Fork me on GitHub
#fulcro
<
2019-02-09
>
tony.kay02:02:16

@nha In fact there is. thheller has something in his set of shadow libraries that is much like that.

tony.kay02:02:06

it’s all react underneath, so custom DOM elements are just a function/macro away

thheller09:02:17

to be honest having used this a while I would no longer recommend it

thheller09:02:34

or at least maintaining a project after 1+ years has the same issues I had with pure SCSS

thheller09:02:01

still prefer it over writing actual SCSS but it is not the-one-true-solution I once thought it was

nha13:02:54

Oh. Interesting. I haven't used something like that in clj/cljs, but I have used react styled components, and was fairly happy with it. Although now that you say that it is true that generated names can be annoying.

nha13:02:54

Out of curiosity, do you mind sharing what you use now?

nha13:02:17

(one problem I had with these was generating css from react component styled and share them afterwards)

nha13:02:06

Reading the wiki, yours doesn't seem to suffer from the generated names issue though

thheller13:02:38

oh I'm still using it

thheller13:02:01

I guess my one main issue with this is: naming is hard!

thheller13:02:21

every damn div/span/whatever has to be given a unique name

thheller13:02:51

so if you want something that is 90% the same as another element but maybe just a different color or so

thheller13:02:06

you either end up with inline styles (eek) or gotta come up with a new name

thheller13:02:21

that gets rather tedious after a while and requires a lot of discipline

thheller13:02:16

I'm considering trying something like tailwind css in a real project

nha11:02:57

Looks interesting yes

grounded_sage11:02:36

I’m currently using it. Works pretty well as a lot of css bloat comes from common styling that utilities can handle. But it doesn’t solve responsive design that well. Media queries are best when content breaks not arbitrary screen sizes. There is no solution for CSS Grid. Component styles aren’t the best as you still have to name them ones

nha12:02:24

started playing with it, agree with the first part and quite like it. It doesn’t play well with the dom shorthand in fulcro, so I am using {:className "a b c"} now.

thheller13:02:56

don't like the classname mayhem very much but something like https://dev.to/chrisfrank/introducing-nanostyled-2p6k may help with that

thheller13:02:04

although it brings back the entire naming issue

😁 5
eoliphant18:02:03

Hi, thought I had an issue with the new dynamic routing, but I think I got it fixed, but I think a note in the doc, might be helpful. (or this might be a bug?) Based on my reading, I assumed default route being the first in the target vector, would result in that component being rendered automatically. I was getting an error about the sm for the router not being found, etc. It was only after I added an initial change-route to the started-callback, per the full example that everything started behaving

tony.kay20:02:29

@eoliphant Yeah, could probably use a comment in the documentation. I don’t remember off the top of my head…It seems like it should probably default if possible, but it might be a limitation required by the generic composition (there is no easy way for a router to locally “know” it is root)

tony.kay20:02:49

Ah, yes, the change route is what starts the internal state machines for the routers

5
tony.kay20:02:57

so at the moment it is a required first action

tony.kay20:02:27

possible that that could be moved to the lifecycle of the router component itself, which would make a “default route” work out of the box…at the moment the workaround is easy (and likely to be what you do anyway), so I consider it low priority

tony.kay20:02:31

but docs should be changed

tony.kay20:02:14

@thosmos @eoliphant how is incubator’s dynamic router working for you in general? Seem solid?

eoliphant21:02:03

Solid in 70's voice I know you've been "separating concerns" but I think an out of the box url integration would be cool. I'm playing around with pulling out the segments automatically into reitit etc

tony.kay04:02:15

Well, the segments are trivial to get with split, and there are numerous ways to integrate with history events…so I’d rather not do that part.

5
Adrian Smith22:02:47

Is fulcro on cljdoc?