feat: decode metamask transaction error

[ariesgun]

Resolves #346

[ubiquity-os-deployer]

eb07da2

2eab98d

62f9e74

980d99f

[github-actions]

Preview Deployment
980d99f1669d9522c9402cb0211041e658f2e715

[ariesgun]

QA:

[Keyrxng]

Looks good. Idk if it's worth it to also log the onchain function name that was called too using the 4byte api .. @rndquu rfc

I think if we can make this robust and re-useable @ariesgun then we can open a task and implement it in rpc-handler as core error handling. That might be a task you'd like to complete if so.

[rndquu]

This QA has some drawbacks:

1. Result is shown in the console instead of a toaster
2. Error is not in a human readable way

Expected behavior: user sees a toaster with a human friendly error instead of a hex string.

Pls use https://github.com/superical/ethers-decode-error/tree/1.x, I'm sure we can simply fork it and set node version to <20.

[ariesgun]

Expected behavior: user sees a toaster with a human friendly error instead of a hex string.

It should be made clear in the spec.

[ariesgun]

[github-actions]

Unused dependencies (1)

Filename	dependencies
package.json	@ariesgun/ethers-decode-error

Unlisted dependencies (2)

Filename	unlisted
static/scripts/rewards/web3/erc20-permit.ts	ethers-decode-error
static/scripts/rewards/web3/erc721-permit.ts	ethers-decode-error

[0x4007]

I think we should self host @rndquu

[ariesgun]

Is it something I can do?

[0x4007]

Perhaps we need to get a copy of your repo and then host it. @rndquu @whilefoo @gentlementlegen

[Keyrxng]

Yeah we do.

Also Ana shared QA on another issue that showed the toaster displaying the same error that aries provided as QA, so there is little difference compared to what prod is running rn as far as I can tell.

We should use 4byte and pull the onchain fn names into the console using the selector from the data in the call that fails. Saves having to manually do it while debugging.

[ariesgun]

For now, waiting for the ubiquite-hosted npm package for ethers-decode-error to be created.

[gentlementlegen]

I opened a PR on that regard here: ubiquity/ethers-decode-error#1
@0x4007 after thought, should I create it there or in ubiquity-os

[0x4007]

I suppose all crypto related software should be under ubiquity org.

[ubiquity-os]

@ariesgun, this task has been idle for a while. Please provide an update.

[ariesgun]

Still ongoing

[ubiquity-os]

@ariesgun, this task has been idle for a while. Please provide an update.

[gentlementlegen]

I had implemented our own ethers-decode-error but every time the publish fails:
https://github.com/ubiquity/ethers-decode-error/actions/runs/11735404423/job/32693050614#step:7:21

@ariesgun Did you manually publish it? If so I do not have the NPM token myself and @0x4007 should publish it manually. If not let me know how you proceeded.

[ariesgun]

I had implemented our own ethers-decode-error but every time the publish fails: https://github.com/ubiquity/ethers-decode-error/actions/runs/11735404423/job/32693050614#step:7:21

@ariesgun Did you manually publish it? If so I do not have the NPM token myself and @0x4007 should publish it manually. If not let me know how you proceeded.

yes, I did it manually

[ubiquity-os]

@ariesgun, this task has been idle for a while. Please provide an update.

[ariesgun]

@ariesgun, this task has been idle for a while. Please provide an update.

Still waiting for the package to be uploaded

[gentlementlegen]

@0x4007 You are the one with NPM credentials if you can please publish the package.

[rndquu]

@0x4007 You are the one with NPM credentials if you can please publish the package.

Isn't NPM_TOKEN relevant? The latest release workflow silently failed with The local branch 1.x is behind the remote one, therefore a new version won't be published. . Should we publish manually first or there's smth wrong with the package.json?

[gentlementlegen]

@rndquu yeah I tried several times:
#353 (comment)

Either we should change the workflow, either manually publish because it seems to compare the version to the main forked repo so we cannot make any modification and publish it. We could copy the publishing action we usually use for our other projects.

[ubiquity-os]

@ariesgun, this task has been idle for a while. Please provide an update.

[ariesgun]

@rndquu yeah I tried several times: #353 (comment)

Either we should change the workflow, either manually publish because it seems to compare the version to the main forked repo so we cannot make any modification and publish it. We could copy the publishing action we usually use for our other projects.

Is there anything I can do to finalize this PR? As far as I understand, the package has not been uploaded yet due to token issue?