feat(orchestration): deposit ERTP payment to ICA #9342

0xpatrickdev · 2024-05-08T17:58:48Z

closes: #9193
closes: #9042

Description

Adds .deposit() method to stakingAccountHolder to enable ERTP payment deposits to an Interchain Account. Flow:
1. LocalChainAccount is created and Payment is deposited there
2. LocalChainAccount sends ibc/MsgTransfer to ChainAccount
  - ChainTimerService is used to create a timeout timestamp. We take the current time and add five minutes to it
3. If anything fails in the wallet offer flow:
  - Attempt is made to deposit live payment back to the user's seat
  - If payment is already in LCA, but failed to make it to the ICA, it is withdrawn back to the user's seat
Adds .withdraw() method to LocalChainAccount
Only accepts deposits of bondDenom until bidirectional lookup of denoms / brands for IBC #9211, orchestration - chain - provide well-known chain #9063 (need to determine denom from Amount/Brand
Hardcodes chain parameters in terms until orchestration - chain - provide well-known chain #9063, vat-orchestration expose client (chain) connection info #8879

Security Considerations

Offer Safety: withdrawFromSeat is needed to retrieve a live payment from a seat. Offer safety is upheld because the user is not allowed to request a Want, but we need to proceed carefully. There is a chance the deposit fails along the way, and we need to ensure the payment can be recovered and deposited back to a users seat.
chainTimerService is needed to create a timeout parameter for MsgTransfer

Scaling Considerations

In order for a ChainAccount to have a deposit() functionality, a LocalChainAccount needs to be created. This could lead to a lot of LCA's
- given the LCA is currently closed over in state, and might wind up with funds the holder needs to access, we might want to consider exposing interfaces for the LCA's (.withdraw, .executeTx, .query)
TimerService is needed to create a MsgTransfer timeout parameter. See Orchestration: optimize timeoutHeight parameter for transactions, queries, and transfers #9324 for more details

Documentation Considerations

Testing Considerations

I wanted to add more explicit tests for the functionality added to LocalChainAccount, but struggled to find a place to add them (and get access to a mint in bootstrap tests). Both deposit() and withdraw() are tested indirectly by tests in boot/../test-orchestration.ts
I would like feedback on the imperative .deposit() flow - what happens if this throws? Can we assume the callers supplied payment is still live if it wasn't deposited to the LCA?

Upgrade Considerations

- refs: #9193

- refs: 9193

- uses a LocalChainAccount to accept an ERTP Payment and transfer it to an ICA account - uses ChainTimerService to create a timeout timestamp for ibc/MsgTransfer

cloudflare-workers-and-pages · 2024-05-08T18:10:07Z

Deploying agoric-sdk with Cloudflare Pages

Latest commit:	`37083e6`
Status:	✅ Deploy successful!
Preview URL:	https://37ad759a.agoric-sdk.pages.dev
Branch Preview URL:	https://9193-deposit-to-ica.agoric-sdk.pages.dev

View logs

0xpatrickdev · 2024-05-08T18:14:05Z

packages/orchestration/src/exos/stakingAccountKit.js

+          trace('Transferring funds to ICA');
+          /**
+           * // TODO can we infer `/ibc.applications.transfer.v1.MsgTransferResponse`?
+           * @type {unknown[]}


My current (unsuccessful) attempt at this:

