Modify portabilities for new Ubuntus

[kwabenantim]

Related Issue

• Create GitHub runner docker images for newer Ubuntus #197

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.99%. Comparing base (e26ac25) to head (e172aaf).
Report is 31 commits behind head on develop.

❗ Current head e172aaf differs from pull request most recent head 6c28b3f

Please upload reports for the commit 6c28b3f to get more accurate results.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #205      +/-   ##
===========================================
- Coverage   100.00%   99.99%   -0.01%     
===========================================
  Files         1029     1029              
  Lines        51386    51377       -9     
===========================================
- Hits         51386    51375      -11     
- Misses           0        2       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fcooper8472]

Following our discussion this morning, I have opened #231 which outlines a strategy for temporarily ignoring (in a very selective and targeted way) some -Wmaybe-uninitialized c_vector warnings.

The discussion indicated a preference to temporarily ignore the warnings, with a view to migrating from Boost uBLAS to Eigen in due course, as part of Milestone 2.2.

[fcooper8472]

I'm stuck on something that appears to be a genuine compiler warning; perhaps something for @jmpf or @mirams:

The following line calls ComputeLinearInUCoeffInSourceTerm with a second parameter as nullptr.

Chaste/pde/src/problem/AbstractLinearEllipticPde.hpp

Line 147 in 7bf9a4d

return ComputeLinearInUCoeffInSourceTerm(rNode.GetPoint(), nullptr);

In AbstractLinearEllipticPde, the function ComputeLinearInUCoeffInSourceTerm is pure virtual, but one place it's defined is in:

Chaste/cell_based/src/cell_based_pde/pdes/AveragedSourceEllipticPde.cpp

Lines 113 to 118 in 7bf9a4d

	template<unsigned DIM>
	double AveragedSourceEllipticPde<DIM>::ComputeLinearInUCoeffInSourceTerm(const ChastePoint<DIM>& rX, Element<DIM,DIM>* pElement) // now takes in element
	{
	assert(!mCellDensityOnCoarseElements.empty());
	return mSourceCoefficient * mCellDensityOnCoarseElements[pElement->GetIndex()];
	}

What should this specialisation do in the case that the element is null?

[jmpf]

What should this specialisation do in the case that the element is null?

What happens if we assert that it's not null? Does any test-suite fall over?