Switch to babbage era .. and all kinds of stuff #408

ch1bo · 2022-06-30T08:31:45Z

Follow up of #355

As described in #406 the goal of this PR is to make all packages compile and target Babbage and PlutusV2 everywhere. From there we can start fixing the tests and make this green. Besides "smaller" issues in the hydra-node tests, bigger tasks are listed in #400.

KtorZ · 2022-06-30T08:37:34Z

hydra-cardano-api/src/Hydra/Cardano/Api.hs

+    txOutAddress
+    (TxOutValue MultiAssetInBabbageEra txOutValue)
+    txOutDatum
+    _txReferenceScript


Add FIXME here to not forget about it.

KtorZ · 2022-06-30T08:40:58Z

hydra-cardano-api/src/Hydra/Cardano/Api/AddressInEra.hs

+-- NOTE: Requires the 'Network' discriminator (Testnet or Mainnet) because
+-- Plutus addresses are stripped off it.
+fromPlutusAddress :: IsShelleyBasedEra era => Network -> Plutus.Address -> AddressInEra era
+fromPlutusAddress network plutusAddress =


Note: unclear why we need that yet, let's see if we can clarify later.

It's from observing commits, there we need to convert from a Plutus.TxOut -> CardanoApi.TxOut ctx era

KtorZ · 2022-06-30T08:56:38Z

hydra-cardano-api/src/Hydra/Cardano/Api/TxOut.hs

+  -- 'TxOutSource' to report corresponding 'TranslationError'. However, this
+  -- value is NOT used for constructing the Plutus.TxOut and hence we hard-code
+  -- it to 0 to not need to bookkeep it and we throw away errors anyway.
+  eitherToMaybe . Ledger.txInfoOutV2 (TxOutFromOutput $ Ledger.TxIx 0) . toLedgerTxOut


Instead of hard-coding a TxIx, we could use a error "TxOutSource" used but shouldn't was we do not expect this to ever be used / evaluated (because of eitherToMaybe)

KtorZ · 2022-06-30T09:16:16Z

hydra-node/src/Hydra/Chain/CardanoClient.hs

@@ -86,9 +86,9 @@ queryProtocolParameters :: NetworkId -> FilePath -> QueryPoint -> IO ProtocolPar
 queryProtocolParameters networkId socket queryPoint =
  let query =
        QueryInEra
-          AlonzoEraInCardanoMode
+          BabbageEraInCardanoMode


This means that those queries can only run in Babbage era. Ideally, we need queries to be able to work on both the current era and the next era. If this isn't something that runQuery does (or some other function from cardano-api) we should extend them to enable that.
An example of such thing can be found here: https://github.com/CardanoSolutions/ogmios/blob/master/server/src/Ogmios/App/Protocol/TxSubmission.hs#L347

KtorZ · 2022-06-30T09:20:35Z

hydra-node/src/Hydra/Chain/Direct/State.hs

@@ -324,7 +324,7 @@ close confirmedSnapshot pointInTime OnChainHeadState{ownVerificationKey, stateMa
    ConfirmedSnapshot{snapshot = Snapshot{number, utxo}, signatures} ->
      CloseWithConfirmedSnapshot
        { snapshotNumber = number
-        , closeUtxoHash = hashTxOuts $ toList utxo
+        , closeUtxoHash = hashUTxO @Tx utxo


Wondering if these are indeed semantically equivalent?

They should be. I will change back the definition of hashUTxO to be not sorting besides using toList.

KtorZ · 2022-06-30T09:22:20Z

hydra-node/src/Hydra/Chain/Direct/Tx.hs

@@ -388,7 +382,7 @@ contestTx vk Snapshot{number, utxo} sig (slotNo, _) ClosedThreadOutput{closedThr
        , parties = closedParties
        , contestationDeadline = closedContestationDeadline
        }
-  utxoHash = toBuiltin $ hashTxOuts $ toList utxo
+  utxoHash = Head.hashTxOuts . mapMaybe toPlutusTxOut $ toList utxo


Should that use the more generic 'hashUTxO' ?

