Update to brotli v1.1.0

[eljobe]

The previous pin was almost 3 years old, and was causing some ugly warnings during builds.

This change also deals with google/brotli@641bec0 which stopped creating the static libraries by default and switched to only building them with -DBUILD_SHARED_LIBS=OFF.

Also, now that the static libraries are unambiguous, the -static suffix was removed from the library names.

[PlasmaPower]

I was a bit worried at first because brotli up to v1.0.5 kept improving the compression, which would be a breaking consensus chain because we measure how large compressed txs are to charge for them. But it seems like since then compression has been stable, so this should work in theory. I don't think we should include it in v3.1.2 though, and I think we should look at doing more specialized testing such as rerunning it against the Arbitrum One chain history to make sure this doesn't change behavior.

[eljobe]

I'd be happy to run it against the Arbitrum One chain history to make sure that this doesn't change behavior.
Do you have any pointers to practical steps which would be required to do that?

The first thing that comes to my mind is:

1. Wait for Validation Inputs wiring #2604 to be submitted and released to produciton.
2. Call the ValidationInputsAt RPC call with a view different blocks to get some bock_inputs_*.json files.
3. Validate those block_inputs_*.json files with the prover built from this branch and confirm the final global state is the same as what it was on mainnet.

Is that the idea, or did you have a different process in mind?