Bump Minimum CMake to 3.13

Add SANS to LOQ sys tests filenames

Replace hardcoded index for MergeSpectra in polaris create_total_scattering_pdf

Add enum34 backport to RHEL rpm

Fix remaining memory leaks in C++ tests

Abins: fix compatibility with CRYSTAL17 molecular calculations

A temporary vector of raw pointers is constructed
instead for passing to splitBytimeVector. It is better
to have the memory handled safely than worry about the
minor cost of this allocation.

This commit adds to the lease notes that the fix for the bug that raised an error when a unicode string was given as a lim file directory has been fixed.

This commit adds to the lease notes that the match spectra in merging workspaces no longer matches by scale.

Re #27444

This commit adds release notes for the changes made in this PR.

Re #27444

This commit adds a catch for if the workspaces or q_lims file have an infinite range

Re #27444

This commit fixes missed line that would allow the index for the max range spectra
to be updated

Re #27444

This commit moves the code for reading the supplies q limits into its own function.
This also fixes a bug that cause an error when trying to read a lim file path that is in
a unicode format.

Re #27444

This commit changes the previously hard coded index for matching spectra to
instead use the workspace with the larges range of x values as these produces the
smoothest match around the edges of the workspaces data.

Re #27444

Add SANS to LOQ system test names so that -R SANS works correctly

This commit changes the usage of ConjoinSpectra to a loop using ConjoinWorkspace.
This is done because using ConjoinSpectra the focused workspace can have a
mismatch between specnumbers and index

Re #27444

Fix issue with Elemental analysis not supporting legacy data

Removes the Py3Compat shim for Enums and converts all associated callers
to use the natively provided Python version (either through a backport
or in Python 3.4 onwards)

Get pulse times as Numpy types

Make OpenMP for loops use max threads value from config

Get ConvertWANDSCDtoQ and FindPeaksMD working for HB3A

Updates the Eigen repo URL to an upstream mirror

Updates the Eigen repo to use an upstream mirror, since the offical
upstream is mid-migration and have not fully updated the Gitlab
permission to make the repo public yet

Since the tar.gz has a different name (they add -git-mirror- into the
name) the md5 must also be updated too

It costs slightly more memory but only bytes per workspace
and it is far safer than trying to manage the arrays
properly with new/delete.

Adds the Python enum backport to the RHEL-7 required packaged, so that
all code bases can use the same functions and we are Python 3 ready

Added done alongside setting the number of threads to ensure that
subsequent parallel regions do not change the number of threads they are
working with from the value set in the config.

Build errors caused by poco automatic library linking, added flag to
disable this, now building fine.

The changes to Goniometer actually fixed a bug that means more peaks are correcly being found.