-
-
Notifications
You must be signed in to change notification settings - Fork 186
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[token-detection-controller] Refactor detectTokens
method
#3938
[token-detection-controller] Refactor detectTokens
method
#3938
Conversation
detectTokens
method
0aed206
to
6fc4688
Compare
a955a1c
to
d4581a9
Compare
6fc4688
to
f2faf4f
Compare
8aaeb9c
to
aee4d13
Compare
4fdd841
to
5fee94d
Compare
aee4d13
to
cf2d53f
Compare
5fee94d
to
c91ff00
Compare
#chainIdAgainstWhichToDetect: Hex; | ||
|
||
#addressAgainstWhichToDetect: string; | ||
|
||
#networkClientIdAgainstWhichToDetect: NetworkClientId; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should these be maintained separately from #chainId
, #selectedAddress
, #networkClientId
? My assumption is that we want the detectTokens
method to be callable independently of the polling/passive detection performed by the controller and messenger. At least, that seems to be what the current implementation's avoidance of class fields in detectTokens
suggests.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You're right that consumers should be able to call detectTokens
independently, and it should just work.
It makes sense that we would cache the currently selected network as well as the currently selected account. I don't think we need to cache the chain ID, because we can always ask the network client. (But removing that is for another PR.)
As for these three variables, though, it seems to me that we've created them so that we can break up detectTokens
into pieces. But, why not pass along these values, then? It seems that this would make it easier for someone to read the code because there would be fewer things for them to track.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Agree that a pure style that avoids using class fields might be more readable here.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Removed the three unnecessary class fields here: ab0e3eb
(#3938)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also removed #chainId
: 12ba664
(#3938)
Token, | ||
'aggregators' | 'image' | 'balanceError' | 'isERC721' | ||
> & { | ||
type LegacyToken = { | ||
name: string; | ||
logo: string; | ||
logo: `${string}.svg`; | ||
symbol: string; | ||
decimals: number; | ||
erc20?: boolean; | ||
erc721?: boolean; | ||
}; | ||
|
||
type TokenDetectionMap = { | ||
[P in keyof TokenListMap]: Omit<TokenListMap[P], 'occurrences'>; | ||
}; | ||
|
||
export const STATIC_MAINNET_TOKEN_LIST = Object.entries<LegacyToken>( | ||
contractMap, | ||
).reduce< | ||
Record< | ||
string, | ||
Partial<TokenListToken> & Pick<Token, 'address' | 'symbol' | 'decimals'> | ||
> | ||
>((acc, [base, contract]) => { | ||
const { logo, ...tokenMetadata } = contract; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Better typing and handling of legacy token list.
await safelyExecute(async () => { | ||
const balances = await this.#getBalancesInSingleCall( | ||
this.#addressAgainstWhichToDetect, | ||
tokensSlice, | ||
this.#networkClientIdAgainstWhichToDetect, | ||
); | ||
|
||
const tokensWithBalance: Token[] = []; | ||
const eventTokensDetails: string[] = []; | ||
for (const nonZeroTokenAddress of Object.keys(balances)) { | ||
const { decimals, symbol, aggregators, iconUrl, name } = | ||
this.#tokenList[nonZeroTokenAddress]; | ||
eventTokensDetails.push(`${symbol} - ${nonZeroTokenAddress}`); | ||
tokensWithBalance.push({ | ||
address: nonZeroTokenAddress, | ||
decimals, | ||
symbol, | ||
aggregators, | ||
image: iconUrl, | ||
isERC721: false, | ||
name, | ||
}); | ||
} | ||
|
||
await safelyExecute(async () => { | ||
const balances = await this.#getBalancesInSingleCall( | ||
addressAgainstWhichToDetect, | ||
tokensSlice, | ||
networkClientIdAgainstWhichToDetect, | ||
if (tokensWithBalance.length) { | ||
this.#trackMetaMetricsEvent({ | ||
event: 'Token Detected', | ||
category: 'Wallet', | ||
properties: { | ||
tokens: eventTokensDetails, | ||
