I am new to this list, but, as way of introduction I'm the guy inside facebook who more or less is guilty if presto releases don't show up in time -- most likely the build/test pipeline would have broke somehow, usually the credit should go when things go well, to the "release oncall" which is a larger rotation of facebook engineers.
TLDR: we need to change how we build presto and I am asking for any and all contribution to coming up with the best way to move forward,
I have a proposal to change things, not because I like change (I rather don't), but because the current situation is far from ideal and if continued in the current form things will break in the future. Even if there's no very immediate danger of things breaking, it's better to discuss things now well ahead of the breakage so that we won't have to discuss it when things break.
I intend to ask the tsc to discuss it more in depth at the next month's TSC meeting but here's the document that I started (it's early and I expect there's many thing may need to change or things that I have missed).