[Debian-med-packaging] Bug#1092522: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1092522

[s3v]

Hi,

I ran test_basic.py from neuron [1] and the failure output is the same:

==================================================================== test session starts ====================================================================
platform linux -- Python 3.13.1, pytest-8.3.4, pluggy-1.5.0
rootdir: /build/neuron-8.2.2
plugins: typeguard-4.4.1
collected 13 items                                                                                                                                          

test_basic.py .......F.....                                                                                                                           [100%]

========================================================================= FAILURES ==========================================================================
_____________________________________________________________ test_HocObject_no_deferred_unref ______________________________________________________________

    def test_HocObject_no_deferred_unref():
        # previous tests should really destroy the model they created
        for sec in h.allsec():
            h.delete_section(sec=sec)
    
        sl = h.SectionList()
        cell = h.Cell()
        assert len(list(cell.all)) == 2
    
        del cell
    
        # When deferred deletion was in effect, following for loop  would segfault
        # because actual deletion of the HOC Cell object would not occur til the next
        # h.functioncall(). I.e. the first call to sl.append below would destroy
        # the HOC Cell which would destroy the two Sections and invalidate the
        # iterator causing a segfault.
        for sec in h.allsec():
            print(sec)
            sl.append(sec=sec)
>       assert len([s for s in sl]) == 0
E       TypeError: Not an iterable HocObject

test_basic.py:208: TypeError

After that, I applied this commit [2] and bmtk builds fine in a sid chroot
environment against the new neuron package
I'm not entirely sure about the change since neuron doesn't run tests during
build, so there might be other bugs involved.

Kind Regards


[1] https://sources.debian.org/src/neuron/8.2.2-7/test/pynrn/
[2] https://github.com/neuronsimulator/nrn/pull/3276