Clojurians Log v2

@tony.kay we are starting a new project from the template here, when we used the bin/rename-project.sh it created duplicates of most of project files, adding "" to the end of the file name, is that expected?

Yeah, the rename is broken.

sorry about that. Would love a patch...just have not had time.

@wilkerlucio ^^^

Taking a look, it seems like a patch about compatibility with older versions of sed is causing this, ill try to have a patch here soon

figwheel isn't reloading the devguide; what's the most common cause of this (playing lotto here)?

watching devguide in figwheel; ran (clean-builds); no dice yet.

shows 'figwheel: socket connection established'.

ran the repl with 'JVM_OPTS="-Ddevguide" lein run -m clojure.main script/figwheel.clj'

@wilkerlucio i just pushed a patch that should fix it, tested it locally

@macrobartfast i cannot reproduce, i just ran your command locally

@adambros thanks, and stand by... may have a solution in the works here.

did you go to localhost:3449 ?

@adambros thank you 🙂

@adambros I did... it looks like I'm editing a copy of the file that's not being watched... pebkap I think

@adambros stand by for 5 and I think I'll have it solved

@adambros, tested with @wilkerlucio and it worked, but it's not on the default branch "develop", could you merge there?

😳

had been editing ./resources/public/js/devguide/untangled_devguide/A_Introduction.cljs instead of ./src/devguide/untangled_devguide/A_Introduction.cljs

@rfhayashi woops! let me fix that, i will rebase as that was a total mistake, so make sure you force pull once im done

@rfhayashi @wilkerlucio ok rebased so develop & master should be correct now you will have to force pull if you’ve been following along with me

So I'm finding that where (om/transact! this '[(widget/new ~(om/tempid))]) works as expected with regards to tempids, (om/transact! this '[(widget/new ~(om/tempid)) (widget/new ~(om/tempid))]) only resolves the second tempid.

Is that expected, and if not, is that more likely on the om.next or untangled side?

@therabidbanana Are you using a backtick for the macro quote? `

doesnt seem like the problem if '[(widget/new …)] works

@therabidbanana i would think the problem is with untangled-client, probably the networking related code

@therabidbanana So, the problem is responses from the server come back in a map

so sending both of those in a single transaction will end up with both going, but only one can response because there is only one slot in the response map

for 'widget/new

{ 'widget/new { :tempids { old new } } }

and our server handling code isn't smart enough for that case

wouldn’t you expect it to just merge the tempids?

so, this is a bug on untangled server's merging of tempids into a response

yes, I would

feel free to open an issue.

Sorry - was away - yes I'm actually quoting with ` but that was harder to show with slack syntax

see https://github.com/untangled-web/untangled-server/blob/master/src/untangled/server/impl/components/handler.clj#L76

Okay - thought that might be the case, somewhere in the server, but wasn't sure where to dig in more

that function needs logic to look for :tempids (mutations can return anything else as well) and do merges on the tempids ONLY

Is the expected case to also merge all keys, or just the tempids?

Heh, one step ahead of me

we can't assume we know what the others are

Stepping back - is this functionality that should work and doesn't, or is it generally unexpected to chain together mutations like this?

so, I'd say that we have a sort-of unsolvable case here. If you invoke a remote mutation more than once in a transaction that wants to return values you'll have to override this logic

*other than tempids

we might provide a hook for people to customize it

right, other than tempids, and many people will never need anything other than tempids

and most won't need to invoke the same mutation more than once per xaction

so, it should be a rare case

so, I'd say fix the tempid logic, and wait for someone to have further need

Yeah - I'm only doing this because it seemed most straightforward - I could easily just write a new multiple mutation

I felt like based on the way transactions work it felt like it should work?

we've got a ton to do here, so no horsepower will be given to this...will accept a PR though

Your case should work the way you wrote it...we just didn't write the plumbing right because no one thought of the case

except David Nolen 😉 I wondered why he wasn't auto-raising return values

well, he could have auto-raised, but you'd still have to decide on the merging of responses