Fork me on GitHub
#om
<
2016-03-29
>
kovasb00:03:22

@anmonteiro: full-query doesn't contain the ident it seems, but i was able to construct what i needed with get-query

kovasb00:03:40

still it seems like somehting that transform-reads should do for me..

tomjack04:03:33

I wonder if it might make sense to support componentless ident metadata in queries -- {:foo ^{:ident (fn [props] ...)} [:bar]}? -- for normalization when referenced data does not need its own component?

threepointone06:03:28

You don't need to actually render the component that you include in the query (iirc). So you can make a component that doesn't have a render, and include its query in a parent.

cjmurphy06:03:15

I suspect tomjack doesn't like the idea of components that don't render, that's all. It does seem a bit odd to have such things. But personally I'd rather have them than have to remember some weird special case metadata syntax.

tomjack06:03:33

yeah. to me the weird thing is using defui/ui for a piece of data and a function, not metadata simple_smile

tomjack06:03:52

but also I wonder whether the question indicates confusion about normalization etc

tomjack06:03:00

also, it need not be a special case, maybe? I imagined that om/get-query might put the :ident metadata on the query

tomjack06:03:22

not that that would help anyone remember :ident

vmarcinko13:03:28

I'll ask again because none replied last time - is there any at least newbish example of server-side query parser for some SQL db? I kidna have hard time realizing this is not hot question because as nice as datomic is, if Om Next is demonstrated with just Datomc examples around, it seems to me it won't find larger adoptance?

hlolli14:03:05

In the file om/dom.clj for svg there is missing filter and feGaussianBlur, if they were added, wouldn't they work with svg's? (and probably more tagged svg functions)

anmonteiro14:03:22

that said, you can use other tags with ReactDOM.dangerouslySetInnerHTML

hlolli14:03:48

allright, I see that they offer react.art I guess I will take a look at that as well.

cjmurphy14:03:34

@vmarcinko: I don't think there is such a thing. Sounds like a useful library for someone that needs it to write.

tomjack18:03:27

anyone happen to have already gotten chrome's password saving to work for an om.next login form?

isak18:03:53

@vmarcinko: i will write one soon for .NET that i could probably show, if you think that will help

vmarcinko19:03:37

@isak: clojure lib on .net ?

isak19:03:50

@vmarcinko: probably just an FSharp lib on .NET. I can't change the backend to JVM at the moment.

vmarcinko19:03:51

@isak: And it will parse Om.Next query, at least some basic stuff?

vmarcinko19:03:37

I asked this very question because being able to specify the query on client, and somehow execute it simply on backend is IMO biggest selling point of Om.Next and other JS libs that Om has been inspired by (Relay/Falcor), but it seems to me that backend story is really up in the air mostly, if one doesn't look at datomic, which is usually the case with systems out there

vmarcinko19:03:32

And it seems to me from quick glance, that this is also the reason that Relay hasn't taken off in big way in JS land also

vmarcinko19:03:50

Of course, I don't doubt that some query backend parser coudl be implemented with a lot of effort, but if this is too big, then one starts to wonder whether it was worth it

isak19:03:33

agreed, and yea simple_smile