Feature/nicaea runtime upgrade#782
Conversation
…d incompatible proposals form nicaea, removed nicaea-compatible proposals from constantinople
…me-joystream into feature/working-group-rejection
…l-gadelshin/substrate-runtime-joystream into feature/working-group-rejection
…/substrate-runtime-joystream into feature/working-group-rejection
|
I'm currently getting some build errors: I think for the first error you we are missing the and in the second case, with the new types library we should remove the |
|
There are some other build errors probably worth looking into as well, perhaps some indication of incorrect versions of @types/mocha and @types/jest with the version of typescript we are using? |
mnaamani
left a comment
There was a problem hiding this comment.
I've only reviewed the ApiWrapper code so far in nicaea. Will continue review of the actual scenarios soon.
I noticed in most places when dealing with Id types, like ApplicationId, WorkerId etc. . you are just using BN instead. Although those types extend u64 which extends BN it is probably better to be type safe and use the real underlying type as defined in the types library. As there a particular reason you did this, and is there any potential problems with this approach?
| return this.sender.signAndSend(this.api.tx.sudo.sudo(this.api.tx.storageWorkingGroup.unsetLead()), sudo, false); | ||
| } | ||
|
|
||
| public async addOpening( |
There was a problem hiding this comment.
when you have such a long list of arguments, I highly recommend passing them in as an object so you can have named arguments. This would be very hard to debug if arguments were passed in the wrong order.
There was a problem hiding this comment.
Done, created in dto/workingGroupOpening.ts
|
I was able to run the tests successfully so so far so good, obviously there will be alot of updates to the scenarios when you update from nicaea branch so will re test when ready. |
…me-joystream into feature/nicaea-runtime-upgrade
|
@mnaamani Does the approach to use |
|
I have to further investigate errors connected with |
It does to some extent yes, although the u32, u64, u128 all extend BN, I think when you loose the exact type information when constructing transaction arguments, it can lead to incorrect encoding of the transaction. So I'm more concerned with that aspect. using primitive javascript number would be safer than just plain BN. But we might as well use the "exact" type. We did have a brief discussion on this with @Lezek123 in #742 (comment) Perhaps i'm being too paranoid, and if you pass a BN as an argument in a |
|
From the linting failure in the CI checks, you will most likely need to checkin the yarn.lock file:
|
|
The |
Co-authored-by: Mokhtar Naamani <mokhtar.naamani@gmail.com>
|
All tests passing 👍 |
Runtime upgrade test from Constantinople to Nicaea.
Includes #769