@tony.kay: can we add this support (https://github.com/omcljs/om/pull/709/files) for chained reading of idents on the Untangled read function?

looking at the plumbing I'm guessing a change here might be needed: https://github.com/untangled-web/untangled-client/blob/master/src/untangled/client/impl/om_plumbing.cljs#L30

what you think?

after more testing I'm thinking its another Om issue, trying to figure it out

@wilkerlucio: pretty sure what you just opened is not an issue

The query doesn't seem right to me

@anmonteiro: why not?

Something like [{:a [:x :foo]}] would work with my fix

But I think you're abusing unions with th query you posted

yes, but in my case I have an union at that point

i.e. The shape of the data doesn't match the union query

I dont understand why not

the way I see it, it has a multiple ident, and at the last point on the chain it has an union

I have multiple types of data, so I really need that union there

I'll have to be in front of a computer to say more :-)

but instead of using the last ident as the union keyer, it's using the first of the ident chain

it's almost working right

except the reference for the union is coming from the first ident of the chain, and I think the correct would be to use the last, makes sense?

Probably makes sense, I'll try it out in a bit and let you know

thanks

and just fyi: I was trying that with your update in place

if we confirm that as an issue, maybe it's a good idea to put in the same PR as the other

@anmonteiro: just to let you know, my actual case is even a bit more complicated than the case I posted on the issue, the minimum for it is this:

I was able to fix the one from my issue (to get the correct ident from a follow), but on this last case it's being more tricky

the difference on this case is that one of the keys on the query path is an ident

so I'm missing the loop at some point

ducky typing is awesome, ok, the minimal case can be reduced, as:

the issue now seems reduced to cases where an ident key points to an union

@wilkerlucio: having a look now

agreed that it should go in the same PR

@anmonteiro: my changes on the db->tree that solves the first issue (the one at github)

I added that v entry on the let, the second one

I'm starting to think those are 2 different issues: 1 - unions on indent chains, 2 - unions for subqueries of an ident query

@wilkerlucio: https://github.com/omcljs/om/pull/709

updated my patch to fix this stuff too

@anmonteiro: thank you so much! that fixed all my issues 😄

@wilkerlucio: there might be one more failing case lying around, there

but I really don’t have time to test it out now

if you want to cover all the bases, you could try to make an example where the join key is an ident chain

although I’m not sure whether that should be allowed or not 🙂

@anmonteiro: my real case is one of that, and it worked here

🙂

👍

@wilkerlucio: Just FYI, you can use the latest Om version with Untangled. We put Om in a provided scope so you can stay current.

@tony.kay: thanks, my current workflow to use Om master is not very good... I've copied the Om files into a folder under my src and added as a source path

Why not use checkouts?

I tried before using lein checkouts, but I couldn't figure how to make it work

the compilation always seems to fallback on some pre-installed version

oh, it should work. We do it for debugging all the time. You have to use the correct version as a dependency in your dependecies AND make checkouts AND include it in your source paths.

And you'll need to publish a local jar of the version listed on master

can you please clarify the last step? how I a publish a local jar?

We use a script to do that last bit:

humm

which requires you have lein localrepo plugin installed

like in your user profile

so, I need to go into the checkouts, make a system link, and also compile the jar, is that correct?

right, from om master, run that script

symlink from checkouts to om

you talked about source paths, do I also need to add any extra source paths too?

add to source paths

you want BOTH the top-level (clojure) source-paths (for macros)

and cljsbuild source paths

what I should add there? checkouts/om?

the source dir

checkouts/om/src/main I think

you need to point source paths at the sources 🙂

checkouts just makes it possible for lein to know to ignore the dependency jar

but it will only do that if the version in checkouts matches the version you say you want

ah, gotcha

I'll try that 🙂

k

@tony.kay: it works, so much better! thank you!

(merge-state! reconciler ui/Root {:current-tab (initial-state (ui/which-tab->tabui tab))})

I get the error no Ident on Root

how do I merge into root?

i opted for this code

(om/merge! reconciler
                        {:current-tab (initial-state (ui/which-tab->tabui tab))}
                        (om/get-query ui/Root))

seems to get the job done

i guess technically this code does less rendering 🙂

(merge-state! reconciler ui/TabUnion (initial-state (ui/which-tab->tabui tab)))
             (reset-ref! env :current-tab [tab :tab]))