Thank you 🙇♀ for wanting to create an issue in this repository. Before you do, please ensure you are filing the issue in the right place. Issues should only be opened on if the issue relates to code in this repository.
If your issue is relevant to this repository, please include the information below:
Describe the bug
We are using exec in our arm-deploy action. Currently, we are using stdout and stdline listeners to collect the error and output. In stdline , it appears that some lines are not being received. We expect a JSON in the output, and we are getting JSON parsing errors. We confirmed this by printing whatever is received in the listener. This issue occurs persistently for large outputs, but not for small outputs. We have customers reporting this issue to us. We changed the stdline to stdout and so far while testing with it, we did not hit this issue.
We used stdline in the first place because since we expect the output to be in JSON format, our assumption is that the libraries could be emitting debug/command messages in between, which will break JSON parsing if we use stdout. So our strategy was to remove any debug/command messages from the output, and collect the actual JSON output.
Questions:
- Is the assumption correct? If no, then we can use
stdout since the issue does not seem to be happening with it.
- If the assumption is correct, will the strategy we are using(i.e. to remove the
[debug] and [command] lines) work?
- Why is the issue happening when we use
stdline listener with large outputs, when it appears to be working with stdout even with large outputs?
- If using stdline listener, some lines are not being received because of any error in sending/listening, then can we use retry anywhere? How to track sequence of the lines?
To Reproduce
Steps to reproduce the behavior:
- Use the
arm-deploy action in a workflow. Deploy a template which will create a large number of resoruces(The example i used to repro our customer issue was a template which 200 resources in a loop with a batchsize of 10 resources, so the output is pretty huge in this case)
- The action fails due to JSON parsing error
Expected behavior
- The action should succeed. It would succeed with small outputs
Screenshots
If applicable, add screenshots to help explain your problem.
Desktop (please complete the following information):
Smartphone (please complete the following information):
Additional context
Add any other context about the problem here.
Thank you 🙇♀ for wanting to create an issue in this repository. Before you do, please ensure you are filing the issue in the right place. Issues should only be opened on if the issue relates to code in this repository.
If your issue is relevant to this repository, please include the information below:
Describe the bug
We are using
execin our arm-deploy action. Currently, we are usingstdoutandstdlinelisteners to collect the error and output. Instdline, it appears that some lines are not being received. We expect a JSON in the output, and we are getting JSON parsing errors. We confirmed this by printing whatever is received in the listener. This issue occurs persistently for large outputs, but not for small outputs. We have customers reporting this issue to us. We changed thestdlinetostdoutand so far while testing with it, we did not hit this issue.We used
stdlinein the first place because since we expect the output to be in JSON format, our assumption is that the libraries could be emitting debug/command messages in between, which will break JSON parsing if we use stdout. So our strategy was to remove any debug/command messages from the output, and collect the actual JSON output.Questions:
stdoutsince the issue does not seem to be happening with it.[debug]and[command]lines) work?stdlinelistener with large outputs, when it appears to be working withstdouteven with large outputs?To Reproduce
Steps to reproduce the behavior:
arm-deployaction in a workflow. Deploy a template which will create a large number of resoruces(The example i used to repro our customer issue was a template which 200 resources in a loop with a batchsize of 10 resources, so the output is pretty huge in this case)Expected behavior
Screenshots
If applicable, add screenshots to help explain your problem.
Desktop (please complete the following information):
Smartphone (please complete the following information):
Additional context
Add any other context about the problem here.