btw the out directory is already generated if you don't want to bother

Can you deal with nix as well...? maybe?

every time I tried nix I ended up crying in a corner

sorry not used to it

lol fair enough. @t4ccer do we have enough budget left here do deal with nix?

Thank you for contributing, a couple of initial comments:

• Scripts do not deserialize, all complain with Codec.CBOR: deserialising failed at offset 0 : expected bytes. Make sure to serialize them the same way other implementations are doing that (Haskell function is serialiseCompiledCode from PlutusLedgerApi.V2)
• Scripts should be using V2 versions of ScriptContext, this is important cause all implementations are passed through the same test suite. I recon this may be difficult though if you do not have V1 and V2 types in plu-ts but we're too low on budget to upgrade all existing implementations to V3. Core issue is not the implementations but our testing framework https://github.com/mlabs-haskell/plutus-test that is a set of testing utils destilled from https://github.com/Liqwid-Labs/liqwid-libs and decoupled from plutarch)
• Once Nix work is set up we should not store compilation artifacts in ./out in the repo (I can handle that)
• I'm not sure how implicit data decoding/encoding works in plu-ts but make sure that enums are encoded as integers and structs with only one constructor like a list of fields (no Constr 0), but once you get script serialization fixed the test suite will yell at you
• At the end decide on TS formatter, but later to not cause merge conflicts now
• Announce it in readme section

OK, I've fixed the serialization and nixified the whole build. Now you can see that tests are failing due to data decoding, once that is fixed and V2 script context is figured out we should be good to merge. You can run tests locally either with cabal test inside ./tests directory, there you can override UPLC_BENCHMARK_BIN_PLU_TS env var to point it somewhere outside nix sandbox or with just nix build .#checks.x86_64-linux.uplc-benchmark to run it like it'll be ran on CI