KtorZ · 2022-06-30T09:24:18Z

hydra-node/src/Hydra/Chain/Direct/Tx.hs

  headRedeemer =
    toScriptData (Head.Fanout $ fromIntegral $ length utxo)

  headTokens =
    headTokensFromValue headTokenScript (txOutValue headOutput)

  fanoutOutputs =
-    map toTxContext $ toList utxo
+    map (toTxContext . snd) . sortOn fst $ UTxO.pairs utxo


This change is in principle semantically equivalent, because toList returns elements ordered by key already. So perhaps we could make the naming clearer or add a comment to clarify this.

KtorZ · 2022-06-30T09:26:32Z

hydra-node/src/Hydra/Chain/Direct/Wallet.hs

@@ -192,15 +166,16 @@ newTinyWallet tracer networkId (vk, sk) queryUTxOEtc = do
 -- checking the output's address.
 applyBlock :: Block -> (Address -> Bool) -> Map TxIn TxOut -> Map TxIn TxOut
 applyBlock blk isOurs utxo = case blk of
-  BlockAlonzo (ShelleyBlock (Ledger.Block _ bbody) _) ->
+  BlockBabbage (ShelleyBlock block _) ->


Same as queries, we probably want to support 2 eras always, the current and next one.

KtorZ · 2022-06-30T09:31:22Z

hydra-node/src/Hydra/Chain/Direct/Wallet.hs

-              let outs = outputs (body tx)
-               in StrictSeq.zip (StrictSeq.fromList [0 .. length outs]) outs
-        forM_ indexedOutputs $ \(fromIntegral -> ix, out@(TxOut addr _ _)) ->
+              let outs = map sizedValue . toList $ outputs (body tx)


Perhaps use outputs' or getField @"outputs" to avoid dealing with sized value.

KtorZ · 2022-06-30T09:32:16Z

hydra-node/src/Hydra/Chain/Direct/Wallet.hs

@@ -262,14 +237,14 @@ coverFee_ pparams systemStart epochInfo lookupUTxO walletUTxO partialTx@Validate
      mkChange
        output
        resolvedInputs
-        (toList $ outputs body)
+        (map sizedValue . toList $ outputs body)


Same here -> outputs'

KtorZ · 2022-06-30T09:36:53Z

hydra-node/src/Hydra/Ledger/Cardano.hs

@@ -122,7 +114,7 @@ instance IsTx Tx where

  txId = getTxId . getTxBody
  balance = foldMap txOutValue
-  hashUTxO = hashTxOuts . toList
+  hashUTxO = fromBuiltin . Head.hashTxOuts . mapMaybe (toPlutusTxOut . snd) . sortOn fst . UTxO.pairs


map snd . sortOn fst . UTxO.pairs should be equivalent to toList

KtorZ · 2022-06-30T09:44:03Z

hydra-node/src/Hydra/Ledger/Cardano/Configuration.hs

@@ -79,5 +77,5 @@ newLedgerEnv pparams =
      -- both unused in Hydra. There might be room for interesting features in the
      -- future with these two but for now, we'll consider them empty.
      Ledger.ledgerAccount = Ledger.AccountState mempty mempty
-    , Ledger.ledgerPp = toLedgerPParams ShelleyBasedEraAlonzo pparams
+    , Ledger.ledgerPp = toLedgerPParams ShelleyBasedEraBabbage pparams


We should make sure to also adjust the sample protocol parameters we have to be Babbage's params instead of Alonzo (i.e. remove d and extraEntropy)

Updated the included protocol-parameters.json with ones from the vasil testnet (also has updated cost models).

Note that the one retrieved from the vasil testnet using the cardano-cli did include d and entropy as null fields:

[...] "decentralization": null, "executionUnitPrices": { "priceMemory": 5.77e-2, "priceSteps": 7.21e-5 }, "extraPraosEntropy": null, [...]

Yes, because the cardano-cli / cardano-api have only one data-type to represent both Alonzo and Babbage params, with Maybe values 😬 ...

github-actions · 2022-07-07T15:24:47Z

Transactions Costs

Sizes and execution budgets for Hydra protocol transactions. Note that unlisted parameters are currently using arbitrary values and results are not fully deterministic and comparable to previous runs.

