[wasm/bug] Handle error when trying to open already opened database

[SirSevenG]

Case: in mobile/desktop/web application user can have multiple wallets with the exact same seed be it intentionally or by mistake.

While libmm2 handles it well. wasm module has issues with it. Trying to switch between different wallets with same seed phrase in web app leads to error Internal error: Database 'MAIN::hash::swap' is open already.

To reproduce you may use build from PR 154 (air_dex) and later dev, when merged. I will notify in comments.

Steps to reproduce:

• create two or more wallets with the same seed (import wallet from the main screen)
• login in one of 'em
• from top right corner of the screen press wallet name -> select from drop down menu other wallet
• enter password, confirm log in
• see error in browser console

[artemii235]

@sergeyboyko0791 Could you please take a look and fix this?

[sergeyboyko0791]

I think it's related to this #1020 issue.
As for now, we can't guarantee that a MarketMaker instance is stopped immediately on stop RPC call. The last time I tested it, it took ~30s to stop every spawned future. So while the old instance is run, databases are open too.

The problem can be also related to hanging pointers: #985
In the PR mentioned above, I fixed the error, but it could appear again since we continue extending and developing atomicDEX-API.

@SirSevenG could you please check if you can log in after 30s-60s timeout?

[SirSevenG]

Tried the following:

2 wallets with different seeds. A and B.

Case 1:

• Log in into A.
• Switch to B - no issues.
• Switch back to A (less than 30s passed)
• database is still open.

Case 2:

• Log in into A.
• Switch to B - no issues.
• Switch back to A (more than 120s passed)
• database is still open.

So 2+ minutes is not enough to prevent database is open already error.

So you have to log out completely (and/or reload web page) to clear databses connections.

[artemii235]

cc @sergeyboyko0791 ☝️

[sergeyboyko0791]

I'll investigate this issue today, thanks for the testing @SirSevenG

[SirSevenG]

Small update: #1162 fixes only one of the issues listed above. Database error still persists in general.

[artemii235]

@SirSevenG @sergeyboyko0791 What is the current status of this issue?

[sergeyboyko0791]

Should be fixed by #1490
@yurii-khi Have you had a chance to test how long mm2 stops after the PR mentioned above?

[onur-ozkan]

Will close this due to #1158 (comment)

Feel free to re-open if the problem still exists.