Significant updates to mass transfer functionality

[ilyamandel]

These updates should really enable slow case A mass transfer:

• Equilibrium zeta and radial response of MS stars to mass loss are now calculated using CalculateRadiusOnPhase() rather than by cloning
• MassLossToFitInsideRocheLobe() and associated functor updated, work more efficiently, no longer artificially fail, and also use CalculateRadiusOnPhase()
• ROOT_ABS_TOLERANCE increased to avoid artificial failures on round-off errors
• code cleanup and bug repairs elsewhere

[github-actions[bot]]

Build Successful! You can find a link to the downloadable artifact below.

Name	Link
Commit	6c2f69f805f2a606b41c94a241cacb2c252d5ab7
Logs	https://github.com/TeamCOMPAS/COMPAS/actions/runs/9139203604
Download	https://github.com/TeamCOMPAS/COMPAS/suites/$SUITE_ID/artifacts/$ARTIFACT_ID

[ilyamandel]

Behaved well on most of my tests such as MS->MS mass transfer ./COMPAS -n 1 --initial-mass-1 2.5 --initial-mass-2 2.1 --semi-major-axis 0.05 --detailed-output --logfile-type CSV |more but misbehaving on accretion onto NS ./COMPAS -n 1 --detailed-output --evolve-pulsars True --random-seed 6944 --common-envelope-alpha 1.0 --pulsar-minimum-magnetic-field 0.1 --common-envelope-mass-accretion-prescription 'MACLEOD' --pulsar-birth-magnetic-field-distribution 'UNIFORM' --pulsar-birth-spin-period-distribution 'UNIFORM' To be investigated further.

[ilyamandel]

Limited nuclear timescale mass transfer to accrete only the smaller of the desired total MT and rate*dt on a timestep of size dt (since timescale to end of MS phase is, by construction, comparable to the nuclear MT timescale, not doing so risks transferring more mass than necessary; furthermore, it does not account for the change in the nuclear timescale as mass is lost).

[ilyamandel]

Thank you, @jeffriley . Yes, I believe this should be fine elsewhere, too.