fix(select): Always wrap select with .input-field

[MaximBalaganskiy]

This will make labelling and validation consistent.
Validator is also modified to reflect the changes to select.

[MaximBalaganskiy]

Please consider #416 as an alternative

[Thanood]

Sorry it took so long this time.
I think #416 is easier to read while this one here seems more straightforward because it always adds a wrapper. So you can expect there is a wrapper and propertly remove it.

I'm leaning to this one.
Do you have any preference?

[MaximBalaganskiy]

I like this one more too. Which got me thinking - shouldn't it be an element, not a attribute?

…

On 9 May 2017 3:43 am, "Daniel Bendel" ***@***.***> wrote: Sorry it took so long this time. I think #416 <#416> is easier to read while this one here seems more straightforward because it always adds a wrapper. So you can expect there is a wrapper and propertly remove it. I'm leaning to this one. Di you have any preference? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#417 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADimvCCk9kSJo7orodqlxNo23YXO7Iu3ks5r31QtgaJpZM4NO3iw> .

[Thanood]

Sure, we should think about that. I'm not sure how to do it, though..
Would be easier if we'd just create an own select for that. 😏

I'm not sure about it because of timing. It would break a few things, I guess.
On the other hand we can provide both and set some "obsolete warning" on the attribute.

[MaximBalaganskiy]

Seems like the element would be almost the same as an attribute. Even simpler if anything, because you would move html to a template. "Obsolete" would work just fine I recon

[MaximBalaganskiy]

ah, one issue though... the following would not be a valid html, would it?

<md-select>
  <option>option</option>
</md-select>

[Thanood]

No, it wouldn't be valid html.
Maybe if you use something like this as the element's template:

<template>
  <select ...>
    <slot></slot>
  </select>
</template>

That would lead to valid HTML. I don't know how this will restrict things if we wanted to add other elements (like custom validation). Then this would have to be a named slot. But until then, I guess it should work. 😃

What we'll have to keep in mind is that all of this will eventually be removed and overwritten by Materialize when it initializes the select proxy.

[MaximBalaganskiy]

The template wouldn't be an issue. The usage of this element will, at least VS editor complains about it. On 9 May 2017 9:12 pm, "Daniel Bendel" <notifications@github.com> wrote: No, it wouldn't be valid html. Maybe if you use something like this as the element's template: <template> <select ...> <slot></slot> </select> </template> That would lead to valid HTML. I don't know how this will restrict things if we wanted to add other elements (like custom validation). Then this would have to be a named slot. But until then, I guess it should work. 😃 What we'll have to keep in mind is that all of this will eventually be removed and overwritten by Materialize when it initializes the select proxy. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#417 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADimvCI5z0rsbbmeRSbGVJPyhcoqXaz7ks5r4Eo2gaJpZM4NO3iw> .

[Thanood]

Really, it complains about option in an unknown tag?
Hmm, what about md-option then? We'd have even omre control in that case.

[MaximBalaganskiy]

That's a good idea :)

…

On 9 May 2017 at 22:05, Daniel Bendel ***@***.***> wrote: Really, it complains about option in an unknown tag? Hmm, what about md-option then? We'd have even omre control in that case. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#417 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADimvDDWRzGn-e9NFuC1ugnH53TdTYNIks5r4FZugaJpZM4NO3iw> .