nrn CellBuild topology page easily crashes when deleting sections with continuous create on

At the last NEURON course, some users experienced issues with deleting sections in cell builder with continuous create on. With this PR, the topology page now seems robust for all operations with continuous create.

There are likely some accumulating memory leaks during CellBuild usage and left over subsidiary objects when a CellBuild is closed. Those should be cleaned up in a separate pr.

The test_secref.py covers 97% of secref.cpp

When this PR is merged, it should be cherry picked into release/8.2

Jul 18 '24 00:07 nrnhines

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 67.35%. Comparing base (e67effd) to head (53b87f7).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3005      +/-   ##
==========================================
+ Coverage   67.26%   67.35%   +0.09%     
==========================================
  Files         573      573              
  Lines      104927   104936       +9     
==========================================
+ Hits        70576    70684     +108     
+ Misses      34351    34252      -99

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

Jul 18 '24 00:07 codecov[bot]

✔️ eb5a542d527c8a83493556d30fc053287fd514aa -> Azure artifacts URL

Jul 18 '24 00:07 azure-pipelines[bot]

The CI documentation error is

TypeError: canonicalize_version() got an unexpected keyword argument 'strip_trailing_zero'

I presume this will go away after a near future merge from master.

There is CI ubuntu error

ubuntu-22.04 - cmake (-DNRN_ENABLE_CORENEURON=ON -DNRN_ENABLE_INTERVIEWS=OFF -DNMODL_SANITIZERS=undefinedundefined)

92 - rxdmod_tests::rxd_tests (Failed)

test/rxd/test_parameter_reaction.py .                                    [ 83%]
4681
test/rxd/test_pltvar.py F                                                [ 84%]

I'm not familiar with the diagnostic process in this area. My presumption is that it is not related to this PR.

Jul 18 '24 11:07 nrnhines

✔️ 6321177125890f9e1fcedded1939d048df477f70 -> Azure artifacts URL

Aug 11 '24 23:08 azure-pipelines[bot]

Logfiles from GitLab pipeline #225633 (:no_entry:) have been uploaded here!

Status and direct links:

Aug 12 '24 00:08 bbpbuildbot

@pramodk ci/gitlab/bbpgitlab.epfl.ch — Pipeline failed on GitLab However, all the status checkmarks in bbpbuildbot above are green. Is everything in fact OK? If so, please approve this PR.

Aug 12 '24 10:08 nrnhines

I just noticed a comment by @1uc at https://github.com/neuronsimulator/nrn/pull/3037#issuecomment-2283234621 Is the gitlab failure an example of that temporary Sonata issue? edit: Ah yes. I just saw the NEURONdev slack general message.

Aug 12 '24 10:08 nrnhines

Yes, this is the same issue. The pipeline that fails is called blueconfigs and it's not covered by the checkmarks. That's likely because blueconfigs is shared CI between several projects: NEURON, NMODL and neurodamus. Therefore, I think it was never taught to "report back" correctly for each of the three projects.

Rerunning should make CI go green again.