During transform, any overflow on the element is clipped/hidden. This caused all dropdown menus to be clipped and unusable. Now, once the animation is complete, the overflow is visible, and menus are usable.
Mistral Vibe AI was used to find this solution. I tried to find a CSS solution last week but failed, then started to consider using JS to remove the class, but decided against it once I realised that the product clone JS was already doing that, and it didn't seem to solve the clipping issue.
So I asked Mistral Vibe and it suggested adding 'forwards' (before it had spent energy on evaluating the existing style rules). As you can see 'forwards' was already there, but removing it helped. So Mistral was wrong, but at least pointed me in the right direction, yay!
The spec was not really testing the order of users appearing on the
page. It's also a UX detail only visible to super admins which is not
important to test. So I'm not investing time to fix it.
The locale config is set in application.rb from environment variables
already. We don't need to repeat that logic in test.rb. And because it
was outdated, the language switcher was actually broken in the test
environment. We did have an English selector for the fallback `en` even
though we were already displaying English as en_TST. And after
switchting to Spanish, we could switch back because en_TST was not in
the available locales.
I now fixed the test with the right assumption and the config to solve
the problem.
I considered a few ways to do this. Cloned products are done with MutationObserver but it doesn't quite sit right with me. A dedicated controller for newly added rows would provide a good general solution. But do we want yet another controller? I'm not sure. This works and is pretty simple (although it requires a quick loop over _every_ form element.. let's see if we can avoid that.)
Several years ago, some checkout features got rewritten and some specs
became invalid. They had been set to pending to keep the option of
rewriting them one day. Some were re-written. But I'm deleting the
remainder.
If we haven't "needed" these specs for several years then I question
their use. System specs are expensive and should only cover the most
common scenarios or the ones we know could go wrong (after a bug
report). We can always write new specs if needed. Otherwise they are
just adding to maintenance cost.
There are two types of linked variant associations: source and target, so we need to keep the name there.
But when cloning a variant and retaining a link as source, we will prefer the general term 'linked variant'. Hopefully this name works well.
Preload the allow list once in the controller. This controller was initially set up to avoid instance variables, and pass variables explicitly to the template. That's a good principle, but in practice we have a growing list of variables passed down the chain to multiple partials which is getting cumbersome. I think instance variables have their place after all.
The permission is effectively the feature toggle. Users can choose to use it, but shouldn't expect it all to work perfectly yet.
When it's considered full featured, we just need to update the translation. Hm... I hope that's not too painful.🤞
For now, we will only be able to create sourced variant from variants that are visible to us (variants that we manage)
In a later commit I will hide the option if you can't use it
