[BUG] - min required UTxO reported by CLI is not correct

[mkoura]

Internal

Area
Other Any other topic (Delegation, Ranking, ...).

Summary
In Babbage, both transaction build and transaction calculate-min-required-utxo report 857690 Lovelace to be minimum required UTxO for Lovelace-only output with no datum. However I am able to create UTxO with just 849070 Lovelace (using transaction build-raw).

Expected behavior
CLI reports precise minimum required UTxO.

System info:

• OS Name: Fedora 35
• Node version: 1.35.0-rc3

[gitmachtl]

The transaction calculate-min-required-utxo is a bit strange, because how to query the real minimum value with this function? In the past i used --tx-out ADDRESS+0 because of the word boundary it didn't make a difference in the outcome and the result was the minUTXO that must be used. But now that the calculation is done down to the bytelevel, the result of a minUTXO request depends on the amount you're doing the request for??? Thats strange. Examples:

./cardano-cli transaction calculate-min-required-utxo --babbage-era \
     --protocol-params-file <(cat babbage-protocol-parameters.tmp) \
     --tx-out "addr_test1vpfwv0ezc5g8a4mkku8hhy3y3vp92t7s3ul8g778g5yegsgalc6gc+0"
Lovelace 840450

the value above is wrong, because if you try to send 840450 lovelaces it really is

./cardano-cli transaction calculate-min-required-utxo --babbage-era \
     --protocol-params-file <(cat babbage-protocol-parameters.tmp) \
     --tx-out "addr_test1vpfwv0ezc5g8a4mkku8hhy3y3vp92t7s3ul8g778g5yegsgalc6gc+840450"
Lovelace 857690

the value above is accepted by the chain

./cardano-cli transaction calculate-min-required-utxo --babbage-era \
     --protocol-params-file <(cat babbage-protocol-parameters.tmp) \
     --tx-out "addr_test1vpfwv0ezc5g8a4mkku8hhy3y3vp92t7s3ul8g778g5yegsgalc6gc+10000000000"
Lovelace 874930

the value above is too high

so why is the amount of the minUTXO depending on the amount i wanna send, that makes no sense IMO. or do i miss something? there is only ONE real minUTXO value on the chain, not a gliding value?

there should be an integrated iteration within cardano-cli to get to the real min value with a starting lovelace amount of zero.

same goes for sending out assets, requesting the minUTXO for
--tx-out ADDR+0+1 POLICY.ASSETNAME (example)
should return the right amount of lovelaces, but if you request it that way, the returned value is too low because of the wrong lower boundary for the lovelaces part.

[JaredCorduan]

The build command might be overestimating the size needed, perhaps using maxBound somewhere.

The min value depends (linearly) on the number of bytes of the output. You cannot therefore figure it out precisely by substituting zero for the value you want. See this CIP: https://github.com/cardano-foundation/CIPs/pull/265/files

But I guess the suggestion is for the CLI to take an output with zero ada, and compute the fix point?

[gitmachtl]

But I guess the suggestion is for the CLI to take an output with zero ada, and compute the fix point?

Yes exactly, like its now working in the alonzo-era. Otherwise external tools would need to do at least two or three requests with different lovelaces values to get the amount that really works for sending it out with that value.

[JaredCorduan]

@lehins has kindly volunteered to make an era-generic function for computing the fixed point for a given transaction output, in the ledger code. After we have that, we figure out how to hook it into the build and build-raw commands.

[Jimbo4350]

In Babbage, both transaction build and transaction calculate-min-required-utxo report 857690 Lovelace to be minimum required UTxO for Lovelace-only output with no datum. However I am able to create UTxO with just 849070 Lovelace (using transaction build-raw).

So this is complaining about the change output. We purposely overestimate the size of the change output because we want to avoid recursively solving for the tx fee (i.e evaluate tx fee, put fee in tx, evaluate the fee again, put the fee in the tx etc). This is expected.

[CarlosLopezDeLara]

Removing from 1.35.1. As @Jimbo4350 mentioned, current behavior is intended and expected.

[lehins]

@Jimbo4350 we now have a function setMinCoinTxOut in cardano-ledger-api (which is a new packahge tha twe started that will provide all the useful functionality for downstream users of ledger).

That function will calculate the true minimum lovelace for any TxOut in any era precisely by finding the fixpoint:

https://github.com/input-output-hk/cardano-ledger/blob/1176affe9b2a7a3b50c1ceda00cc92f237ea6360/libs/cardano-ledger-api/src/Cardano/Ledger/Api/Tx/Out.hs#L32-L40

@CarlosLopezDeLara You maybe want to consider reopening this ticket, since now it will be very easy to adopt this in cardano-node thus minimize the amount and make everyone happy