Set upload fees for contracts higher

[JakeHartnell]

It should cost more to upload a contract. I believe there is already a parameter for this? Let's update it on the Uni testnet.

If there is not a parameter, we should add this feature. Transactions should be cheap, but uploading a contract should be expensive.

[blockcreators]

Governance can change it later but what do you think the initial cost per byte should be?
And how much would the average contract be? (byte cost * average contract bytes)

To make a simple token via junomint, how much do we expect it to cost?

[the-frey]

Will have a look at this later today to work out what the param change would be. I think we also are going to need to document the param changes that this and #82 mean we will need to do on mainnet before we can upgrade. Moneta is gradually becoming quite hairy and we should have a pretty foolproof checklist for the upgrade to run imo with backout points if possible

[the-frey]

...which I guess in practical terms means I need to write the checklist lol

[JakeHartnell]

...which I guess in practical terms means I need to write the checklist lol

Appreciate all your help on this. 🙏

[JakeHartnell]

...which I guess in practical terms means I need to write the checklist lol

Let's try to keep the list here in GitHub issues. Let's try to use this Milestone: https://github.com/CosmosContracts/juno/milestone/1

[the-frey]

okay so there's also the upload size to consider too.

In wasm;

// DefaultMaxWasmCodeSize limit max bytes read to prevent gzip bombs
DefaultMaxWasmCodeSize = 600 * 1024 * 2

and

junod query params subspace wasm maxWasmCodeSize --chain-id uni

confirms we are using default

key: maxWasmCodeSize
subspace: wasm
value: '"1228800"'

[the-frey]

While the various costs are specified https://github.com/CosmWasm/wasmd/blob/57517b0c333e23aedd2bbccf80579cf75081241b/x/wasm/keeper/gas_register.go#L13

With reasoning https://github.com/CosmWasm/cosmwasm/blob/v1.0.0-beta/docs/GAS.md

And some further discussion: CosmWasm/wasmd#634 (comment)

[JakeHartnell]

While the various costs are specified https://github.com/CosmWasm/wasmd/blob/57517b0c333e23aedd2bbccf80579cf75081241b/x/wasm/keeper/gas_register.go#L13

With reasoning https://github.com/CosmWasm/cosmwasm/blob/v1.0.0-beta/docs/GAS.md

Good find! Looks like this is maybe what we want:

// DefaultCompileCost is how much SDK gas is charged *per byte* for compiling WASM code.
// Benchmarks and numbers were discussed in: https://github.com/CosmWasm/wasmd/pull/634#issuecomment-938056803
DefaultCompileCost uint64 = 3

[the-frey]

Yup, although it's hard coded AFAICT which would mean forking. AFAIUI changes there are consensus-breaking, hence why they aren't params. That said, it's lateish in the day and my caffiene level is drooping so may have that wrong.

[faddat]

hate that caffeine drooping feeling

[the-frey]

I have a path to doing this, but given that Confio (as per the reasoning & discussion links above) have thought about this a fair bit, I am... still not sold on overriding these values.

Moreover, if we were to override these values we'd have to do a few more rounds of testing. Which isn't necessarily a bad thing, but November is flying by...

[JakeHartnell]

I have a path to doing this, but given that Confio (as per the reasoning & discussion links above) have thought about this a fair bit, I am... still not sold on overriding these values.

Moreover, if we were to override these values we'd have to do a few more rounds of testing. Which isn't necessarily a bad thing, but November is flying by...

Confio thought about these values in the context on Confio and tgrade, we need to think about them in the context of Juno.

Yup, although it's hard coded AFAICT which would mean forking. AFAIUI changes there are consensus-breaking, hence why they aren't params.

Well, we already have a fork. Would be great to make some of these params and contribute that upstream. Having DefaultCompileCost as a param would be the best option for us going forward.

[the-frey]

FWIW I'm not against changing this and then testing, it's a bunch more stuff to test. It's potentially doable in base Juno, no fork required - but I am against making them params. At the minute I'm dubious about exposing something that's an obvious attack vector to governance. Could be convinced in future however.

[the-frey]

Okay have spiked something on https://github.com/CosmosContracts/juno/tree/override_cw_default_compile_costs - don't think it's quite there yet. Need to test a couple of things before PR