Metadata
Generated at	2022-07-08 08:46:25.217429909 UTC
Max. memory units	14000000.00
Max. CPU units	10000000000.00
Max. tx size (kB)	16384

Cost of Init Transaction

Parties	Tx size	% max Mem	% max CPU
1	4831	9.02	3.55
2	5042	10.92	4.28
3	5241	13.34	5.22
5	5651	17.36	6.77
10	6677	28.62	11.15
46	14058	99.99	38.65

Cost of Commit Transaction

Currently only one UTxO per commit allowed (this is about to change soon)

UTxO	Tx size	% max Mem	% max CPU
1	5771	19.85	8.02

Cost of CollectCom Transaction

Parties	Tx size	% max Mem	% max CPU
1	13057	20.68	8.29
2	13380	36.34	14.72
3	13595	52.93	21.58
4	14094	77.82	31.92
5	14237	98.58	40.59

Cost of Close Transaction

Parties	Tx size	% max Mem	% max CPU
1	9331	8.19	3.36
2	9560	9.85	4.16
3	9663	9.75	4.27
5	10021	11.75	5.34
10	10912	16.67	7.99
30	14226	32.58	17.14
70	16199	41.35	15.65

Cost of Contest Transaction

Parties	Tx size	% max Mem	% max CPU
1	9331	8.16	3.34
2	9493	8.94	3.79
3	9694	10.16	4.42
5	10017	11.72	5.32
10	10882	16.01	7.73
30	14222	32.42	17.07
43	16362	42.37	22.86

Cost of Abort Transaction

Parties	Tx size	% max Mem	% max CPU

Cost of FanOut Transaction

Involves spending head output and burning head tokens. Uses ada-only UTxO for better comparability.

UTxO	Tx size	% max Mem	% max CPU
1	13424	9.88	4.31
2	13459	11.40	5.19
3	13561	13.30	6.22
5	13694	16.72	8.12
10	13878	24.88	12.75
50	15250	84.88	47.60
59	15574	98.85	55.63

github-actions · 2022-07-07T16:05:35Z

Unit Test Results

234 tests +4 232 ✔️ +4 12m 37s ⏱️ + 3m 41s
  84 suites ±0     2 💤 ±0
    5 files ±0     0 ❌ ±0

Results for commit 254cb9c. ± Comparison against base commit 799da77.

This pull request removes 19 and adds 23 tests. Note that renamed tests count towards both.

Hydra.Chain.Direct.Contract/TxOut hashing ‑ OffChain.hashTxOuts == OnChain.hashTxOuts
Hydra.Chain.Direct.Contract/TxOut hashing ‑ serialize' commutes with hashing
Hydra.Chain.Direct.State/close ‑ validates within ExecutionUnits {executionSteps = 10000000000, executionMemory = 14000000}
Hydra.Chain.Direct.State/collectCom ‑ validates within ExecutionUnits {executionSteps = 10000000000, executionMemory = 14000000}
Hydra.Chain.Direct.State/contest ‑ validates within ExecutionUnits {executionSteps = 10000000000, executionMemory = 14000000}
Hydra.Chain.Direct.State/fanout ‑ validates within ExecutionUnits {executionSteps = 10000000000, executionMemory = 14000000}
Hydra.ClientInput/JSON encoding of (ReasonablySized (ClientInput (Tx AlonzoEra))) ‑ allows to encode values with aeson and read them back
Hydra.ClientInput/JSON encoding of (ReasonablySized (ClientInput (Tx AlonzoEra))) ‑ produces the same JSON as is found in golden/ReasonablySized (ClientInput (Tx AlonzoEra)).json
Hydra.Ledger.Cardano/JSON encoding of (Tx AlonzoEra) ‑ allows to encode values with aeson and read them back
Hydra.Ledger.Cardano/JSON encoding of (Tx AlonzoEra) ‑ produces the same JSON as is found in golden/Tx AlonzoEra.json
…

