register any chain info #9452

turadg · 2024-06-04T22:13:49Z

refs: #9187

Description

Progress on #9187

Plus incidental improvements

Security Considerations

Scaling Considerations

Documentation Considerations

Testing Considerations

Upgrade Considerations

cloudflare-workers-and-pages · 2024-06-04T22:13:51Z

Deploying agoric-sdk with Cloudflare Pages

Latest commit:	`6d597fd`
Status:	✅ Deploy successful!
Preview URL:	https://88827a70.agoric-sdk.pages.dev
Branch Preview URL:	https://9187-any-chaininfo.agoric-sdk.pages.dev

View logs

dckc

oops; I started reviewing without noticing the DRAFT status

dckc · 2024-06-04T22:19:32Z

packages/orchestration/src/chain-info.js

@@ -0,0 +1,61 @@
+/**
+ * @file static declaration of known chain types will allow type support for


I'd expect this as a test fixture, but not a source module.

This info will vary between a3p, devnet, mainnet, etc.

oh. weird. this is just a .ts->.js rename. it was already here.

yeah, I figure it's all prototype code still. though the .ts was less risky to have in src because it would never show up in a bundle.

can we leave this here until #8879? I'll put a TODO

dckc · 2024-06-04T22:21:24Z

packages/orchestration/src/cosmos-api.ts

+  /**
+   * cf https://github.com/cosmos/chain-registry/blob/master/chain.schema.json#L117
+   */
+  stakingTokens?: [{ denom: string }];


That syntax gave me a strange error about [X] vs. X[]. This made it go away:

Suggested change

stakingTokens?: [{ denom: string }];

stakingTokens?: Array<{ denom: string }>;

I imagine that's your linter. That is a better type though, because what's here is a tuple rather than an array of arbitrary length. I'll change it.

dckc · 2024-06-04T22:22:44Z

packages/orchestration/src/facade.js

 */

 /** @type {any} */
 const anyVal = null;

 // FIXME should be configurable
+// ??? could the localchain service offer this?
 const mockLocalChainInfo = {


mockX looks like test code, not src code.

it's not test code either; it's an incomplete implementation. just like all the "'fixme'" strings we've had to date

I've put issues at all the fixmes to be sure they'll be addressed

dckc · 2024-06-04T22:31:13Z

packages/orchestration/src/facade.js

-        deposit(payment) {
+        async deposit(payment) {


orthogonal to this PR: In asyncFlow discussion yesterday, somebody said the facade methods can't be async.

dckc · 2024-06-04T22:32:50Z

packages/orchestration/src/facade.js

+ * @template {CosmosChainInfo} CCI
+ * @param {CCI} chainInfo


sorry; no... it just hand't occurred to me that a type could be parameterized by a whole structure, and not just a primitive value

dckc · 2024-06-04T22:35:50Z

packages/orchestration/src/facade.js

+          const chainInfo = chainInfos.has(name)
+            ? chainInfos.get(name)
+            : // @ts-expect-error may be undefined
+              wellKnownChainInfo[name];


well known chain should come from agoricNames, yes?

eventually, yes. that's #9063. I'll add a TODO

dckc · 2024-06-04T22:38:26Z

packages/orchestration/src/orchestration-api.ts

+  CosmosChainAccountMethods,
+  CosmosChainInfo,
+  IBCMsgTransferOptions,


This file shouldn't depend on cosmos types, should it?

ideally not, but it already does. I see that as independent cleanup

dckc · 2024-06-04T22:41:25Z

packages/orchestration/test/facade.test.ts

+  const result = (await handle()) as Chain<any>;
+  t.deepEqual(await result.getChainInfo(), mockChainInfo);
+
+  // FIXME allow re-registering. perhaps this should be called `provide`.


init / register seems like the right thing to me; I wouldn't expect to be able to re-register.

