sapphire-wagmi-v2: Add EIP-6963 support #550
Conversation
✅ Deploy Preview for oasisprotocol-sapphire-paratime canceled.
|
lubej
force-pushed
the
ml/wagmi-eip-6963
branch
5 times, most recently
from
183085f to
401322a
Compare
lubej
force-pushed
the
ml/wagmi-eip-6963
branch
3 times, most recently
from
720f252 to
2cd6377
Compare
|
Playwright test seems to be failing, but I could only reproduce it once. Needs investigation, seems there is an alert that need to be dismissed upon deploying the contract. |
Playwright seems flaky. Also failed on @ZigaMr PR https://github.com/oasisprotocol/sapphire-paratime/actions/runs/14178490717/job/39718975456?pr=552 |
The issue here is that |
| const _addEventListener = EventTarget.prototype.addEventListener; | ||
| Object.defineProperty(EventTarget.prototype, "addEventListener", { | ||
| value: function ( | ||
| type: string, | ||
| callback: EventListenerOrEventListenerObject | null, | ||
| options?: AddEventListenerOptions | boolean, | ||
| ): void { | ||
| if (type === EIP6963_ANNOUNCE_PROVIDER_EVENT_NAME) { | ||
| _addEventListener.call( | ||
| this, | ||
| type, | ||
| (event: Event) => { | ||
| let patchCustomEvent = null; |
There was a problem hiding this comment.
This looks like it breaks removeEventListener(EIP6963_ANNOUNCE_PROVIDER_EVENT_NAME, callback, options) because callback function won't match inline patch function
There was a problem hiding this comment.
If removing EIP6963_ANNOUNCE_PROVIDER_EVENT_NAME listeners isn't common then I would just add a headsup comment
There was a problem hiding this comment.
I haven't seen a codebase, that would try to remove the event. But yeah, I see the potential issue. Thats why I didn't add "restore" to original _addEventListener, which could cause memory leaks, because I did not expect that someone would try to remove the event listener.
Either way, I added the comment here: #565. I hope that is enough for now, and if there is an use-case for it. I am happy to implement more robust solution, that would "reverse" event patching.
Closes #499