token_standard: 'ERC20', | ||
asset_type: 'TOKEN', | ||
}, | ||
}); | ||
|
||
await this.messagingSystem.call( | ||
'TokensController:addDetectedTokens', | ||
tokensWithBalance, | ||
{ | ||
selectedAddress: this.#addressAgainstWhichToDetect, | ||
chainId: this.#chainIdAgainstWhichToDetect, | ||
}, | ||
); | ||
const tokensWithBalance: Token[] = []; | ||
const eventTokensDetails: string[] = []; | ||
for (const nonZeroTokenAddress of Object.keys(balances)) { | ||
const { decimals, symbol, aggregators, iconUrl, name } = | ||
tokenListUsed[nonZeroTokenAddress]; | ||
eventTokensDetails.push(`${symbol} - ${nonZeroTokenAddress}`); | ||
tokensWithBalance.push({ | ||
address: nonZeroTokenAddress, | ||
decimals, | ||
symbol, | ||
aggregators, | ||
image: iconUrl, | ||
isERC721: false, | ||
name, | ||
}); | ||
} | ||
if (tokensWithBalance.length) { | ||
this.#trackMetaMetricsEvent({ | ||
event: 'Token Detected', | ||
category: 'Wallet', | ||
properties: { | ||
tokens: eventTokensDetails, | ||
token_standard: 'ERC20', | ||
asset_type: 'TOKEN', | ||
}, | ||
}); | ||
await this.messagingSystem.call( | ||
'TokensController:addDetectedTokens', | ||
tokensWithBalance, | ||
{ | ||
selectedAddress: addressAgainstWhichToDetect, | ||
chainId: chainIdAgainstWhichToDetect, | ||
}, | ||
); | ||
} | ||
}); | ||
} | ||
} | ||
}); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Large diff, but no actual changes here other than being moved to the new #addDetectedTokens
method, and variable name updates (this.#{address,networkClientId,chainId}AgainstWhichToDetect
, this.#tokenList
).
8a73dde
to
e9bbf1d
Compare
606211e
to
70f1a26
Compare
const { selectedNetworkClientId } = this.messagingSystem.call( | ||
'NetworkController:getState', | ||
); | ||
const newNetworkClientId = this.messagingSystem.call( | ||
'NetworkController:findNetworkClientIdByChainId', | ||
chainId, | ||
const { | ||
configuration: { chainId }, | ||
} = this.messagingSystem.call( | ||
'NetworkController:getNetworkClientById', | ||
selectedNetworkClientId, | ||
); | ||
return { | ||
chainId, | ||
networkClientId: newNetworkClientId, | ||
networkClientId: selectedNetworkClientId, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Replaces getProviderConfig
, findNetworkClientIdByChainId
actions with getState
, getNetworkClientById
I used getNetworkClientById
instead of getNetworkConfigurationByNetworkClientId
to avoid having to use this.#chainId
as a fallback value for the case where the returned network configuration is undefined.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bergeron @jiexi @adonesky1 Thoughts on this approach for retrieving networkClientId
, chainId
? c75fb3b
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That approach looks okay to me, but wondering if you could get both in 1 call:
const {
selectedNetworkClientId,
providerConfig: { chainId },
} = this.messagingSystem.call('NetworkController:getState');
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me as long as there's no edge case where the selected network client's chainId
is inconsistent with providerConfig
. @mcmire What do you think?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Applied here: d6caee2
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That makes sense. We might as well do the same thing for lines 418-428, huh?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm a little confused by that section. For detectTokens
to work independently of the class state, shouldn't the user be able to pass in a network client ID that's not the currently selected network's and get the corresponding chain ID?
It seems like 418-428 is handling this case where the user-supplied network client ID is the source of truth, instead of the selectedNetworkClientId
in NetworkState.
Otherwise, if the only network client ID ever used in detectTokens
is the selectedNetworkClientId
from NetworkState, shouldn't we remove the networkClientId
params from detectTokens
, restartTokenDetection
, and getCorrect...
