API Addition: findByClass

[danny-andrews-snap]

I think it would be really nice to have a findByClass (or equivalent) method. 99% of the time, I select portions of the DOM by class, and when using css-modules, you end up having to do awkward things
like

const subject = wrapper.find(`.${styles.slotNumberCol}`);

to select them.

const subject = wrapper.findByClass(styles.slotNumberCol);

would be much nicer.

Even in the normal case, it would nicer (albeit slightly) to do

const subject = wrapper.findByClass('slot-number-col');

rather than

const subject = wrapper.find('.slot-number-col');

[danny-andrews-snap]

Similar to #349 but different in that it matches class names exactly, not by RegExp. I would argue this is a much more common scenario and so warrants a special method in the API. So common, in fact that there exists a special native method to do just that.

[ljharb]

The "special native method" you mention is a legacy one; the modern API is to only use querySelector/querySelectorAll.

I think that "find by class" is a pattern that shouldn't be common; either way, I think adding a specialized convenience method to avoid prepending a . is complexity we probably don't want.

[danny-andrews-snap]

I think that "find by class" is a pattern that shouldn't be common;

What? Why? How else should you find elements?

I think adding a specialized convenience method to avoid prepending a . is complexity we probably don't want.

You're ignoring the first scenario, which is actually really common. wrapper.find(`.${styles.slotNumberCol}`); is pretty ugly to have to do every time you want to find an element by class name.

[serhii-yakymuk]

@danny-andrews-snap
Just use this, if you're using jest
https://facebook.github.io/jest/docs/en/webpack.html#mocking-css-modules

[craigkovatch]

@ljharb

The "special native method" you mention is a legacy one; the modern API is to only use querySelector/querySelectorAll.

Not sure that's the whole story. On my computer (macOS 10.12 Chrome 66) getElementsByTagName is literally 32x more performant than querySelectorAll. There's useful reasons to have both.

https://jsperf.com/getelementbyid-vs-queryselector/25

[ljharb]

@craigkovatch if you're selecting enough DOM elements that 3200x even matters, you've got something subpar with your architecture :-)

[ljharb]

I'm going to close this; the added complexity in enzyme isn't worth it for the simple sugar of avoiding a template literal with a leading dot.