refactor: Add mem module to allow tests to override allocators.

[iphydf]

This will allow us to do more interesting things with memory allocation within toxcore, and allow fuzzers to explore various allocation failure paths.

---

This change is 

[codecov]

Codecov Report

Patch coverage: 89.54% and project coverage change: +64.49% 🎉

Comparison is base (6133fb1) 9.99% compared to head (a549807) 74.49%.

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #2393       +/-   ##
===========================================
+ Coverage    9.99%   74.49%   +64.49%     
===========================================
  Files         141       87       -54     
  Lines       32530    26194     -6336     
===========================================
+ Hits         3250    19512    +16262     
+ Misses      29280     6682    -22598     

Files Changed	Coverage Δ
toxcore/group_moderation.c	80.69% <50.00%> (+72.52%)	⬆️
toxcore/onion_announce.c	85.61% <62.50%> (+81.82%)	⬆️
toxcore/TCP_common.c	78.72% <75.00%> (+72.34%)	⬆️
toxcore/tox_private.c	49.27% <75.00%> (+36.04%)	⬆️
toxcore/mono_time.c	94.33% <77.77%> (+73.58%)	⬆️
toxcore/tox.c	63.69% <80.95%> (+55.89%)	⬆️
toxcore/tox_events.c	78.70% <81.81%> (+27.07%)	⬆️
toxcore/TCP_server.c	78.77% <83.33%> (+70.65%)	⬆️
toxcore/TCP_connection.c	70.86% <84.61%> (+60.48%)	⬆️
toxcore/DHT.c	84.50% <87.17%> (+74.73%)	⬆️
... and 21 more

... and 112 files with indirect coverage changes

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

[nurupo]

A bit confused by this PR and #2397 both introducing the mem module and using new system_{random,network,memory}. Is one PR build on top of another? Or perhaps these are alternative implementations of the same thing?

[Green-Sky]

@nurupo this one introduces the mem, #2397 should just be moving stuff around (afaik)