PTAAL; I pushed a change after you started reviewing

dckc · 2024-06-04T22:44:20Z

packages/orchestration/test/facade.test.ts

+    message: '"chainInfos" already registered: "mock"',
+  });
+
+  // FIXME support running again in the same zone (as will happen after an upgrade)


FIXME when? oops... I started reviewing without noticing the DRAFT status.

I requested it, thank you. I think it's okay to land FIXME comments in master until we're ready to publish the package

dckc · 2024-06-04T22:44:44Z

packages/orchestration/test/facade.test.ts

+      }),
+    {
+      message:
+        'key "chainInfos" has already been used in this zone and incarnation',


darn it; I thought zone.mapStore('chainInfos', ...) already had provide semantics.

dckc · 2024-06-04T22:48:45Z

packages/orchestration/src/facade.js

-  const chainInfos = zone.mapStore('chainInfos', {
+  const chainInfos = makeScalarBigMapStore('chainInfos', {


did you disconnect it from the zone on purpose?

yes, to reduce state to maintain across upgrades. however the contract registered the chainInfo the first time, it can do again. and in many cases it won't want to because it'll have made it into the upstream wellKnown registry

Co-Authored-By: Dan Connolly <dckc@madmode.com>

dckc

I'm surprised that the way the test makes zcf passes lint. But I suppose it doesn't matter.

dckc · 2024-06-05T15:01:42Z

packages/orchestration/src/cosmos-api.ts

+    ? IcaAccount
+    : {}) &
+    CCI extends {
+    stakingTokens: {};


does it matter that stakingTokens is an Array? should this be...

Suggested change

stakingTokens: {};

stakingTokens: [];

Is an array accurate? I know this is how chain-registry has it modeled, but I am skeptical. The bond_denom field in x/staking params seems to indicate just a single entry: https://buf.build/cosmos/cosmos-sdk/docs/main:cosmos.staking.v1beta1#cosmos.staking.v1beta1.Params

Yes, it is curious. https://github.com/search?q=repo%3Acosmology-tech%2Fchain-registry+stakingTokens shows only examples of a single element in the array.

I consider this out of scope that will be resolved in already scheduled issues.

dckc · 2024-06-05T15:08:02Z

packages/orchestration/test/facade.test.ts

+
+  const zone = bootstrap.rootZone;
+
+  const { zcf } = await setUpZoeForTest();


does setUpZoeForTest make a zcf? I get an error when I try this (in another context).

good catch. I guess the zcf isn't used in this test. I've changed it to get a working one though,

const { zcf } = await setupZCFTest();

But I don't know why TS didn't catch this. It typed zcf as any. It goes away with noImplicitAny: true but that causes many other linting errors

dckc · 2024-06-05T15:11:45Z

packages/orchestration/test/facade.test.ts

+  const handle = orchestrate('mock', {}, async orc => {
+    return orc.getChain('mock');


thinking out loud: passing the chain object outside the context of the orchestrate callback seems odd somehow. Is that consistent with the asyncFlow rules? (where do we find those rules?)

IDK. But if it isn't allowed I'd like to see a test where this would break.

0xpatrickdev

LGTM. It seems like #8879 is out of scope for the follow up PR(s) for #9187, but bringing this mocked file to your attention to maybe update in the course of your efforts:
https://github.com/Agoric/agoric-sdk/blob/5659362b247593793bd30a6b726b261587701487/packages/orchestration/src/utils/mockChainInfo.js

0xpatrickdev · 2024-06-05T15:53:49Z

packages/orchestration/src/chain-info.js

+        allowedMessages: [],
+        allowedQueries: [],


Are you planning to use these for generating Orchestration Account facets? Regardless, looking forward to seeing the approach for that in the next iteration.