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh, sorry. What I meant was that it looks like further up we are using getNetworkConfigurationByNetworkClientId
to get the configuration for the given network client, then we are checking to see whether we have a configuration, then we are getting the chain ID from the configuration. But in your changes here you are using getNetworkClientById
. So do we want to be consistent? I'm thinking this:
#getCorrectChainIdAndNetworkClientId(networkClientId?: NetworkClientId): {
chainId: Hex;
networkClientId: NetworkClientId;
} {
if (networkClientId) {
const {
configuration: { chainId },
} = this.messagingSystem.call(
'NetworkController:getNetworkClientById',
networkClientId,
);
return {
chainId,
networkClientId,
};
}
const { selectedNetworkClientId } = this.messagingSystem.call(
'NetworkController:getState',
);
const {
configuration: { chainId },
} = this.messagingSystem.call(
'NetworkController:getNetworkClientById',
selectedNetworkClientId,
);
return {
chainId,
networkClientId: selectedNetworkClientId,
};
}
And if we can do this, then we can flip it around:
#getCorrectChainIdAndNetworkClientId(givenNetworkClientId?: NetworkClientId): {
chainId: Hex;
networkClientId: NetworkClientId;
} {
const networkClientId =
givenNetworkClientId ??
this.messagingSystem.call(
'NetworkController:getState',
).selectedNetworkClientId;
const {
configuration: { chainId },
} = this.messagingSystem.call(
'NetworkController:getNetworkClientById',
networkClientId,
);
return { chainId, networkClientId };
}
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The reason I made that distinction was that getNetworkConfigurationByNetworkClientId
returns undefined
if there's no network configuration for the given network client ID.
So by using that first instead of getNetworkClientById
(which never returns undefined), I thought we'd be able to rule out any user-supplied networkClientId
that's not registered in the network state. If that's the case, we default to the selected network client ID.
But if that's not a meaningful edge case then the second option looks good I can apply it rn.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh I see what you're saying. Okay, that makes sense!
…dates (#3923) ## Explanation As a preparatory step for fully replacing the extension `DetectTokensController` with the consolidated core repo `TokenDetectionController`, `TokenDetectionController` needs to be updated with changes made to the extension `DetectTokensController` since #1813 was closed. #### Diff of `DetectTokensController` state - [MetaMask/metamask-extension@`5d285f7be5f7be981995dfa725aad97d81cc990a..85cd1c89039e900b452edb704ec37e9ccbd3e76a`#diff-323d0cf464](https://github.com/MetaMask/metamask-extension/compare/5d285f7be5f7be981995dfa725aad97d81cc990a..85cd1c89039e900b452edb704ec37e9ccbd3e76a#diff-323d0cf46498be3850b971474905354ea5ccf7fa13745ad1e6eba59c5b586830) ### Differences from extension `DetectTokensController` - Refactors logic for retrieving `chainId`, `networkClientId` into `this.#getCorrectChainIdAndNetworkClientId` - Uses `getNetworkConfigurationByNetworkClientId` action instead of `getNetworkClientById` to retrieve `chainId`. - If `networkClientId` is not supplied to the method, or it's supplied but `getNetworkConfigurationByNetworkClientId` returns `undefined`, finds `chainId` from `providerConfig`. - `detectTokens` replaces `detectNewTokens` - `detectTokens` accepts options object `{ selectedAddress, networkClientId }` instead of `{ selectedAddress, chainId, networkClientId }`. - Does not throw error if `getBalancesInSingleCall` fails. Also does not exit early -- continues looping. - Passes lists of full `Token` types to `TokensController:addDetectedTokens` instead of objects containing only `{ address, decimals, symbol }`. - `#trackMetaMetricsEvents` is a private method instead of protected. - Passes string literals instead of extension shared constants into `_trackMetaMetricsEvent`. ## References - Partially implements #3916 - Blocking #3918 - Changes adopted from: - MetaMask/metamask-extension#22898 - MetaMask/metamask-extension#22814 - #3914 - MetaMask/metamask-extension#21437 - Blocking (Followed by) #3938 ## Changelog ### [`@metamask/assets-controllers`](https://github.com/MetaMask/core/pull/3923/files#diff-ee47d03d53776b8dd530799a8047f5e32e36e35765620aeb50b294adc3339fab) ## Checklist - [x] I've updated the test suite for new or updated code as appropriate - [x] I've updated documentation (JSDoc, Markdown, etc.) for new or updated code as appropriate - [x] I've highlighted breaking changes using the "BREAKING" category above as appropriate --------- Co-authored-by: jiexi <jiexiluan@gmail.com>
…IdAndNetworkClientId`, `#getTokenListAndSlicesOfTokensToDetect`, `#addDetectedTokens` - Fixes #1614 - Maintains distinction between private class fields `#{chainId,selectedAddress,networkClientId}{,AgainstWhichToDetect}`, so that `detectTokens` method can be used independently of polling/passive detection (convert to static method?). Is this expected behavior?
