Add GasmanDisableMemoryLimits

[ChrisJefferson]

This gives a way in GAP code of disabling GAP entering the break loop when memory is full, or stopping altogether if the upper memory limit is reached.

This might seem weird, but I find it useful for two reasons.

1. When writing code to be run automatically it stops me having to figure out upper limits to give.

2. It is useful in Windows, where it is hard to add command line parameters. I have had this requested by at least 3 windows users over time, particularly the "don't enter the break loop" part.

Of course, GAP will still eventually run out of memory, but most other languages just allocate memory until the OS refuses.

[codecov[bot]]

Codecov Report

Merging #4217 (13f3e0f) into master (1a9d58b) will decrease coverage by 0.02%. The diff coverage is 33.33%.

@@            Coverage Diff             @@
##           master    #4217      +/-   ##
==========================================
- Coverage   93.40%   93.38%   -0.03%     
==========================================
  Files         717      717              
  Lines      812469   818935    +6466     
==========================================
+ Hits       758915   764729    +5814     
- Misses      53554    54206     +652     

Impacted Files	Coverage Δ
src/gap.c	75.86% <0.00%> (-0.31%)	:arrow_down:
src/sysmem.c	53.11% <ø> (+0.73%)	:arrow_up:
src/system.c	58.67% <ø> (+4.62%)	:arrow_up:
lib/system.g	75.97% <100.00%> (ø)
lib/autsr.gi	50.00% <0.00%> (-20.06%)	:arrow_down:
lib/gpfpiso.gi	70.66% <0.00%> (-5.09%)	:arrow_down:
lib/twocohom.gi	91.61% <0.00%> (-3.24%)	:arrow_down:
lib/csetgrp.gi	70.51% <0.00%> (-1.92%)	:arrow_down:
hpcgap/lib/hpc/stdtasks.g	70.15% <0.00%> (-0.22%)	:arrow_down:
lib/kbsemi.gi	98.00% <0.00%> (-0.11%)	:arrow_down:
... and 106 more

[fingolfin]

@ChrisJefferson any plans to revise this to be ready for merging?

[ChrisJefferson]

will do, once it gets up my list to the top again :)

[fingolfin]

@ChrisJefferson this has been sitting without update for 3 years. Any plans to finish it? Or should we just close it for now (you can always resubmit in the future)