packages/cosmic-proto/src/helpers.ts

git diff packages/cosmic-proto/src/helpers.ts diff --git a/packages/cosmic-proto/src/helpers.ts b/packages/cosmic-proto/src/helpers.ts index ca192be60..00ebfea30 100644 --- a/packages/cosmic-proto/src/helpers.ts +++ b/packages/cosmic-proto/src/helpers.ts @@ -1,6 +1,15 @@ -import type { QueryAllBalancesRequest } from './codegen/cosmos/bank/v1beta1/query.js'; -import type { MsgSend } from './codegen/cosmos/bank/v1beta1/tx.js'; -import type { MsgDelegate } from './codegen/cosmos/staking/v1beta1/tx.js'; +import type { + QueryAllBalancesRequest, + QueryAllBalancesResponse, +} from './codegen/cosmos/bank/v1beta1/query.js'; +import type { + MsgSend, + MsgSendResponse, +} from './codegen/cosmos/bank/v1beta1/tx.js'; +import type { + MsgDelegate, + MsgDelegateResponse, +} from './codegen/cosmos/staking/v1beta1/tx.js'; import { RequestQuery } from './codegen/tendermint/abci/types.js'; import type { Any } from './codegen/google/protobuf/any.js'; import { @@ -13,17 +22,32 @@ import { * `unknown` but it's actually this. The `value` string is a base64 encoding of * the bytes array. */ -export type AnyJson = { typeUrl: string; value: string }; - -// TODO codegen this by modifying Telescope +export type AnyJson = { + typeUrl: string; + value: string; +}; export type Proto3Shape = { '/cosmos.bank.v1beta1.MsgSend': MsgSend; + '/cosmos.bank.v1beta1.MsgSendResponse': MsgSendResponse; '/cosmos.bank.v1beta1.QueryAllBalancesRequest': QueryAllBalancesRequest; + '/cosmos.bank.v1beta1.QueryAllBalancesResponse': QueryAllBalancesResponse; '/cosmos.staking.v1beta1.MsgDelegate': MsgDelegate; + '/cosmos.staking.v1beta1.MsgDelegateResponse': MsgDelegateResponse; '/ibc.applications.transfer.v1.MsgTransfer': MsgTransfer; '/ibc.applications.transfer.v1.MsgTransferResponse': MsgTransferResponse; }; +export type ResponseType<T> = + T extends '/ibc.applications.transfer.v1.MsgTransfer' + ? '/ibc.applications.transfer.v1.MsgTransferResponse' + : T extends '/cosmos.bank.v1beta1.MsgSend' + ? '/cosmos.bank.v1beta1.MsgSendResponse' + : T extends '/cosmos.staking.v1beta1.MsgDelegate' + ? '/cosmos.staking.v1beta1.MsgDelegateResponse' + : T extends '/cosmos.bank.v1beta1.QueryAllBalancesRequest' + ? '/cosmos.bank.v1beta1.QueryAllBalancesResponse' + : never; + /** * The encoding introduced in Protobuf 3 for Any that can be serialized to JSON. *

packages/vats/src/localchain.js

git diff packages/vats/src/localchain.js diff --git a/packages/vats/src/localchain.js b/packages/vats/src/localchain.js index 5a07e8ad3..6ab60c379 100644 --- a/packages/vats/src/localchain.js +++ b/packages/vats/src/localchain.js @@ -5,6 +5,8 @@ import { AmountShape } from '@agoric/ertp'; const { Fail, bare } = assert; +/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */ + /** * @typedef {{ * system: import('./types.js').ScopedBridgeManager; @@ -82,8 +84,9 @@ const prepareLocalChainAccount = zone => return E(purse).withdraw(amount); }, /** - * @param {import('@agoric/cosmic-proto').TypedJson<unknown>[]} messages - * @returns {Promise<import('@agoric/cosmic-proto').TypedJson[]>} + * @template {keyof Proto3Shape} T + * @param {TypedJson<T>[]} messages + * @returns {Promise<TypedJson<ResponseType<T>>[]>} */ async executeTx(messages) { const { address, powers } = this.state; @@ -178,8 +181,8 @@ const prepareLocalChain = (zone, makeAccount) => * the query fails. Otherwise, return the response as a JSON-compatible * object. * - * @param {import('@agoric/cosmic-proto').TypedJson} request - * @returns {Promise<import('@agoric/cosmic-proto').TypedJson>} + * @param {TypedJson} request + * @returns {Promise<TypedJson>} */ async query(request) { const requests = harden([request]); @@ -197,11 +200,11 @@ const prepareLocalChain = (zone, makeAccount) => * system error, will return all results to indicate their success or * failure. * - * @param {import('@agoric/cosmic-proto').TypedJson[]} requests + * @param {TypedJson[]} requests * @returns {Promise< * { * error?: string; - * reply: import('@agoric/cosmic-proto').TypedJson; + * reply: TypedJson; * }[] * >} */

0xpatrickdev · 2024-05-08T18:14:40Z

packages/orchestration/src/exos/stakingAccountKit.js

+          ]);
+          trace('MsgTransfer result', result);
+
+          return /** @type {{ sequence: number }} */ (result);


Was expecting to see a bigint or string here. Surprised to see a number, but this is what I observed with local chains.

is that observed data captured in a unit test?

