Oracle price fetching refactoring #512
Conversation
| 'ETH-A': { | ||
| nextUnitPrice: '1712.2106886', | ||
| currentUnitPrice: '1712.2106886', | ||
| nextPriceChange: new Date('2022-09-09T10:00:00.000Z'), | ||
| nextUnitPrice: '1602.1803800999999', | ||
| currentUnitPrice: '1602.1803800999999', | ||
| nextPriceChange: new Date('2022-09-14T13:00:00.000Z'), | ||
| }, | ||
| 'ETH-B': { | ||
| nextUnitPrice: '1602.1803800999999', | ||
| currentUnitPrice: '1602.1803800999999', | ||
| nextPriceChange: new Date('2022-09-14T13:00:00.000Z'), | ||
| }, | ||
| 'ETH-C': { | ||
| nextUnitPrice: '1208.0159951', | ||
| currentUnitPrice: '1227.067888375', | ||
| nextPriceChange: new Date('2022-06-13T11:00:00.000Z'), | ||
| nextUnitPrice: '1602.1803800999999', | ||
| currentUnitPrice: '1602.1803800999999', | ||
| nextPriceChange: new Date('2022-09-14T13:00:00.000Z'), |
There was a problem hiding this comment.
@KirillDogadin-std Why was those values different? Those collateral types has the same oracle (confirmed via calling MCD_SPOT.ilks), so it should return the same values. Moreover, executing just one test passes as expected:
npm run test -- --grep "Sound values are extracted"
Any idea on which of the tests introduces side effects/modifies them?
There was a problem hiding this comment.
within several minutes: i did not find the source of side effect:
- checked if we
pokethe oracle - does not seem so - checked if we overwrite the price - does not seem so
- checked if there's additional modifier for determination of unitPrice - does not seem source
I would have to look longer/deeper to determine the reason
There was a problem hiding this comment.
Thanks, I think it might be because something else is executed at the time of this test, block number changed and it somehow affects the numbers (but I still can't make sense of it). Will also look into it
core/src/constants/COLLATERALS.ts
Outdated
| { | ||
| slotAddress: '0x3', | ||
| format: ['uint128 isCurrentUnitPriceValid', 'uint128 currentUnitPrice'], | ||
| }, | ||
| { | ||
| slotAddress: '0x4', | ||
| format: ['uint128 isNextUnitPriceValid', 'uint128 nextUnitPrice'], | ||
| }, |
There was a problem hiding this comment.
@KirillDogadin-std What do you think about this format?
There was a problem hiding this comment.
just from looking at the selected code piece i did not figure out why you need format / for what purpose it's used.
Nothing against the datastructure, but perhaps a clearer key name can be an option?
There was a problem hiding this comment.
It's used by ABI coder. It's basically a type + variable name. See examples on how hex being decoded here https://docs.ethers.io/v5/api/utils/abi/coder. Basically it allows us to define all those variables in unified format. But I might change it to better suit overwrite functionality that haven't yet addressed
There was a problem hiding this comment.
if it remains, i would go with something like abiCoderFormat
Closes #513
Checklist:
#)