Heya all, Has anybody, since the last thread quite a while ago, investigated thread safety for concurrency in Soft?
Of course if I completely remove any data from Soft, work in isolation, and then brutally dump it back in afterwards I expect things to work, but that's seldom ideal, and even then I've found it's somehow possible to mess things enough that the main thread times out fatally (entirely my fault, but it was kind of scary to find out I could :p). So, anybody figured out if there actually is anything that is even remotely thread safe, or things that are particularly susceptible to compromising the host? This is outside of ICE, and it's prominently update in ops on kine and meshes I'm interested in. Mr Chia, if you happen to read this: are there any plans in those regards you could share? At this point even something really crappy ala old Maya OpenMP to just ensure thread safety reading mesh data would be a pretty big step forward (I'd consider a single wake-up from lazy call as single threaded before pulling as an acceptable compromise to be able to ref read afterwards) -- Our users will know fear and cower before our software! Ship it! Ship it and let them flee like the dogs they are!