It's currently captured in packages/boot/tools/supports.ts: https://github.com/Agoric/agoric-sdk/pull/9342/files#diff-5f4f231637e90b9cdfecc29b28c80beb222607ad154bbc266cbf52c6a7c246e9R438-R446

I considered adding a new type helper to @agoric/cosmic-proto that's like JsonSafe, but wanted to be sure of this behavior before proceeding. @michaelfig - do you have any insights here? Are bigints cast to numbers for the LocalChainAccount implementation?

0xpatrickdev · 2024-05-08T18:15:32Z

packages/orchestration/src/exos/chainAccountKit.js

+        /**
+         * see stakingAccountKit.js for an example until #9212
+         * @param {Payment} payment
+         */


Given Scaling Considerations, we may want to reconsider including deposit() on ChainAccount.

0xpatrickdev · 2024-05-08T18:17:06Z

packages/vats/src/localchain.js

@@ -5,6 +5,8 @@ import { AmountShape } from '@agoric/ertp';

 const { Fail, bare } = assert;

+/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */


Suggested change

/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */

/** @import {TypedJson, Proto3Shape} from '@agoric/cosmic-proto'; */

i wouldn't worry about extra types for now. especially because ambients can make it look like imports that aren't used that actually are during build time

0xpatrickdev · 2024-05-08T18:18:04Z

packages/vm-config/decentral-devnet-config.json

+                "sourceChannelId": "channel-1",
+                "sourcePortId": "transfer"
+              },
+              "icqEnabled": false


set true for tests, where query responses are mocked, but won't work on a local simulated chain

0xpatrickdev · 2024-05-08T18:22:42Z

packages/boot/tools/supports.ts

@@ -295,6 +295,7 @@ export const makeSwingsetTestKit = async (

  let inbound;
  let ibcSequenceNonce = 0;
+  let icaExecuteTxSequence = 0;


Suggested change

let icaExecuteTxSequence = 0;

let lcaExecuteTxSequence = 0;

0xpatrickdev · 2024-05-08T18:25:52Z

packages/boot/test/bootstrapTests/test-orchestration.ts

@@ -215,3 +339,5 @@ test.serial('stakeAtom - smart wallet', async t => {
    'delegate fails with invalid validator',
  );
 });
+
+test.todo('deposit to LCA fails, payment should be returned');


Currently only simulating failed LCA -> ICA transfer. This captures the last nested block of the Deposit() try / catch, but we have nothing testing the first depositToSeat, which depends on await Object.values(payments)[0]. Suggestions on how to simulate this welcome.

dckc

time for a meeting; better share my notes so far

dckc · 2024-05-08T20:31:04Z

packages/ERTP/src/typeGuards.js

@@ -77,6 +77,11 @@ export const AmountShape = harden({
  value: AmountValueShape,
 });

+export const NatAmountShape = harden({
+  brand: BrandShape,
+  value: NatValueShape,


there's a NatValueShape? even though it's the same as M.nat()? odd.

dckc · 2024-05-08T20:32:21Z

packages/boot/test/bootstrapTests/test-orchestration.ts

@@ -1,3 +1,4 @@
+/* eslint-disable camelcase */


it's not obvious why camelcase should be disabled. is it easy to add a few words? no big deal for test code, I guess

dckc · 2024-05-08T20:35:09Z

packages/boot/test/bootstrapTests/test-orchestration.ts

@@ -150,6 +151,7 @@ test.serial('stakeAtom - smart wallet', async t => {
    'agoric1testStakAtom',
  );

+  // 1. Make Account


consider turning such comments into t.log() calls.

odd... makeAcountInvitationMaker? with 1 c? oops!

dckc · 2024-05-08T20:38:31Z

packages/boot/test/bootstrapTests/test-orchestration.ts

+    },
+    proposal: {
+      give: {
+        // @ts-expect-error BoardRemote is not assignable to Brand<any>


I/we really should fix that. the getBoardId() kludge is a bad idea. the mapping from value to slot (board id) belongs in a table external to the object; we learned this in makeClientMarshaller but we haven't back-ported it all the way to here.

