multiple viable caches from parent commits

Stream: general

[Filippo A. E. Nuccio]

OK, this is something that I would like to understand better: namely, how are commits relative to mathlib treated on branches? What happens to me is that I have a mathlib folder, and if I am on branch master everything is nice (and up to date). When I checkout fae_H_space and try a leanproject get-cache I get the multiple viable caches from parent commits and then, even if downloading all of them with --fallback download-all, VSCode still takes ages. So I would like some help both in making this work again and, most importantly, in understanding "how" my local set-up knows what mathlib a certain branch is depending on (and why sometimes it thinks it depends on several ones, which is how I interpret the error multiple viable caches from parent commits).

[UPDATE] After 15 minutes or so, I did a new leanproject get-cache and got no multiple viable caches from parent commits, nor any compiling issue. Yet, I would like to understand the overall picture better.

[Eric Wieser]

Every commit that gets built by CI gets a cache attached to it; the caching infrastructure doesn't know anything about master being more special than any other branch

[Filippo A. E. Nuccio]

Ah good, this means that my branch gets compiled and when I do get-cache I get the oleans from that compilation on the branch?

[Filippo A. E. Nuccio]

OK, but then what does this multiple viable caches from parent commits mean?