Hydra.Chain.Direct.Contract/Serializing commits ‑ deserializeCommit . serializeCommit === id
Hydra.Chain.Direct.Contract/TxOut hashing ‑ OffChain.hashUTxO == OnChain.hashTxOuts (on sorted tx outs)
Hydra.Chain.Direct.Contract/TxOut hashing ‑ OnChain.hashPreSerializedCommits == OnChain.hashTxOuts (on sorted tx outs)
Hydra.Chain.Direct.State/close ‑ validates within maxTxExecutionUnits
Hydra.Chain.Direct.State/collectCom ‑ validates within maxTxExecutionUnits
Hydra.Chain.Direct.State/contest ‑ validates within maxTxExecutionUnits
Hydra.Chain.Direct.State/fanout ‑ validates within maxTxExecutionUnits
Hydra.ClientInput/JSON encoding of (ReasonablySized (ClientInput (Tx BabbageEra))) ‑ allows to encode values with aeson and read them back
Hydra.ClientInput/JSON encoding of (ReasonablySized (ClientInput (Tx BabbageEra))) ‑ produces the same JSON as is found in golden/ReasonablySized (ClientInput (Tx BabbageEra)).json
Hydra.Ledger.Cardano/Generators ‑ arbitrary @(Hash PaymentKey) is reasonably collision resistant
…

♻️ This comment has been updated with latest results.

ch1bo

Some minor comments on recent changes

ch1bo · 2022-07-07T16:26:30Z

hydra-cluster/test/Test/EndToEndSpec.hs

+                    , "datum" .= Null
+                    , "datumhash" .= Null
+                    , "inlineDatum" .= Null
+                    , "referenceScript" .= Null


😮 Is this our life now?

cf commit message: 13626d5

ch1bo · 2022-07-07T16:27:51Z

hydra-cluster/test/Test/DirectChainSpec.hs

+              alicesCallback `shouldSatisfyInTime` \case
+                Observation OnCloseTx{snapshotNumber, remainingContestationPeriod} ->
+                  -- FIXME(SN): should assert contestationDeadline > current
+                  Just (snapshotNumber == 1, remainingContestationPeriod)


The Maybe (Bool, b) is a bit of an odd type.. how about Maybe b and do a guard $ snapshotNumber == 1 here?

ch1bo · 2022-07-07T16:32:07Z

hydra-test-utils/src/Test/Hydra/Prelude.hs

+ where
+  -- NOTE: Somehow, since 1.35.0, cleaning-up cardano-node database directory
+  -- _sometimes_ generates an empty 'clean' file which prevents the 'db' folder
+  -- to be fully removed and triggers an 'UnsatisfiedConstraints' IOException.


This is an odd workaround and refers to the cardano-node behavior, while this function is very generic and lives in hydra-test-utils. Can't we do some node-specific cleanup prior to this in the action?

I know. If you feel like debugging this, go ahead. Spent already an hour chasing this thing and gave up.

ch1bo · 2022-07-07T16:33:47Z

hydra-cluster/src/CardanoNode.hs

+  kesKeyFilename i =
+    dirname i </> "kes.skey"
+  opCertFilename i =
+    dirname i </> "opcert.cert"


We can get rid of all those is by simply defining dirname = "stake-pool-" <> show (nodeId cfg)

Addressed this myself.

Mostly cardano-node and consensus from open PRs + plutus a bit further ahead on release/1.0.0 to get serialiseData

…crypto-class

This is mostly updating TxOut, TxOutDatum and TxBodyContent patterns.

We picked some interesting periods like day, week etc. as well as the horizon when a transaction is definitely final on mainnet (after k blocks)

This allows us again to get hold of the comittedUTxO for a counterexample

This makes it individual committedUTxO explicit. In some cases the individual entries "collapsed" because of identical TxIn values.

For some reason the Arbitrary instance for Ledger.TxIn seems to have degraded in likelihood of collisions, but Arbitrary instance on the Cardano.Api.TxIn has not (we defined that one).

They need to be collision resistant (stronger) or do not generate "collapsing" foldable values (weaker).

This makes us depend (hopefullly) temporarily on a fork of cardano-node that fixes ToJSON insteance for ledger scripts.

Also drop Num instance for ContestationPeriod and instead define a fixture 'cperiod'.