IOU an issue, I guess.

dckc · 2024-05-08T20:42:06Z

packages/boot/test/bootstrapTests/test-orchestration.ts

+  });
+  t.like(wd.getLatestUpdateRecord(), {
+    status: { id: 'request-deposit-success', numWantsSatisfied: 1 },
+  });


Hm. I'd like to see a check that the money got there. Can we capture calls across the bridge and check for a suitable one here?

The best we can do in this environment is mock, which doesn't give the highest confidence or seem worth the cost. I am going to pursue better unit testing at the exo level per your suggestion to get more confidence around these failure cases

dckc · 2024-05-08T20:44:49Z

packages/boot/test/bootstrapTests/test-orchestration.ts

+      },
+    }),
+    {
+      message: 'Deposit failed, payment returned.',


again, it would be good to have evidence that the right vbank messages went over the bridge

ah... but we do have payouts below

dckc · 2024-05-08T20:47:59Z

packages/builders/scripts/orchestration/init-stakeAtom.js

    bondDenom = 'uatom',
+    bondDenomLocal = 'ibc/C4CFF46FD6DE35CA4CF4CE031E643C8FDC9BA4B99AE598E9B0ED98FE3A2319F9',


it would be nice to get this from hashing... or to test that this matches what we get from hashing

Agree. Something i think we'll tackle as part of #9211

dckc · 2024-05-08T20:51:27Z

packages/orchestration/src/examples/stakeAtom.contract.js

@@ -50,25 +73,40 @@ export const start = async (zcf, privateArgs, baggage) => {
    zcf,
  );

+  /** @type {BrandToIssuer} */
+  const brandToIssuer = zone.mapStore('brandToIssuer');


a durable store is over-kill for this. it costs a syscall for each access.

dckc · 2024-05-08T20:55:36Z

packages/orchestration/src/examples/stakeAtom.contract.js

+    const icqConnection = icqEnabled
+      ? await E(orchestration).provideICQConnection(controllerConnectionId)
+      : undefined;


Maybe it's ok by the new await safety style, but I've picked up an aversion to code that sometimes represents a turn boundary and sometimes doesn't.

Suggested change

const icqConnection = icqEnabled

? await E(orchestration).provideICQConnection(controllerConnectionId)

: undefined;

const icqConnection = await (icqEnabled

? E(orchestration).provideICQConnection(controllerConnectionId)

: undefined);

I'm also reminded to check for a clear commit point, as in the prepare / commit pattern.

If makeAccount() succeeds but provideICQConnection() fails, should we release the account?

turadg · 2024-05-08T20:44:41Z

packages/ERTP/src/typeGuards.js

@@ -77,6 +77,11 @@ export const AmountShape = harden({
  value: AmountValueShape,
 });

+export const NatAmountShape = harden({


Suggested change

export const NatAmountShape = harden({

/** @see {makeNatAmountShape} to match a specific brand */

export const NatAmountShape = harden({

turadg · 2024-05-08T20:51:37Z

packages/ERTP/src/typeGuards.js

  numerator: AmountShape,
  denominator: AmountShape,


Suggested change

numerator: AmountShape,

denominator: AmountShape,

numerator: NatAmountShape,

denominator: NatAmountShape,

turadg · 2024-05-08T20:52:35Z

packages/boot/test/bootstrapTests/test-orchestration.ts

+
+  // 2. Deposit to Account
+  await wd.executeOffer({
+    id: 'request-deposit-success',


success is independent of the offer identity

Suggested change

id: 'request-deposit-success',

id: 'request-deposit',

reading further I see you have other variants. Still I think success or failure are independent of identity. E.g.

request-deposit request-deposit-bad-want request-deposit-two-gives

I tripped over that name/id on 1st reading too.

turadg · 2024-05-08T20:59:46Z

packages/boot/test/bootstrapTests/test-orchestration.ts

+    status: { id: 'request-deposit-success', numWantsSatisfied: 1 },
+  });
+
+  await t.throwsAsync(


good to have these validation tests, but they don't need to integrate with smart-wallet. Spinning up a contract in isolation is a bit of work, I don't know that they even need to integrate with stakeAtom. Could these just be tests of the StakingAccountKit exo? Those would run very fast

