This is closer to the original design:
* 6px between inputs
* 6px vertical padding on condensed rows
* 12px vertical padding on relaxed rows
Note that 'relaxed' rows are now smaller than the regular rows, which was not the original intention. But we haven't got spare time to do a broader review of table styles right now.
Just don't. It makes life hard and doesn't solve the problem properly.
Now, when the content doesn't fit within the screen width, it will flow naturally and not jump up and down all over other elements.
Well, that made the JS way simpler.
Adds a lot of classes though. Maybe we could do it based on column index instead, but this will do for now.
table.hide-col0 { td:nth-child(0) { display: none; } }
The flash container was set to 100% width to center the messages on the
screen. The messages were covering only part of the screen though. So
the container beyond the actual message box was covering part of the
page, blocking clicks on elements.
A new way of centering the container with CSS translate means that the
width of the container can be the same as the content, not covering
anyting accidentally.
And moving the whole container up instead of only moving the contained
message allows us to interact with elements below the flash message as
well.
This also improves the styling of the orders action dropdowns (on index and edit pages). It adds the new chevron icon, but needed some fiddling to make it look right.
Use the new design for checkboxes and fix alignment.
Removes redesigned-input, which is a small regression on the old design, but I think it's acceptable bcause we're going to shut it down soon.
I developed this while going down a slightly different path. I tested it visually, but the case I tested doesn't exist. I'm confident it will work if we ever have an error on another select though.
That was surprisingly easy. Note that it's still shared with SR.
It hides a bit early though: when the web response returns, but before the DOM has been rendered. Something to optimise in the future.
There's still an odd 1px height change on hover that I can't track down. I think it would be better to just give new variants a default of 1 (blank is not valid anyway).
Thankfully I was able to use basic DOM features, so there's no coupling of the logic with tom-select.
It wasn't going to be simple to get tom-select to listen for the 'changed' class on the original select, so I found a simple solution with a CSS sibling selector instead.
