Fix issue

[anxolin]

Summary

<<if there's an issue>>Fixes #issueNumber

High-level description of what your changes are accomplishing

Add screenshots if applicable. Images are nice :)

To Test

1. <> Open the page about

• <<What to expect?>> Verify it contains about information...
• Checkbox Style list of things a QA person could verify, i.e.
• Should display Text Input our storybook
• Input should not accept Numbers

2. <> ...

Background

Optional: Give background information for changes you've made, that might be difficult to explain via comments

[github-actions]

CLA Assistant Lite All Contributors have signed the CLA.

[github-actions]

• 🔭 GP Swap: CoW Protocol v2 Swap UI

[alfetopito]

I'll document the explanation of what's happening in this issue.

Without code changes just adding a few debugging entries, we get to this:

The first part is logging stuff from https://github.com/gnosis/cowswap/blob/d42aabe6f0/src/custom/state/claim/hooks/index.ts#L970-L975

This is where the total max investment amount in investment currency is calculated.
This is where the weird amount is coming from

The second part is logging stuff from

cowswap/src/custom/state/claim/hooks/utils.ts

Line 181 in d42aabe

return { vCowAmount: price.quote(amount), investmentCost: amount }

This is where the source for the missing atom is coming, causing amounts to be 99.99% claimed, but not quite 100%

[alfetopito]

Why is the amount not precise in the first place? (the weird 25000000000000050000000000000)

Without diving deep into how JSBI/BigNumber and etc handle the amounts, my assumption would be something funny with its math.

From the contract, we get that 1 vCOW = 150000 atoms of USDC == 0.15 USDC
Inverting the price we get 1 USDC = 1000000000000000000/150000 == 6666666666666 (18 decimals from vCOW - 6 decimals from USDC) == 6.666666667 vCOWs

Which is not a precise price, but that's how we represent it in the UI and also how we work with internally

Also, let's take a look at the amounts.

The raw claim data for this case we get is 166666666666667000000000 in vCOW atoms == 166,666.666666667 vCOW

Applying the price to it, we get the weird representation of 25000000000000050000000000000/1000000000000000000 == 25000000000 atoms of USDC == 25000 USDC

[alfetopito]

Here's regarding the last missing amount on the way back.

As can be seen for the second claim, the original claim amount is 166666666666667000000000, while the amount we get by applying the price (0.15 USDC per vCOW) we get 166666666666666666666666

[W3stside]

but what is the difference between inverting above and now here?

[W3stside]

is the math found in the method used on line975 fixing this? if so, damn, that's sketchy af for us to be using this math lib...

[W3stside]

careful here about the package (just note)

[anxolin]

@alfetopito let me know if we close this one or u want it open

[alfetopito]

I was supposed to do a post-morten after finishing with the uni-merge.
Keeping it open will be better to remind something must still be done :)