…yChainId` actions with `getState`, `getNetworkClientById`
70f1a26
to
a09fcda
Compare
- Fixes #3661 Co-authored-by: Brian Bergeron <brian.e.bergeron@gmail.com>
f5c3194
to
f060ced
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Makes sense overall, but I had some comments.
NetworkClientId, | ||
} from '@metamask/network-controller'; | ||
import { defaultState as defaultNetworkState } from '@metamask/network-controller'; | ||
import type { AutoManagedNetworkClient } from '@metamask/network-controller/src/create-auto-managed-network-client'; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Heads up: #3998 will make it impossible to import subpaths, so this will need to be changed when that happens, but is okay for now.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes I was surprised to learn that AutoManagedNetworkClient
isn't a package-level export. Is this an intentional choice? Otherwise I could create a ticket for exporting this.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It wasn't. I don't think I expected anyone to have to use this type, as it's low-level. But there's no harm in exporting it or anything.
#chainIdAgainstWhichToDetect: Hex; | ||
|
||
#addressAgainstWhichToDetect: string; | ||
|
||
#networkClientIdAgainstWhichToDetect: NetworkClientId; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You're right that consumers should be able to call detectTokens
independently, and it should just work.
It makes sense that we would cache the currently selected network as well as the currently selected account. I don't think we need to cache the chain ID, because we can always ask the network client. (But removing that is for another PR.)
As for these three variables, though, it seems to me that we've created them so that we can break up detectTokens
into pieces. But, why not pass along these values, then? It seems that this would make it easier for someone to read the code because there would be fewer things for them to track.
const { chainId } = | ||
this.#addressAgainstWhichToDetect = this.#selectedAddress; | ||
|
||
const { chainId, networkClientId: correctNetworkClientId } = | ||
this.#getCorrectChainIdAndNetworkClientId(networkClientId); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
With your changes to #getCorrectChainIdAndNetworkClientId
, it will retrieve the currently selected network from NetworkController. So does the constructor need to take a networkClientId
? It seems like we don't need to pass it here.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm looks like this is also related to #3938 (comment). If we allow #getCorrectChainIdAndNetworkClientId
to accept a non-selected network client id and then return its chain id, then keeping the networkClientId
constructor option would also allow token-detection-controller's internal state to diverge from network-controller's.
Removing this will be fairly disruptive for the tests, but it looks like it'll be necessary.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Applied here: d0f6f69
(#3938)
const { selectedNetworkClientId } = this.messagingSystem.call( | ||
'NetworkController:getState', | ||
); | ||
const newNetworkClientId = this.messagingSystem.call( | ||
'NetworkController:findNetworkClientIdByChainId', | ||
chainId, | ||
const { | ||
configuration: { chainId }, | ||
} = this.messagingSystem.call( | ||
'NetworkController:getNetworkClientById', | ||
selectedNetworkClientId, | ||
); | ||
return { | ||
chainId, | ||
networkClientId: newNetworkClientId, | ||
networkClientId: selectedNetworkClientId, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That makes sense. We might as well do the same thing for lines 418-428, huh?
Co-authored-by: Elliot Winkler <elliot.winkler@gmail.com>
… class field is always initialized with `selectedNetworkClientId` property of network-controller state
…gainstWhichToDetect` with input params
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good to me!
Explanation
Adds refactors and cosmetic, typing fixes to #3923.
TokenDetectionController
'sdetectTokens
method:#getCorrectChainIdAndNetworkClientId
#getTokenListAndSlicesOfTokensToDetect
#addDetectedTokens
#selectedAddress
,#networkClientId
and corresponding parameters used indetectTokens
and its helper methods, so thatdetectTokens
method can be used independently of polling/passive detection.#getCorrectChainIdAndNetworkClientId
to removefindNetworkClientIdByChainId
which might return inconsistent/unexpected results, and replace it withgetState
,getNetworkClientById
NetworkClientId
LegacyToken
,TokenDetectionMap
#chainId
class field.References
DetectTokensController
updates #3923TokenDetectionController
up-to-date and release new version ofassets-controllers
#3916Changelog
Checklist