Create a script to validate the binary is runnable

[jennijuju]

Now pre-built binaries are produced in release assets (including rcs), we want to create a script to make sure it actually works. test can cover:
(add a self-check commend)

• daemon can be started
• cli works
• version is updated
• can make a proof call

[placer14]

Some scope clarification:

• What is the signal for knowing when a daemon is "started"? (This will be the objective of the test. Kindness please!)
• How do I make a proof call?

Assumptions I'm making:

• I will assume we're just going to ensure version reported on the CLI matches the current tag.
• "CLI works" probably has a bunch of individual unit tests, I can pick a few just to establish the pattern here for more later.

[jennijuju]

What is the signal for knowing when a daemon is "started"? (This will be the objective of the test. Kindness please!)
Assumptions I'm making:

I will assume we're just going to ensure version reported on the CLI matches the current tag.
"CLI works" probably has a bunch of individual unit tests, I can pick a few just to establish the pattern here for more later.

for both i think we can just make sure lotus version is returning the right version.

How do I make a proof call?

id say

lotus/extern/sector-storage/ffiwrapper/sealer_test.go

Line 78 in aae9bf3

p1, err := sb.SealPreCommit1(context.TODO(), id, s.ticket, []abi.PieceInfo{s.pi})

can be a good example?

but @magik6k probably have a better answer to both q

[placer14]

@jennijuju Some clarification is needed around the "make a proof call" requirement. What would this script do that the above test does not do? What more would adding the lotus context provide for us? (I'm ultimately trying to understand what behavior we're attempting to protect with this check. It would seem the above test already incorporates the latest proof logic with 2k params and validates the proof.)