Fork me on GitHub
#om
<
2016-07-29
>
ag00:07:52

hey guys, any React experts here? Can someone help me here with css-transition? I have a table, whenever user clicks on a col-header I am changing class of sort icon “fa-sort-desc”/“fa-sort-asc". Now I would like to animate that thing. Tried using vanilla css, by attaching keyframe animation to specified classes. But it only works very first time, when clicked on a header. I tried using CSSTransitionGroup - still no luck, there’s no DOM update happening, - class gets removed and a different one added

ag00:07:05

Anyone has any ideas how get it to work?

ag00:07:17

Maybe I can do something right there in onClick handler?

dankweaver02:07:31

CSSTransitionGroup should only be needed if the element is entering or leaving dom.. Sounds like your element is staying in the DOM just changing the className

peeja16:07:52

I asked a while back if an ident's second value could be a vector (essentially creating a compound key). The answer was no. But the only things that seem to not support it are the prospective specs (https://gist.github.com/swannodette/842326cbb6ab35e0d1aa51fb41c5697e) and Datomic.

peeja16:07:32

I understand why Datomic lookup refs can't support compound keys, but it would be really helpful to allow them for other users.

peeja16:07:55

unless there's some upcoming work in Om that they'll actually be incompatible with

peeja16:07:31

Case in point: on CircleCI, we have orgs which are identified by an org name and a VCS provider (GitHub or Bitbucket). It would be nice to continue to use that pair as the key, rather than introduce a new id key just to satisfy a spec that's based on conforming to Datomic's technical limitations.

peeja16:07:58

(unless I'm missing a theoretical issue, in which case I'd love to know what that is)

peeja16:07:17

I suppose it comes down to our API, which is parameterized by a VCS provider and an org name, not by an opaque id. We prefer that design, and so it would be convenient (possibly necessary?) for our idents to match.

anmonteiro16:07:59

@peeja: even though idents like that are not officially supported, I think your app will work if you use something like that

anmonteiro16:07:32

from what I remember, I don’t think it’ll break anything?

anmonteiro16:07:35

I might be wrong

peeja16:07:36

Well, I'd rather not do something so fundamental that's not supported.

peeja16:07:44

but in a toy app, it does appear to work

anmonteiro16:07:29

I suggest you open an issue with your use case so that it just doesn’t get forgotten

peeja16:07:37

What I'm asking is: would the team consider calling it supported (but inadvisable if you're using Datomic on the backend)?

anmonteiro16:07:51

might even be a good suggestion to support it, but I’m not the person to say

peeja16:07:38

Er, where's the Om JIRA again? Is it linked from somewhere?

peeja16:07:13

Oh, I guess I'm thinking of filing ClojureScript issues.

anmonteiro16:07:14

not JIRA, just github repo

chris-andrews17:07:34

@peeja I also found it useful to use maps as the second element of an ident a while back. I didn’t know about this spec. I think it would be worth filing an issue to discuss it at least

chris-andrews18:07:43

And using maps worked without any bugs for me

peeja18:07:59

@chris-andrews: Yeah, someone else on our team also suggested maps, and that seems like an even better idea. I've filed an issue: https://github.com/omcljs/om/issues/737

chris-andrews18:07:39

@peeja Awesome, thanks for starting that discussion. I chimed in on your issue.

colindresj20:07:53

For those using om next, or really writing any code that involves client-side querying, how are you addressing the security issues involved?

dnolen22:07:49

@anmonteiro: are there any outstanding PRs you want to get in?

dnolen22:07:53

I just merged the last 3 I saw from you

anmonteiro22:07:19

@dnolen: thanks, those were the ones

anmonteiro22:07:04

@dnolen: if you’re looking at Om, it might be a good time for me to say that there are a bunch of issues that I think are no longer valid

anmonteiro22:07:14

or already solved

dnolen22:07:34

@anmonteiro: cool, do you have some in mind for me to look at?

dnolen22:07:06

thanks! closed all of those

dnolen22:07:15

1.0.0-alpha41 is on Clojars

dnolen22:07:32

oops thanks

marcin.cylke22:07:21

Hi, I'm going through Om basic tutorial (https://github.com/omcljs/om/wiki/Basic-Tutorial), now at "Higher Order Components". I always see only single entry from the dataset, the top one. Is this some known issue? I can't find any difference in what I've written and what's in the tutorial, so could you point me to the right place? I've pasted this code into a Gist: https://gist.github.com/zygm0nt/4f27cb733552f79ccf31257185b95143