New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Process connection:local
after processing the magic variables
#24083
Conversation
This is needed if we expect the task `connection: local` to supersede a group/host `ansible_connection: winrm`. This fixes ansible#24076.
The test
|
Since I don't want to impact other task attributes, this commit only affects `connection` and `ansible_connection`.
The test
|
CI failure in unit tests:
|
@mattclay The PEP8 check tells me a line is too long, but it's not something I did. So once more we should not confuse contributors with stuff from other people. Unless you expect people to fix that too in their PR (which will only introduce conflicts if they all do...) |
That is resolved in devel, please rebase needs_rebase |
@sivel I know the issue itself is resolved, but it shouldn't be exposed to people who do not know what is expected from them in the first place. I am not reporting the PEP8 error, I am reporting the problematic workflow... |
Can't rebase now BTW as I can only work through the GitHub web-interface now. |
I've removed the broken command and label from this PR. The broken command for the bot is used to stop the bot from further processing on an issue when a bug in the bot is encountered. The bot is currently functioning correctly, so there's no need to disable it. |
No rebase is required for CI to pick up the fixes in devel. |
@dagwieers Determining which test failures are due to a given PR and which are unrelated is a complex problem. The current approach is to comment on In the case of the PEP 8 issue above, the comment was made by the bot, but the If you have any suggestions on how to improve this workflow, please let me know, as I'm continuing to refine the workflow to improve CI feedback to users. |
@mattclay Thanks for the feedback. I was thinking that PEP8 errors on files that are not part of the PR, should not be reported as part of the PR. Why are we testing any other files than those from the PR for PEP8 ? |
@jctanner @nitzmahone @bcoca @sivel Your feedback is appreciated :-) |
-1 this reverses precedence and breaks existing workflows, |
@bcoca So give me an example use for Because with the current behaviour, |
connection:local
after processing the magic variables
I'm fine with |
closing as per public meeting |
SUMMARY
In a heterogeneous environment where more than one connection-method is being used, one has to set
ansible_connection
for sets of hosts deliberately. In this case a play/taskconnection: local
stops working as expected because it is superseded by a group/hostansible_connection
.There is no situation where a deliberate play/task
connection: local
should be replaced with another connection method. (Think Windows and Unix systems mixed)This fixes #24076.
ISSUE TYPE
COMPONENT NAME
play_context / task_executor
ANSIBLE VERSION
v2.4