feat: SDK beta release

feat(sdk): start using sdk in integration tests

feat(sdk): implement message wait time estimation

integration-tests: Remove superfluous nightly env tests

Fix Typo In Comment

These aren't needed anymore now that we've released the Berlin HF on Kovan and nightly has been reset.

feat(sdk): refactor into single messenger class

Extract core BSS logic into bss-core package

fix: actually allow BSS log level to be set via flags

fix: mimic BSS timestamp bug fix from #2093

crtiical -> critical

See https://github.com/ethereum-optimism/optimism/pull/2093 for more
details.

l2geth: bring back unsupported RPC methods

This was a known bug, but recently caught the reason in combing through
the new copy in the specs repo. The old hack that auto defaulted to
debug has been removed.

Removed constants from data-transport-layer

feat(sdk): implement finalize message

Depreciated unused variable from Integration tests

removed unused variable

removed-sequencer

removed-sequencer references

remove sequencer variable

removed unused variable

removed-sequencer

removed-sequencer references

remove sequencer variable

feat(sdk): introduce token bridge adapters

test(integration): withdrawing a fake L2 token

fix(bss,dtl): correctly submit and parse timestamp information for L1 to L2 txs

go/batch-submitter: Downgrade Alpine version

Alpine 3.14 has an issue (https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.14.0#faccessat2) that prevents it from working on CircleCI at this time.

integration-tests: Increase withdrawal timeout

This commit adds support for ERC20 deposits and withdrawals by introducing the
concept of token bridge adapters. This gives the SDK a standard
interface for interacting with token bridges even though the bridges may
each have slightly different internal logic.

This commit makes sure that the DTL correctly returns timestamps using
the new scheme when the first batch submitter hardfork block is
activated. If the block is not activated, the DTL will operate as it
previously did.

Monotonicity violations were being reported in two locations, once in
the sync service and then later in the worker. Because of recent updates
in the sync service, there's never a case where monotonicity violations
would not be caught first within the sync service. This code path only
triggered within L1 syncing verifiers and caused issues with unintended
timestamp manipulation.

.env files were not being ignored by docker. This meant that any local
environment variable changes you placed into an .env file would be
pulled into docker images built inside of the ops package. Docker images
should only be using environment variables as specified in their
respective environment folders and should not copy over .env files.