Codecov Report

Merging #2037 (9a4f017) into master (fac194c) will increase coverage by 0.28%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##           master    #2037      +/-   ##
==========================================
+ Coverage   48.96%   49.24%   +0.28%     
==========================================
  Files         520      520              
  Lines      115736   115736              
==========================================
+ Hits        56667    56994     +327     
+ Misses      59069    58742     -327     

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

✔️ e27da32 -> Azure artifacts URL

✔️ dd10bda -> Azure artifacts URL

✔️ 4c83468 -> Azure artifacts URL

✔️ ce25ac7 -> Azure artifacts URL

✔️ 064ab21 -> Azure artifacts URL

✔️ 1e7f326 -> Azure artifacts URL

@alkino 9602b9b adds a FUNCTION_TABLE test to this PR.

✔️ 9602b9b -> Azure artifacts URL

Maybe we can add example for a constant function table, and one with only 1 arg, and 2 runtime hoc vectors?

Maybe we can add example for a constant function table, and one with only 1 arg, and 2 runtime hoc vectors?

I'll do that.

Logfiles from GitLab pipeline #82951 (:white_check_mark:) have been uploaded here!

Status and direct links:

• ✅ clean_workspace
• ✅ set_homebrew_variables
• ✅ install_python_dependencies
• ✅ patches
• ✅ cmake_build: [cmake, ON, OFF, OFF, address]

Logfiles from GitLab pipeline #82962 (:white_check_mark:) have been uploaded here!

Status and direct links:

• ✅ clean_workspace
• ✅ set_homebrew_variables
• ✅ install_python_dependencies
• ✅ patches
• ✅ cmake_build: [cmake, ON, OFF, OFF, address]

✔️ 4a66d11 -> Azure artifacts URL

Logfiles from GitLab pipeline #83385 (:no_entry:) have been uploaded here!

Status and direct links:

• ⛔ cmake_build: [cmake, ON, OFF, OFF, address]

✔️ a442c95 -> Azure artifacts URL

Logfiles from GitLab pipeline #83588 (:no_entry:) have been uploaded here!

Status and direct links:

• ⛔ cmake_build: [cmake, ON, OFF, OFF, address]

✔️ 9bfd8b0 -> Azure artifacts URL

Logfiles from GitLab pipeline #84315 (:white_check_mark:) have been uploaded here!

Status and direct links:

• ✅ cmake_build: [cmake, ON, OFF, OFF, address]

✔️ 9a4f017 -> Azure artifacts URL

I just have one suggestion. Maybe the expsynspine.ses could be turned into a python script so it's easier to convert it to a proper test tested using pytest?

I can do that in the sense of replacing the last line in expysynspine.py with

h(r'''
// the contents of expsynspine.ses
''')

but I'm not sure that's what you mean. I.e is it beneficial to avoid extra files read by a test executed by pytest?
Also, until there is a way for pytest to execute a a bunch of test_foo.py files one at a time with separate invocations of a clean from neuron import h the test authors need to do extra work in cleaning up after each test, and if HOC variables are used, the HOC namespace gets more and more polluted. For now it seemed easiest to just run each mod test with a separate invocation of nrniv.

I just have one suggestion. Maybe the expsynspine.ses could be turned into a python script so it's easier to convert it to a proper test tested using pytest?

I can do that in the sense of replacing the last line in expysynspine.py with

h(r'''
// the contents of expsynspine.ses
''')

but I'm not sure that's what you mean. I.e is it beneficial to avoid extra files read by a test executed by pytest? Also, until there is a way for pytest to execute a a bunch of test_foo.py files one at a time with separate invocations of a clean from neuron import h the test authors need to do extra work in cleaning up after each test, and if HOC variables are used, the HOC namespace gets more and more polluted. For now it seemed easiest to just run each mod test with a separate invocation of nrniv.

I agree with what you said and I came across that before. What I meant was instead of the .ses file to run the same thing in python so that we can add asserts of certain values to check the the test is running correctly. If I understood correctly with just the .ses we can only check the the simulation is running without error, right?

I agree with what you said and I came across that before. What I meant was instead of the .ses file to run the same thing in python so that we can add asserts of certain values to check the the test is running correctly. If I understood correctly with just the .ses we can only check the the simulation is running without error, right?

Ok. I don't think the ses file hurts in this case as it gives a visual indication of how things look. It will be no problem for
expsynspine.py to subsequently add a test_run() (although pytest is not used for any of the tests in this folder) to do its own run with a few judiciously chosen Vector.record and compare the SHA hash of those to a "standard".

Ok. I don't think the ses file hurts in this case as it gives a visual indication of how things look.

Yes, I agree.

(although pytest is not used for any of the tests in this folder)

I had the same impretion however I see that every python file in this directory gets tested when we run ctest and the NRN_ENABLE_TESTS=ON. For example check the output of ctest -R example_nmodl

I had the same impretion however I see that every python file in this directory gets tested when we run ctest and the NRN_ENABLE_TESTS=ON. For example check the output of ctest -R example_nmodl

Correct, but they are not run with pytest. Each one is executed as python foo.pyand the exit value is checked.