contracts: fix intermittent failures in auth.ts tests

[CedarMist]

fixes: #361

Also fixes typescript warnings in auth.ts

[netlify]

✅ Deploy Preview for oasisprotocol-sapphire-paratime canceled.

Name	Link
🔨 Latest commit	a4beb2a
🔍 Latest deploy log	https://app.netlify.com/sites/oasisprotocol-sapphire-paratime/deploys/66fad74aaa73a00008a39bc2

[CedarMist]

Well... F...

This is still running into the problem! We're waiting for bearer token to expire, push a transaction through... And it still thinks the bearer token is valid?

  1) Auth
       Should call authenticated method:
     TypeError: Expected a valid transaction hash, but got 'Very secret message'
      at onSuccess (/home/runner/work/sapphire-paratime/sapphire-paratime/node_modules/.pnpm/@[email protected]_@[email protected][email protected]_djiantmfwzd4h5mvs2d3iacurm/node_modules/@nomicfoundation/hardhat-chai-matchers/src/internal/reverted/reverted.ts:32:17)
      at processTicksAndRejections (node:internal/process/task_queues:95:5)
      at async Context.<anonymous> (test/auth.ts:165:5)

[CedarMist]

@matevz Any idea why this test is still failing intermittently?

[lukaw3d]

This is still running into the problem

I tried repeating the test
https://github.com/oasisprotocol/sapphire-paratime/actions/runs/10692954396/job/29642295984
failed 6/10 repeats

I tried double await siweAuthTests.doNothing();
https://github.com/oasisprotocol/sapphire-paratime/actions/runs/10693129388/job/29642755780
failed 1/10 repeats

I tried await delay(6000);
https://github.com/oasisprotocol/sapphire-paratime/actions/runs/10694436081/job/29646226736
it succeeded 20/20 repeats

[lukaw3d]

https://github.com/oasisprotocol/sapphire-paratime/actions/runs/10692954396/job/29642295984
also randomly failed "Delegate then begin Undelegate (with receipts)"

sapphire-paratime/contracts/test/subcall.ts

Lines 318 to 320 in 58be8a6

	expect(await ethers.provider.getBalance(await contract.getAddress())).eq(
	initialContractBalance + parseEther('100'),
	);

(I tried repeating it 20 times but it just started failing differently after 10th success 🤷 https://github.com/oasisprotocol/sapphire-paratime/actions/runs/10694614566/job/29646662356 )

[matevz]

Nice catch, thanks.

[matevz]

@matevz Any idea why this test is still failing intermittently?

#375 (comment)

[CedarMist]

So, even if I wait for the tx receipt from doNothing, I still get intermittent failures.

Ohh.. I know why, because block timestamps are in seconds, we're waiting less than a second. If I change it to wait at least 1 second then rounding won't cause it to succeed when it should fail.

[CedarMist]

So to summarize:

• Wasn't waiting for tx to be mined by waiting for receipt
• Was providing milliseconds, which rounded to seconds
• On-chain time is in seconds
• Wasn't waiting an extra second, to account for possible round-up

Have now run this locally 20 times in a row with no failures.

Yet CI failed again...

Will revisit this tomorrow or Friday. For now we can continue re-running failed CI tests if we encounter this bug.

[matevz]

Interesting. Maybe the new block wasn't propagated yet to the oasis web3 gw (and postgres db) and it used the old block number when performing the query. Did you try adding sleep after waiting for the receipt and before making a query?

[CedarMist]

I'll add in more debugging to check:

• timestamp in SIWE message
• timestamp, block number & epoch returned from view call
• timestamp of block etc.