Acknowledged and agree: #9342 (comment)

turadg · 2024-05-08T21:01:27Z

packages/orchestration/package.json

@@ -60,7 +61,8 @@
    },
    "files": [
      "test/**/*.test.js",
-      "test/**/*.test.ts"
+      "test/**/*.test.ts",
+      "test/**/test-*.js"


reminds me to revive #8653

turadg · 2024-05-08T21:11:31Z

packages/orchestration/src/exos/stakingAccountKit.js

+          // messages from the LCA (e.g., withdraw funds)
+          await E(localAccount).deposit(payment);
+
+          const timeoutTimestamp =


this is a conditional await. I think it's safe because of the await above it, but please add a // @jessie-check at the top to be sure. all modules that run in a vat should pass jessie-check

turadg · 2024-05-08T21:46:30Z

packages/orchestration/src/exos/stakingAccountKit.js

+          trace('Transferring funds to ICA');
+          /**
+           * // TODO can we infer `/ibc.applications.transfer.v1.MsgTransferResponse`?
+           * @type {unknown[]}


turadg · 2024-05-08T21:47:11Z

packages/orchestration/src/proposals/start-stakeAtom.js

+  const chainTimerService = await chainTimerServiceP;
+  const chainTimerBrand = await E(chainTimerService).getTimerBrand();


why not pipeline?

Suggested change

const chainTimerService = await chainTimerServiceP;

const chainTimerBrand = await E(chainTimerService).getTimerBrand();

const chainTimerBrand = await E(chainTimerServiceP).getTimerBrand();

turadg · 2024-05-08T21:47:44Z

packages/orchestration/src/types.d.ts

@@ -266,7 +267,7 @@ export interface ChainAccount {
    opts?: Partial<Omit<TxBody, 'messages'>>,
  ) => Promise<string>;
  /** deposit payment from zoe to the account*/
-  deposit: (payment: Payment) => Promise<void>;
+  deposit: (payment: Payment) => Promise<{ sequence: number }>;


@dtribble wants to review changes to this API

turadg · 2024-05-08T21:48:32Z

packages/vats/src/localchain.js

@@ -5,6 +5,8 @@ import { AmountShape } from '@agoric/ertp';

 const { Fail, bare } = assert;

+/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */


i wouldn't worry about extra types for now. especially because ambients can make it look like imports that aren't used that actually are during build time

dckc

I'm running out of steam for today.

But at least the bankManager change is critical.

dckc · 2024-05-08T22:44:51Z

packages/orchestration/src/exos/stakingAccountKit.js

+              this.state.brandToIssuer.has(giveAmount.brand) ||
+                Fail`${giveAmount.brand} not registered`;


Contracts can rely on Zoe to ensure that brands in proposals are registered.

dckc · 2024-05-08T22:47:48Z

packages/orchestration/src/exos/stakingAccountKit.js

+                  await depositToSeat(zcf, seat, give, payments);
+                  throw Fail`Deposit failed, payment returned.`;
+                } catch (error) {
+                  if (error.message.includes('not a live paymen')) {


Our style for errors and control flow is to not peek at errors.

Why do we care what sort of error it is?

dckc · 2024-05-08T22:50:22Z

packages/orchestration/src/exos/stakingAccountKit.js

+                  await depositToSeat(zcf, seat, give, payments);
+                  throw Fail`Deposit failed, payment returned.`;
+                } catch (error) {
+                  if (error.message.includes('not a live paymen')) {


Also, are we testing all these failure modes?

I wonder if it would help to hoist much of this logic to a function that we can unit test.

dckc · 2024-05-08T23:00:36Z

packages/orchestration/src/typeGuards.js

+/**
+ * - `give` allows any `Nat` `issuerKeyword` record. Must be exactly one entry.
+ * - `exit` must be `{ waived: null }`
+ * - `want` must be empty
+ */
+export const DepositProposalShape = M.splitRecord(


waived: null merits some explanation: once somebody makes such an offer, undoing it may be infeasible, so the client can't honor exit requests; clients must waive their right to exit.

docs should bear a WARNING notice about lack of offer safety

maybe the name should be scarier too... GiveAndHope :)

maybe these docs belong on the deposit() method and/or the Deposit invitationMaker.

dckc · 2024-05-08T23:03:45Z

packages/vats/src/localchain.js

+        const { address, powers } = this.state;
+        const bankManager = getPower(powers, 'bankManager');
+
+        const acctsBank = E(bankManager).getBankForAddress(address);


This getBankForAddress() call should happen before this account object is created. No account should have access to everybody else's purses. critical

(Didn't I give this feedback a while ago? hm.)

