Fix rounding issues by upgrading decimal maths library

[macanudo527]

What? Why?

BigDecimal was version locked, but with some small modifications to a test and another small modification to display logic, it doesn't need to be. This allows us to use default gems on Alpine Docker images much easier. It also allows dependabot to automatically upgrade the gem when necessary.

What should we test?

• Visit http://localhost:3000/admin/products
• Edit any product's unit to be 500.1, which causes a rounding error.
• Save and verify that the display unit is 500.1 and not 500.09, which happened when we used trunc to create the value instead of round.

Release notes

BigDecimal will now be updated to the latest version more easily.

Changelog Category (reviewers may add a label for the release notes):

• User facing changes
• API changes (V0, V1, DFC or Webhook)
• Technical changes only
• Feature toggled

The title of the pull request will be included in the release notes.

[dacook]

I confess I've never needed to understand numbers at this level! I've added my thoughts, but I think this is ok as it is.

[mkllnk]

Nice one!

I'm glad we have this in a separate pull request now because there's quite a bit to discuss. Awesome work and I think that this update will fix some rounding bugs in our app as well. I just have two little requests to make the code more readable and robust. Otherwise good to go.

[rioug]

Nice, thanks @macanudo527 🙏

[macanudo527]

@mklink is there anything else needed for this one?

[dacook]

Sorry for the delay, as Maikel has been offline for the last week.
I can see that his feedback has been addressed, and this PR has already been added to the testing queue, we are just waiting for someone to verify with manual testing.

[filipefurtad0]

Hey @macanudo527,

Same on my side, apologies for the delay in testing this one.

I think it should close #7504. I've spotted this issue in two places in the app (pics taken before staging):

Products page

Variant edit page

Interestingly, the first test case seems fixed:

While the second one does not:

I've pulled this PR to my local env and tested it locally, but still obtained the same result. I'm very tempted to merge this as it addresses the first test case. Any idea why the second case is not working @macanudo527 ?

Adding the feedback needed label.