0xpatrickdev · 2024-06-05T15:56:27Z

packages/orchestration/src/facade.js

+      const [{ denom: bondDenom }] = chainInfo.stakingTokens || [
+        {
+          denom: null,
+        },
+      ];
+      assert(bondDenom, 'missing bondDenom');


Mentioned here - but i don't expect we'll see an array of staking staking tokens for anyone using cosmos.staking.v1beta1

## Description In a recent discussion we noticed that the test wasn't showing a type error where it should have: #9452 (comment) It would have with `noImplicitAny: false` but that pulls in way too many other errors. While investigating I noticed some other things that could be cleaned up and tackled those here. The remaining blocker for `skipLibCheck: false` is the Telescope output in `dist`. The .js files have `@ts-nocheck` but the .d.ts don't. Surprising but not worth more effort. ### Security Considerations n/a, types ### Scaling Considerations n/a, types ### Documentation Considerations none ### Testing Considerations CI ### Upgrade Considerations Potential cherry-pick complexities but we're moving to release based on master

turadg requested review from dckc and 0xpatrickdev June 4, 2024 22:14

dckc reviewed Jun 4, 2024

View reviewed changes

turadg and others added 4 commits June 5, 2024 07:48

feat(typeGuards): CosmosChainInfoShape

183c072

Co-Authored-By: Dan Connolly <dckc@madmode.com>

feat(typeGuards): DelegationShape

f7ed627

Co-Authored-By: Dan Connolly <dckc@madmode.com>

test: clarify intentional type errors

303318d

chore: include stakingTokens denom in CosmosChainInfo

8c6c757

turadg force-pushed the 9187-any-chainInfo branch from adcc359 to fd70556 Compare June 5, 2024 14:52

turadg requested a review from dckc June 5, 2024 14:53

turadg marked this pull request as ready for review June 5, 2024 14:53

dckc approved these changes Jun 5, 2024

View reviewed changes

0xpatrickdev approved these changes Jun 5, 2024

View reviewed changes

turadg added 4 commits June 5, 2024 09:25

feat: registerChain

9d38e2e

chore(types): Chain parameterized by ChainInfo

bdd68cf

fix: get bondDenom from chainInfo

82115f1

chore: facade reconciliation

6d597fd

turadg force-pushed the 9187-any-chainInfo branch from fd70556 to 6d597fd Compare June 5, 2024 16:26

turadg added the automerge:rebase Automatically rebase updates, then merge label Jun 5, 2024

mergify bot merged commit 21a664a into master Jun 5, 2024
66 checks passed

mergify bot deleted the 9187-any-chainInfo branch June 5, 2024 17:00

turadg mentioned this pull request Jun 5, 2024

fix tsc errors detected by skipLibCheck: false #9457

Merged

		@@ -0,0 +1,61 @@
		/**
		* @file static declaration of known chain types will allow type support for

	stakingTokens?: [{ denom: string }];
	stakingTokens?: Array<{ denom: string }>;

		const chainInfos = zone.mapStore('chainInfos', {
		const chainInfos = makeScalarBigMapStore('chainInfos', {


		const zone = bootstrap.rootZone;

		const { zcf } = await setUpZoeForTest();

		const handle = orchestrate('mock', {}, async orc => {
		return orc.getChain('mock');

register any chain info #9452

register any chain info #9452

Conversation

turadg commented Jun 4, 2024 • edited Loading

Description

Security Considerations

Scaling Considerations

Documentation Considerations

Testing Considerations

Upgrade Considerations

cloudflare-workers-and-pages bot commented Jun 4, 2024 • edited Loading

Deploying agoric-sdk with Cloudflare Pages

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

turadg Jun 5, 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

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

Choose a reason for hiding this comment

turadg Jun 5, 2024 • edited Loading

Choose a reason for hiding this comment

0xpatrickdev left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

turadg commented Jun 4, 2024 •

edited

Loading

cloudflare-workers-and-pages bot commented Jun 4, 2024 •

edited

Loading

turadg Jun 5, 2024 •

edited

Loading

turadg Jun 5, 2024 •

edited

Loading