-
Notifications
You must be signed in to change notification settings - Fork 117
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
Testing new way to specify azure tests #20060
Conversation
I like this idea. One thing to check with the trigger: We need to make sure it triggers if the contents of symlink change. The symlink itself is a version controlled entity, so I am not 100% sure changes to the destination will be seen. |
Doh! OMG...you are right. ChatGPT says the |
Ok, so I did get a pass with this approach in baee914. This approach makes it easier to add tests to Azure CI without having to change the I am trying a different approach now. |
…sit-dav/visit into task-mcm86-19nov24-expand-azure-tests
@@ -1,5 +1,5 @@ | |||
# ---------------------------------------------------------------------------- | |||
# CLASSES: nightly | |||
# CLASSES: nightly, prci |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Your first condition for PRCI tests isn't true for blueprint: The test typically takes less than 20 seconds to run "normally".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The condition is about adding new tests...not about anything preexisting this work.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I see
I feel like we could use some dev docs explaining how to go about adding new tests to the CI. If you haven't seen this PR, the process is mysterious. |
@markcmiller86 this is growing a bit complex. I think simply triggering CI if any py file changes is not the end of the world. Don't really like the idea of extra commits from the robot to updated the checksum. (those will be in our history) Also, worried about the timing interaction between github actions and azure. I really don't like the fact that the CI could trigger on a later, unrelated PR. |
How is this more complex than checking the BLT hash as part of something called the "BLT Time Travel Check"?
Triggering CI is not really what this PR is about. I mean, its certainly part of it but a substantially smaller part than...
But those are apparently ok from LC cron jobs during from nightly testing
Nothing to worry about. As explained in
Either I don't understand this statement or the statement is false. |
you also have to add a symlink for it too
Either the list is there or it is in this new tests/azure folder. I think I prefer it in
I don't know that we want even more robo commits. I agree with Cyrus that this all feels complicated. The barrier to entry to understanding how all this works is higher than before. |
Thats not true. Whoops...forgot to remove that stuff. Not needed anymore in this PR |
@@ -158,8 +126,8 @@ stages: | |||
# add to ld_lib path (rpaths are missing?) | |||
export LD_LIBRARY_PATH=${VTK_LIB_DIR}:${OSPRAY_LIB_DIR}:${QT_LIB_DIR} | |||
# run test suite on silo + blueprint tests | |||
export TESTS="$(python_test_files)" | |||
./run_visit_test_suite.sh --fuzzy --pixdiff 10 --avgdiff 10 --numdiff 0.1 -v ${TESTS} | |||
export TESTS="$(find ../../src/test/tests -name '*.py' -exec grep -H -m 1 CLASSES {} \; | grep prci | cut -d':' -f1)" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am not familiar enough with all of these commands to know exactly what is happening... if someone names a variable in a python test prci
will the test get unwittingly triggered in CI?
That seems to fly in the face of "more automation" to make our lives easier. Something I've also heard discussed a lot in our team recently. I don't know what to do about vague statements such as "this feels complex". I can respond to specific issues for which a technical argument has been offered. Ok, I see another email...feels like people are just looking for reasons to decline this. I will close. |
It looks like this could create a commit for every push to a PR, which is not the same as the test suite results (max of one commit per day to the develop branch). Are those pushed to the current branch? Will those re-trigger the CI? I don't understand it and can't predict how the multiple CI jobs and commits will interact. Could we still have the symlink folder to organize the ones we want to run on azure and then trigger if any python file in all of the test suite has changed? Or just run our own script to see if any of those files were changed at the start of the main CI for an easy + fast exit with a pass. |
Description
This PR adjusts how Azure CI testing is handled.
.py
file insrc/test/tests
to Azure CI, just add the stringprci
to the comma separatedCLASSES
list in the.py
file header comments. Be sure to follow requirements specified insrc/test/prci_checksum.txt
though..py
files and holds it insrc/test/prci_checksum.txt
src/test/prci_checksum.txt
has changed.Type of change
How Has This Been Tested?
Reminders:
Checklist: