Luc Grosheintz

I see the logic behind the change and agree that the INCREF looks very suspicious. IIUC a section is removed from a model, by letting the python object go out-of-scope....

Is there anything we're currently interested in that's failing only for Python 3.8? If not, I'd say continue to check until it breaks.

I think there's two (or three) abstractions: 1. An interface for writing to files/streams. This interface needs to supports injecting a callback that intercepts `stderr` and `stdout` and sends them...

I tried with the pre-release for 3.0 using the following file: ``` $ cat issue-1045.cpp #include #include int main() { xt::xarray array({2, 2}, {1, 2, 3, 4}); auto file =...

Can't be reproduced on recent versions of HighFive; and since there's no details w.r.t. versions of the involved libraries there's nothing we can do. Windows CI includes XTensor and when...

I agree with the motivation, but I'm not sure I see why this moves the `assert` close enough to where the issue occurs to make a difference. The pattern is...

> It has been added because sometime this is a double Okay, so it's not just for readability, but it's a bug fix, even better.

We have code that looks like this: ``` nmodl_eigen_j[static_cast(120)] = -nmodl_eigen_x[static_cast(0)] * _nt->_dt * kf11_ / (pow((*inst.diam[id]), 2.0) * (inst.global->beta + 1.0) * (_thread_vars.vol_ptr(id))[static_cast(0)]); nmodl_eigen_j[static_cast(132)] = _nt->_dt * kb11_ /...

@mgeplf what does one do about the coverage drop? It seems like the usual issue of reducing the total number of lines by removing covered lines, leads to a decrease...

Thank you for your reply. If we look at `setdata`: ``` static void _setdata(Prop* _prop) { _extcall_prop = _prop; _prop_id = _nrn_get_prop_id(_prop); neuron::legacy::set_globals_from_prop(_prop, _ml_real, _ml, _iml); _ppvar = _nrn_mechanism_access_dparam(_prop); }...