Dropped that function and module because we do not intend to run multiple cardano nodes, but instead a single block producer holding all the stake going forward. We separated the things contained into Hydra.Cluster.{Faucet,Fixture,Util}.

No more BFT nodes, so we now fork into the last era on 'epoch 0' and run a single stake pool.

We would like to use the cardano-ledger's generator to test our internal ledger behaves correctly but this is currently problematic because it's hard to tweak the generation process to produce only the kind of txs we want in a Head. Given we are only using those generators for "smoke" testing the Ledger's validation function over a sequence of txs, it's probably fine but we should revisit this in the future.

The node is no longer a BFT one though, will rename eventually.

This is needed until we leverage reference scripts and inputs.

Somehow... cardano-node database now include an empty file 'clean', of 0 byte, which doesn't gets deleted (or is maybe re-created?) and causes the 'removePathForcibly' function to raise an awkward exception: removeDirectory: unsatisfied constraints (Directory not empty) despite having deleted everything else. If we inspect temporary directories after the exception is raised, they look like the following: run/user/1000/hydra-cluster-b5c7621dbb5c60a1 └── db └── clean With that 'clean' file being empty and of size 0. :shrugging-man:

Instead of computing a delay from hard-coded parameters and error-prone equation. This was now apparently triggered too fast? Maybe the computation was bonkers from the start and we didn't noticed somehow because BFT nodes took longer to process the transaction? This is more correct anyway so, good to do.

I still don't like this instance. It's cumbersome, inconsistent and ugly. I wish we also had more control over our JSON API, there's too much things that can break because we update cardano-api at the moment.

Also left a FIXME about the wrong promises made by the function name and its arguments.

ch1bo · 2022-07-08T08:57:11Z

It was a team effort.. no approval necessary ;)

ch1bo mentioned this pull request Jun 30, 2022

Spike: Babbage preview #355

Closed

33 tasks

KtorZ reviewed Jun 30, 2022

View reviewed changes

ch1bo force-pushed the ch1bo/babbage-era branch from df70d82 to 0fadac6 Compare June 30, 2022 11:00

ch1bo linked an issue Jun 30, 2022 that may be closed by this pull request

Rebase babbage-preview and make everything compile (mark failing tests as pending) #406

Closed

ch1bo requested a review from a user June 30, 2022 11:15

ch1bo marked this pull request as ready for review June 30, 2022 11:15

ch1bo force-pushed the ch1bo/babbage-era branch from 07e62b7 to 7931427 Compare July 5, 2022 08:06

ch1bo mentioned this pull request Jul 6, 2022

Use cardano-api Key interface for Hydra keys #398

Merged

ch1bo force-pushed the ch1bo/babbage-era branch 2 times, most recently from 8994ef0 to 1b1819f Compare July 7, 2022 12:21

ghost mentioned this pull request Jul 7, 2022

Use new/updated transaction sequence generators from the cardano-ledger #420

Closed

ch1bo commented Jul 7, 2022

View reviewed changes

ch1bo mentioned this pull request Jul 8, 2022

Rebase babbage-preview and make everything compile (mark failing tests as pending) #406

Closed

ch1bo added 4 commits July 8, 2022 08:01

Update most dependencies to latest & greatest

a5a2bc6

Mostly cardano-node and consensus from open PRs + plutus a bit further ahead on release/1.0.0 to get serialiseData

Use more recent nixpkgs and ensure secp256k1 is in scope for cardano-…

e8c5954

…crypto-class

Re-use nixpkgs between shell.nix and default.nix

170f142

Make hydra-cardano-api babbage ready, but stay on Alonzo

8b01c2d

This is mostly updating TxOut, TxOutDatum and TxBodyContent patterns.

ch1bo and others added 23 commits July 8, 2022 08:02

Detail definition of Arbitrary ContestationPeriod

2be344a

We picked some interesting periods like day, week etc. as well as the horizon when a transaction is definitely final on mainnet (after k blocks)

Undo factoring of genCollectComTx to debug

cb01dd8

This allows us again to get hold of the comittedUTxO for a counterexample

Return [UTxO] from executeCommits to debug

09017cc

This makes it individual committedUTxO explicit. In some cases the individual entries "collapsed" because of identical TxIn values.

