on scripts
* Build zeromq earlier, fix symlinks
* Bring 61446 to 3004.1 branch
* Add changelog and tests
* Fix schedule test flakiness
* Retry with new port if in use
* fixing failing tests, ensuring that the correct path is used.
* fixing failing tests, ensuring that the correct path is used.
* fixing failing tests, ensuring that the correct path is used.
* Re-enable tiamat-pip on windows
Signed-off-by: Pedro Algarvio <palgarvio@vmware.com>
* Bump duration time for windwos for test_retry_option_success
* Skip test cauing hangs
* go go pylint disable
* more pre-commit
* oh lint
* so many weird hook failures
* Add unit tests for PAM auth CVE
We could add functional tests if it's important enough, but this is the
narrowest place to test.
* Fix PAM auth CVE
Credit to @ysf
Previously we weren't checking the result of PAM_ACCT_MGMT.
* pylint disable
* rewrite hook changes
* Skip PAM auth tests on Windows
Since Windows ends out lacking the correct bits, no need to run tests
there.
* pre-commit fixes
* docs 3004.2 release
* Fix bug in tcp transport
* Fix the test_zeromq_filtering test
* skip test_npm_install_url_referenced_package on centos 7 and 8.
* Swapping CentOS Linux-8 for CentOS Stream-8
* Update build scripts to use pyenv
* Fix tests on MacOS
* Fix bug in tcp transport
* Fix test failures
* Update release notes and man pages for 3003.5
* Add 3002.9 changelog, release notes, man pages
* Update doc/topics/releases/3002.9.rst
Co-authored-by: Megan Wilhite <mwilhite@vmware.com>
* Fix requirements
* Fix imports
* Test fixup
* Fix merge warts
* fix merge wart in changelog
* Fix merge warts in tests
Co-authored-by: krionbsd <krion@FreeBSD.org>
Co-authored-by: Megan Wilhite <megan.wilhite@gmail.com>
Co-authored-by: Alexander Kriventsov <akriventsov@nic.ru>
Co-authored-by: Megan Wilhite <mwilhite@vmware.com>
Co-authored-by: Wayne Werner <wwerner@vmware.com>
Co-authored-by: Gareth J. Greenaway <gareth@saltstack.com>
Co-authored-by: David Murphy < dmurphy@saltstack.com>
Co-authored-by: Twangboy <leesh@vmware.com>
Co-authored-by: MKLeb <calebb@vmware.com>
Co-authored-by: Pedro Algarvio <pedro@algarvio.me>
Co-authored-by: Pedro Algarvio <palgarvio@vmware.com>
Co-authored-by: Thomas Phipps <tphipps@vmware.com>
Co-authored-by: Frode Gundersen <frogunder@gmail.com>
Co-authored-by: Alyssa Rock <alyssa.rock@gmail.com>
Co-authored-by: Alyssa Rock <43180546+barbaricyawps@users.noreply.github.com>
CVE-2020-27783
moderate severity
Vulnerable versions: < 4.6.2
Patched version: 4.6.2
A XSS vulnerability was discovered in python-lxml's clean module. The module's parser didn't properly imitate browsers, which caused different behaviors between the sanitizer and the user's page. A remote attacker could exploit this flaw to run arbitrary HTML/JS code.
