An error was apparent in specs when trying to assign a string as the price. It's not a problem when submitting the form in the browser, I don't know why.
But in any case, it shouldn't be trying to modify a variable passed as a parameter.
Now we don't have to specify the field names on products, yay!
I tried desparately to get it working for the nested variant forms too, but sadly the form builder refuses to acknowledge the relationship. The form builder simply doesn't support a collection of objects in this way.
We could try creating a fake model similar to ProductSet that accepts_nested_attributes_for :products. But it woudl be better to create a custom form builder to do it.
Or, just manually specify field names for now!
I have to admit I don't fully understand why, but it seems to work, even though the rails guide says "only one level of arrayness is allowed (https://guides.rubyonrails.org/form_helpers.html#combining-them). Maybe it's ok here because it's not an array of arrays.
(I think this format is what the reflex spec was already testing).
The DFC doesn't actually specify which cup it means. I don't expect
anyone providing "cup" as unit to measure their produce and expect it to
calculate as a regular volume. It can just be seen as items.
This spec was nested within another block that executed a form submit
which we don't actually need here. So I flattened the structure and
repeated the few missing lines of code. This speeds up the execution.
The fee type is important in the setup because it determines the order
of the fees on the page and we access the rows by their row index.
This commit is best viewed without whitespace changes.
The problem here was that the second fee was created in a `let!` block
that came after a `before` block which visited the page. In some cases
it worked but sometimes the fee wasn't created yet when the page was
loaded. Form changes in the second row were not affecting the second fee
as intended.
We still have the scale stored in two places but in our current system
that's part of a unit's "id".
If the DFC adds that value to its standard then we can use it for lookup
and don't need to repeat it.
* https://github.com/datafoodconsortium/taxonomies/issues/7
I observed new data from the DFC Prototype. It now uses the DFC 1.8
ontology with the hasQuantity object.
It now also uses PUT requests for updates because PATCH is not as well
supported. Rails doesn't care though.
I couldn't observe a request for the CatalogItem yet because the
Prototype failed to send it.
This means you can click on the label name to select the field, and accessibility software can read the label name. I was motivated to do this so that we could update the spec to use label text rather than IDs. But it seems we can't do that with tom-select.
I've updated the styles for all form labels (in admin_style_v3 toggle) to match the design.
An inner join with the billing address was hiding some orders when
sorting by billing address name. Telling Rails that those fields are
referenced triggers an outer left join including orders without billing
address.
But when the Bulk Order Management page sorts by `bill_address_lastname`
then Ransack does most of the magic, except that we need to override the
select in combination with distinct results.
'let!' is executed in the same way as 'before', so it matters which order they are defined in.
For contexts with a single example, I just put it straight into the example.
