#Technical question for you @Erik Sipsma
vocal totem
I'm guessing that you are running this comparison where each run has a totally empty engine cache, right? just to eliminate the possibility that it's like re-using layers or something even if it's pulling some
uneven ivy
I don't know if t's totally empty but it's right after doing ./hack/dev
And it happened several time so I would assume each image is re-using layers if they exist
vocal totem
That is quite interesting though, the biggest difference I can see is that the alpine image has likely more layers? at least it requires more http requests to pull (could also be multiplatform manifest indirection)
vocal totem
I don't know if t's totally empty but it's right after doing ./hack/dev
And it ...
to make sure this is real, I'd do docker rm -fv dagger-engine.dev; docker volume rm dagger-engine.dev; ./hack/dev to start with a fresh cache
uneven ivy
Hmm maybe it's because I pull bun after pulling alpine (in another function)
and if bun depends on alpine it may reuse layers 😮
vocal totem
Hmm maybe it's because I pull bun after pulling alpine (in another function)
and...
definitely possible too, good point