Refs: #9342 ## Description Make it so executeTx gives the proper response message for each request message. ### Security Considerations  ### Scaling Considerations  ### Documentation Considerations  ### Testing Considerations  ### Upgrade Considerations

refs: #9342, #9193 stacked on - #9353 ## Description Provide localchain accounts only with the `bank` for the relevant address, rather than giving them access to all accounts in the form of the `bankManager`. earlier discussion: https://github.com/Agoric/agoric-sdk/pull/9342/files#r1594791187 ### Security Considerations bankManager was excess authority ### Scaling / Documentation / Testing Considerations Nothing significant: no scaling changes; existing docs/tests suffice. ### Upgrade Considerations code is not released / deployed

0xpatrickdev added 5 commits May 8, 2024 13:49

fix(orchestration): test-* files are not picked up by ava

5936a2d

feat(localchain): add withdraw() to LocalChainAccount

aac8e1c

- refs: #9193

feat(orchestration): add DepositProposalShape

f22c3fd

- refs: 9193

feat(orchestration): add deposit() to stakingAccountKit

6088aa1

- uses a LocalChainAccount to accept an ERTP Payment and transfer it to an ICA account - uses ChainTimerService to create a timeout timestamp for ibc/MsgTransfer

fixup! feat(orchestration): add deposit() to stakingAccountKit

37083e6

0xpatrickdev requested review from turadg and dckc May 8, 2024 18:10

0xpatrickdev commented May 8, 2024

View reviewed changes

dckc reviewed May 8, 2024

View reviewed changes

turadg reviewed May 8, 2024

View reviewed changes

turadg mentioned this pull request May 8, 2024

proto msg response types #9351

Merged

dckc requested changes May 8, 2024

View reviewed changes

erights self-requested a review May 8, 2024 23:59

0xpatrickdev marked this pull request as draft May 9, 2024 17:35

dckc mentioned this pull request May 9, 2024

fix(vats): don't give all of bankManager to 1 account #9354

Merged

0xpatrickdev closed this Jun 6, 2024

		@@ -5,6 +5,8 @@ import { AmountShape } from '@agoric/ertp';

		const { Fail, bare } = assert;

		/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */

	/** @import {TypedJson, ResponseType, Proto3Shape} from '@agoric/cosmic-proto'; */
	/** @import {TypedJson, Proto3Shape} from '@agoric/cosmic-proto'; */

		bondDenom = 'uatom',
		bondDenomLocal = 'ibc/C4CFF46FD6DE35CA4CF4CE031E643C8FDC9BA4B99AE598E9B0ED98FE3A2319F9',

	export const NatAmountShape = harden({
	/** @see {makeNatAmountShape} to match a specific brand */
	export const NatAmountShape = harden({

		const chainTimerService = await chainTimerServiceP;
		const chainTimerBrand = await E(chainTimerService).getTimerBrand();

		this.state.brandToIssuer.has(giveAmount.brand) \|\|
		Fail`${giveAmount.brand} not registered`;

feat(orchestration): deposit ERTP payment to ICA #9342

feat(orchestration): deposit ERTP payment to ICA #9342

Conversation

0xpatrickdev commented May 8, 2024 • edited Loading

Description

Security Considerations

Scaling Considerations

Documentation Considerations

Testing Considerations

Upgrade Considerations

cloudflare-workers-and-pages bot commented May 8, 2024 • edited Loading

Deploying agoric-sdk with Cloudflare Pages

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

0xpatrickdev May 8, 2024 • 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

dckc 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

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

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

dckc 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

0xpatrickdev commented May 8, 2024 •

edited

Loading

cloudflare-workers-and-pages bot commented May 8, 2024 •

edited

Loading

0xpatrickdev May 8, 2024 •

edited

Loading