Here's the blog post I was promising someone a while back, complete with some code samples (and a bit of selling for our coworkers who haven't seen live-reloading in action): http://tech.adstage.io/2016/09/12/how-we-test-full-stack-clojure.html
Nice post @therabidbanana
Thanks! Think I covered everything - @currentoor and @kenbier helped proof-read/make suggestions
@therabidbanana: Really interesting post, I think it's worth posting to some non-untangled channels as well - lots in there that applies to pretty much every front end CLJS project.
Good job. So u guys are using boot instead of leiningen. Great! Any ideas to share the experiences?
I think that'd probably be something worth writing up, but I'm more of a clojure newcomer so I haven't really dealt with leiningen all that much for comparison
Oh! I guess I missed that one: https://github.com/currentoor/untangled-starter
Oh yeah, forgot @currentoor released that.
Doesn't look like it takes much advantage of being in boot from the looks of the build.boot file - we have a lot more in our actual project. Might be worth updating a bit.
But it's a very good place to start with boot and untangled. Thanks @currentoor
Ok, after reviewing the files an update would be nice ;)
sth with boot-reload and so on. I'm happy to talk and help out if I can.
Hey @baris I've actually got a boot-untangled template in the works. I'll post it here as soon as it's ready!
the visual regression test with advanced compilation is a brilliant idea
@currentoor: Iโve already got untangled-template going, have you looked at that and/or is there a reason you couldnโt put boot in that?
@ethangracer thanks!
@adambros sorry it's not a template like that, a simple toy application
like TODO MVC but hopefully more silly
ah
ok well i would like to see it as maybe i can put it in that project
for sure, i'll keep you posted. i've just setup the skeleton so far
@therabidbanana Nice article! Thanks for sharing.
@ethangracer it's kind of funny the devcards visual regression test idea were the direct result of spite and anger with a former teammate rather than some good idea we deliberately cooked up
I agree with @ethangracer, the visual regression stuff is nice
nice to know something good comes from spite ๐
i basically got furious one night and had to prove the clojure stack was worth us using ๐
i don't think i would have been as motivated otherwise lol
yeah, new obscure frameworks are a tough sell. I was exploring the competition today...still not seeing anything I like nearly as well. I know I'm biased, but the full-stack story we have is pretty darn compelling once you understand it.
it has kinks to work out still...
anyone playing with anything to componentize CSS with respect to their Om/Untangled components?
I've been brainstorming on that particular topic because we're starting an untangled-components library, and I want people to be able to style the things easily.
I've got a meeting to go to now, but if anyone has ideas, feel free to share them here or in PM
oh that would be cool!
yeah, co-located CSS? Compiler integrations the extract base CSS like the i18n stuff? Naming conventions with CSS name generation macros, etc.
I'm trying to weigh the different options and see what seems to have the best long-term promise
i used react-material-ui once in cljs
you specified the theme with a js object
and then apply it to the context feature in react
but maybe here it would make most sense to just store it in the app-state?
Other approach would be use something like garden, I've played with that recently, not too bad to use (but in CLJ, not sure if it works well in CLJS)
Now that our designer is getting more into the CLJS files maybe we can subject him to a few experiments.
@tony.kay @currentoor have you looked at https://github.com/ladderlife/om-css ?
shameless plug right there, but it works quite nicely for some cases
@anmonteiro I have not but it looks pretty cool.
it is now on my list to evaluate
I'm not sure it is the right solution...e.g. how would a designer override that?
the approach there is to generate css based on the colocated styles
I understand. I'm wondering how well that can be made to work where you have components some CSS person is trying to style
generated classes are ns-name prefixed by default, but you can specify raw classes without prefixes
in that case you can write CSS against those classes
I'll look it over...glad to see you're already working on a solution
I expect to be pleasantly surprised ๐
letโs hope so!
Iโve also recently added support for Sablono
oh, and when did you want that job??? ๐
๐