Releases need vigerous testing - kinda wip, ReleaseDef - Update
Maybe add the alternative way? I.E. if it's changed in real resources
Teams.get_members(recursive=True) Not sure that Teams are the right thing, maybe Groups? Idk
We can maybe use run id to delete all by run id? "prevent_destroy", "ignore_changes" MORE WORK ON LIFECYCLE!
https://stackoverflow.com/questions/77522387/approving-pipeline-stage-azure-devops-via-api Auto approve via token ^
https://learn.microsoft.com/en-us/azure/devops/pipelines/process/approvals?view=azure-devops&tabs=check-pass https://stackoverflow.com/a/61519132 Pipeline perms, currently our pipelines are approval-able by almost anyone, we should be able to set the perms
Hmmmmm, it appears that when creating a variable group, it doesn't return the created by or modified by with enough data. This is different from the get_by_id, because the get_by_id gets the creator, but the creation return doesn't the data, grr...
More tests around:
Service connections perms on pipelines (done, not tested)
TestStateManager needs some work
build_defs/allow_on_environment Needs testing - Maybe better to use Build.approve_environment?
AgentPools create/delete and testing
AuditLogs Tests - Only have perms in UK, dang... New systems???
Test state_manager.py more?
Test main.py
Add all the required perms with @required_perms - On hold, perms seem to be user based, not token based
Can Jobs have logs, not just the tasks? Maybe the stage can have logs?
Docs have functions which raise NotImplemented, inspect and remove if they do?
Pat token stuffs, artifact stuffs
Somehow detect expired tokens? simple_ado??? simple_ado.exceptions.ADOHTTPException: ADO returned a non-200 status code, configuration=<simple_ado.http_client.ADOHTTPClient object at 0x1025de310>, status_code=401, text=Access Denied: The Personal Access Token used has expired. UNSURE HOW TO DO ABOVE But because we can fetch access_tokens, maybe we can instead check if the current one's expired? Somehow? I CANT REMEMBER HOW I GOT THAT ERROR, SO ANNOYING SINCE IT EXACTLY TELLS ME ITS EXPIRED
When creating yaml, if you don't provide stages, it creates one called "__default"
ado_wrapper.errors.ConfigurationError: Wrong stage name or job name combination (case sensitive), received Jobs/EchoEnvironmentVariables/Echo Modified Keys and ValuesOptions were __default/EchoEnvironmentVariables/Initialize job, __default/EchoEnvironmentVariables/Echo Modified Keys and Values
For each test file, have a zzz_cleanup() function which empties state? Maybe each test could inherit from a parent class which has this?
Retry system, if any of the requests raise ConnectionError, retry somehow? Have it call a generic _request(type) func, which has the retry logic? We can now made this better, since we have the logging function which overrides the request.
Maybe we can do some clever stuff with the StateManagedResource which trackes which functions takes how long? Might help tests, detecting calls, etc
rather than "requires initialisation", maybe make .ado_project_id a property and put it there? Perhaps have some system relating to intents? Which pre-loads a bunch of stuff (e.g, pat_author) from enum import Flag
ProjectRepositorySettings settings - MEHHH, Project creation sucks PullRequest.delete_by_id is untested. PullRequest.add_reviewer PullRequest.get_reviewers PullRequest.get_my_pull_requests PullRequest.set_my_pull_requests_included_teams? Never test get_environment_approvals
Project -> get_build_queue_settings, get_retention_policy_settings
PullRequest -> Variable Groups
Add more changelog messages for the older versions
Tests for build timelines Test utils
Work more on maintain
Rollback a commit? Tricky... I guess I could fetch the repo by a certain commit, and find the difference for all the files, and then make a commit like that?
PullRequestComment.link still doesn't work...
System messages definitely include: https://dev.azure.com/VFCloudEngineering/Platform/_git/ef514837-0d89-4da3-b156-57229447b890/pullRequest/18648#1724413175 This kind of thing ^ Check if it's fixed...
Get all runs, figure out which stage takes the longest, basically profile our runs
We do 3 code-web.branch-policies-data-provider in a row...
Build, Fetch: &maxTime={maxTime}&minTime={minTime}"
AdoUser - Get all, doesn't work with pagination, implement pagination
PullRequest.from_link()????
Commands:
python3.11 -m ado_wrapper --delete-everything --creds_file "credentials.txt"
python3.11 -m ado_wrapper --delete-everything --creds_file "credentials.txt" --state-file "tests/test_state.state"
coverage run -m pytest tests/ coverage html && open htmlcov/index.html
coverage run -m pytest && coverage html && open htmlcov/index.html
python3.11 -m pip install ado_wrapper --upgrade
python3.11 -m pytest tests/ -vvvv -s
black . --line-length 140
mypy . --strict && flake8 --ignore=E501,E126,E121,W503,W504,PBP --exclude=script.py && ruff check && pylint . bandit -c pyproject.toml -r .