Build legacy multisig transaction programatically

[LGLO]

Pros:

• it would be easier to keep this test if genesis utxo changes

Cons:

• costs many lines of code and still requires updating genesis utxo buried in applied scripts

[AmbientTea]

still requires updating genesis utxo buried in applied scripts

can't we also programmatically apply the utxo?

[LGLO]

still requires updating genesis utxo buried in applied scripts

can't we also programmatically apply the utxo?

We could by copy-pasting some code from csl module here or by making some pub(crate) functions/structs pub. WDYT?

[kpinter-iohk]

csl (and plutus_scripts for that matter) should be in their own crate (crates?) as they have useful things. I'm pretty sure it happened more than once that I wanted to use something from them and ended up having to copy-paste it. If you look at those modules they are pretty generic and would even make sense outside of PC (they do reference PC in comments or one error message but nothing relevant).

[gilligan]

Any thoughts on putting blobs like that in external files and including them via include_str or include_bytes (depending on how the content is stored in the file obviously) ?

Not a major concern but something like the following seems appealing:

let applied_version_oracle_script = !include_bytes("./applied_version_oracle_script.bin")
plutus_scripts.add(&PlutusScript::new_v2(applied_version_oracle_script.to_vec()));

Maybe I'm ignoring some downsides though? WDYT?

[LGLO]

There is a question if we want to use applied script or we prefer to apply parameters to the base script in a programatic manner. In the latter case we can import it use raw_scripts::VERSION_ORACLE_POLICY

I think it it worth trying to include it from some other file.

[gilligan]

There is quite a lot going on in this test - I wonder if it would be feasible to maybe extract some pieces and get a bit closer to the common structure of setup / "action" / assertion?

let something = setup_something();
let whatever = build_whatever();

let result = run_upsert_d_param(...)

assert!(result, ...)

Might help to make it clearer what is being asserted? WDYT?

[LGLO]

Of course, but this is polishing. The main question for this draft was if it is worth to build this programatically.

[LGLO]

csl (and plutus_scripts for that matter) should be in their own crate (crates?) as they have useful things. I'm pretty sure it happened more than once that I wanted to use something from them and ended up having to copy-paste it. If you look at those modules they are pretty generic and would even make sense outside of PC (they do reference PC in comments or one error message but nothing relevant).

I don't think so. We take only offchain code needs into account there and we also take advantage of having things there usually pub(crate).
Especially csl is a set of offchain needs regarding mostly CSL and also our Ogmios implementation. I don't think we are in business of providing (and if we are, we lack workforce) general purpose improvements to CSL.