Not generate colliding TxIn values in genCommit

7d89b8a

For some reason the Arbitrary instance for Ledger.TxIn seems to have degraded in likelihood of collisions, but Arbitrary instance on the Cardano.Api.TxIn has not (we defined that one).

Define properties on UTxO generators

e49ce2d

They need to be collision resistant (stronger) or do not generate "collapsing" foldable values (weaker).

Fix genUTxoFor and drop unused genUTxO

8b298ee

Update golden files

a669712

Change cardano-api upstream dependency to use correct ToJSON

df462f4

This makes us depend (hopefullly) temporarily on a fork of cardano-node that fixes ToJSON insteance for ledger scripts.

Fix test on alternative ClientInput format

3ca62d9

Fix use of ContestationPeriod in TUI

3d3f91e

Copy alice, bob & carol fixtures into hydra-cluster

49c5334

Also drop Num instance for ContestationPeriod and instead define a fixture 'cperiod'.

Bump expected cardano-node version

8db5b39

Update README of hydra-cluster

5f9cc8d

Rework devnet configuration for Babbage

2ed3133

No more BFT nodes, so we now fork into the last era on 'epoch 0' and run a single stake pool.

Fix compilation errors in TUI

b570847

Fix pathnames for starting 'BFT Node'

8df84ad

The node is no longer a BFT one though, will rename eventually.

Bump e2e tests' transaction max size back to 32K

a252ee2

This is needed until we leverage reference scripts and inputs.

Fix expected snapshot JSON instance in e2e

b1cf9a6

I still don't like this instance. It's cumbersome, inconsistent and ugly. I wish we also had more control over our JSON API, there's too much things that can break because we update cardano-api at the moment.

Factor away some complexity of withBFTNode

fa5d3de

Also left a FIXME about the wrong promises made by the function name and its arguments.

ch1bo force-pushed the ch1bo/babbage-era branch from e9c44cd to fa5d3de Compare July 8, 2022 06:02

abailly and others added 3 commits July 8, 2022 08:28

Stop-gap solution to trim-down length of Model traces

284dacb

Only stop in Open state (safety belt + suspender)

e9cc0d6

Add some details to changelog for babbage update

254cb9c

ch1bo merged commit 11c015b into master Jul 8, 2022

ch1bo deleted the ch1bo/babbage-era branch July 8, 2022 08:56

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Switch to babbage era .. and all kinds of stuff #408

Switch to babbage era .. and all kinds of stuff #408

ch1bo commented Jun 30, 2022 •

edited

Loading

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

ch1bo Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

ch1bo Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022 •

edited

Loading

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

KtorZ Jun 30, 2022

ch1bo Jun 30, 2022

ch1bo Jun 30, 2022

KtorZ Jun 30, 2022

github-actions bot commented Jul 7, 2022 •

edited

Loading

github-actions bot commented Jul 7, 2022 •

edited

Loading

ch1bo left a comment

ch1bo Jul 7, 2022

KtorZ Jul 7, 2022

ch1bo Jul 7, 2022

ch1bo Jul 7, 2022

KtorZ Jul 7, 2022

ch1bo Jul 7, 2022

ch1bo Jul 7, 2022

ch1bo commented Jul 8, 2022

Switch to babbage era .. and all kinds of stuff #408

Switch to babbage era .. and all kinds of stuff #408

Conversation

ch1bo commented Jun 30, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

KtorZ Jun 30, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Jul 7, 2022 • edited Loading

Transactions Costs

Cost of Init Transaction

Cost of Commit Transaction

Cost of CollectCom Transaction

Cost of Close Transaction

Cost of Contest Transaction

Cost of Abort Transaction

Cost of FanOut Transaction

github-actions bot commented Jul 7, 2022 • edited Loading

Unit Test Results

ch1bo left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ch1bo commented Jul 8, 2022

ch1bo commented Jun 30, 2022 •

edited

Loading

KtorZ Jun 30, 2022 •

edited

Loading

github-actions bot commented Jul 7, 2022 •

edited

Loading

github-actions bot commented Jul 7, 2022 •

edited

Loading