shim: simulate about blobs HOT 2 OPEN

Can you elaborate a bit more about how the user is meant to submit commands while the simulation is running? Do you imagine there being multiple processes which communicate with each other somehow?

from blobs.

I envision that the user (e.g. rollup dev) would spin up the shim simulator, e.g. by running

shim simulate serve

then would run their rollup

./dn-chain

then the user would interact with the simulation by issuing the commands below.

# submit a blob
shim simulate submit blob.txt

# pause the chain
shim simulate pause

# revert the chain
shim simulate revert 1

Note those commands are fire-and-forget.

Under the hood, the command that is the part of simulate that controls simulation (maybe better described as simctl), must be able to connect to the simulation server, perform a command, and then disconnect (well, unless it's a listener of some kind). That implies that there is some kind of endpoint published by the simulation component that allows controlling the sim. I think normal HTTP/RPC should suffice for that.

I guess one take aways from the issue, is that it's probably not worth it to reuse the existing shim query subcommand for the same purpose. I mean they do share some similarities, e.g. shim simulate submit is very similar to shim query submit, but I believe they are in the end super different.

E.g. shim query revert doesn't make sense, because a normal RPC node simply doesn't provide such functionality. Also even shim simulate submit could differ from shim query submit by simulation specific flags.

from blobs.