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
Weird mixup of test data in test suite run #6058
Comments
Maybe b8e4d1c is the trouble maker? The weird data looks like a repetition |
What might be good though, is that it might not be a big problem other than the fact that this output includes irrelevant files. |
I guess this is the main root cause of all CI troubles with Windows, nested test suite processes lurking around still "using"/locking files. My changes in #6049 at least reduce the number of times a test server is "unexpectedly alive". |
Is there a way to check which process is still locking a file? Could be fun to try if |
In previous attempts to solve this I tried to use
I don't think there is a native utility on Windows anymore, previous versions had |
I don't think so. This feature is used for multiple test cases and just affects the generated output. |
Just FYI: this is still on my agenda, just didn't have the time for this yet. |
For now only required and implemented for Windows. Ref: curl#6058
Seems like the TFTP tests are the first during the run causing files to be locked:
|
For now only required and implemented for Windows. Requires Sysinternals handle[64].exe to be on PATH. Ref: curl#6058
@bagder
|
For all I can see, the mqtt server is using the same logging function as several of the other test servers and it makes Lines 133 to 136 in fa6bbbe
|
Uhm, it is the test file that is locked, not the log... |
PR pending |
@bagder it seems like the TFTP server is locking the uploaded file even after the test has finished:
|
Given a first look at the code that seems accurate. I'll fire up a PR asap. |
Thanks, I am currently trying to understand the relevant code, but I guess you are much faster than myself regarding this. |
Make sure the log file is not logged once a test has finished and align with the behaviour of our logmsg. Rename curl_test_data.py to be a general util.py. Bug: curl#6058
Found a similar issue in the Python-based tests servers and for now added a commit to PR #6179 about that. Will split that one out after testing. |
I was wrong. I have not found the reason for the locking after a completed test, |
Make sure the log file is not logged once a test has finished and align with the behaviour of our logmsg. Rename curl_test_data.py to be a general util.py. Bug: curl#6058
Make sure the log file is not logged once a test has finished and align with the behaviour of our logmsg. Rename curl_test_data.py to be a general util.py. Bug: curl#6058
Make sure the log file is not logged once a test has finished and align with the behaviour of our logmsg. Rename curl_test_data.py to be a general util.py. Bug: curl#6058
Make sure uploaded file is not locked after transfer. Bug: curl#6058
Make sure uploaded file is not locked after transfer. Bug: curl#6058
Make sure uploaded file is not locked after transfer. Bug: curl#6058
Is this issue still valid? It is simply too generic for anyone to understand what it refers to unless we have specific examples pointed out. Do we have any such? |
The examples were talked above at the top of this issue and I think the most prominent ones regarding MQTT and TFTP were fixed. #6179 will also help avoid this issue in the future. We can also already close this issue or keep it open until the PR is ready and merged, which I still plan to do soon once I have the Unix variant implemented. |
Issues work best when they identify an issue and thus also indirectly when the issue is fixed (when the issue/bad thing no longer occurs). The way this issue is phrased it is more of a placeholder for you as nobody else knows what it is about nor when it is done (since the only explicitly mentioned ones are fixed already). As such, I don't think it serves much purposes to be left open. |
I did this
While working on #6049 this weird test suite run occurred.
For some unknown reason the tests 2055 3002 3003 3004 3005 3006 3007 failed while showing irrelevant test data.
I expected the following
No irrelevant test data, like MQTT related things, for these test cases.
cc @bagder any idea?
The text was updated successfully, but these errors were encountered: