The current state of the repo is some code snippets I managed to create recently combined into a single signal-driven hot-reloadable application-loop. The current implementation has a an atomic ring buffer with a depth of 16 tasks. It supports growing signal amount (currently 3) and currently hard codes up to 8 attached callbacks to every signal. It does all of this snuggely in 512 bytes. Still too much imo.
- consider relative pointers to cut memory requirement easily in 2 and maybe by 4
- consider demand-driven vtable allocation so that signals grow on attach for a more dynamic callback distribution
- consider 4-hastrie for enabling constant time callback detachment (in stead of current bulk detach)