Log output while Popen is in progress #45
Conversation
Currently the `execute_command` function will only log the output after the command has completed, considering that the resources being created may take a long time, getting feedback during execution would be very valuable.
There was a problem hiding this comment.
Hey, thanks for this it makes complete sense.
I dropped a couple minor comments. Can you also lint with autopep8 using 2 spaces [1], and add a simple test? [2]
[1] use a ~/.config/pep8 file with these two lines
[pep8]
indent-size = 2[2] Something like this in test_plan.py
def test_plan_stdout(fixtures_dir):
tf = tftest.TerraformTest('plan_no_resource_changes', fixtures_dir)
result = tf.plan(output=False)
assert 'just_an_output = "Hello, plan!"' in result
|
I switched to using GitHub workflows, you should be able to check lint and test errors here in the PR. Thanks again for submitting this! |
- fixed formatting - used list instead of string concat - add encoding and ignore error params to popen
|
Thanks for the changes, you should be able to merge now. Let me know if anything is preventing you from merging! |
|
Merging is blocked. It says |
Hm, as a first time contributor I think you're prevented from running checks and merging. Guess you'll have to seend another PR soon to verify that. :) Let me merge for you! |
|
Thank you for your help and providing this tool. We're using this library extensively for our terraform module testing framework. I am hoping to make improvements in the future when i can. |
Thank you for this PR, and sorry if I lagged a bit! |
|
When do you expect to release next? We are looking forward to this enhancement |
|
Good point, let me try doing it in the next few minutes. :) |
|
https://pypi.org/project/tftest/1.6.5/ Thanks again! |
Currently, the output is logged after the process has completed, this could be a problem for longer running terraform/terragrunt actions. It may take over 10 minutes to create a resource, or be locked on deleting a resources. The user would not know what's happening until the completion of the process.
I propose polling and logging the output during the process's runtime. this will provide valuable feedback to the user.