Fork me on GitHub
#fulcro
<
2021-01-04
>
Jakub Holý (HolyJak)16:01:44

Hello! I have remade https://blog.jakubholy.net/2020/troubleshooting-fulcro/#_frontend_fulcro into a more mindmap-like experience, where child nodes in the troubleshooting tree are collapsed by default and you must click to expand them. I believe this will make it clearer as you will see all the children of a node at once, one per line, and can click on the one that is most relevant to your problem. What do you think about it? Is it better or worse than the static version? (Simply disable JavaScript to see the original, static version.) Or perhaps I should add an "expand all" button to the top? Thank you!

👍 6
alex-eberts20:01:10

I think it’s a great way to simplify the process - nice addition! I would recommend standardizing all the headings some that they represent the problem the user is having. E.g. “Data in the client DB is missing or wrong” is great. The heading “Fulcro Inspect” doesn’t describe the problem the user is having so it’s less clear if my problem is contained under that branch.

mss18:01:58

has anyone had any luck building 1.0.15 of workspaces? getting an odd error TBD, classpath indexing is gone

tony.kay19:01:47

you using my fork or the original?

tony.kay19:01:58

that’s a shadow-cljs change I think

tony.kay19:01:06

so one way is to downgrade shadow

👍 3
thheller19:01:18

https://github.com/nubank/workspaces/pull/50 dunno if there is a new version out yet?

tony.kay20:01:24

@thheller I have not had time to patch my fork for F3, but I thought I saw a PR on the nubank one.

tony.kay20:01:51

it would be a trivial port from the nubank one if there is one

JAtkins20:01:33

@tony.kay Quirky behavior: com.fulcrologic.fulcro.algorithms.normalized-state/sort-idents-by returns a seq, not a vector of idents. If you want I can pr either a doc or fn fix.

JAtkins20:01:25

which? I can wrap in vec or add it to the docstring

❤️ 3
tony.kay20:01:30

to vec

👍 3
tony.kay21:01:06

Released in 3.4.12

fjolne22:01:12

@mss @tony.kay these versions work fine:

{thheller/shadow-cljs        {:mvn/version "2.11.4"}
 com.github.awkay/workspaces {:mvn/version "1.0.1"}}

njj22:01:12

How do I use fs/mark-complete with an ident that is a function (i.e. :ident (fn [] [:COMPONENT/by-id :foo])

njj22:01:28

logging.cljc:50  [ 55.154s] [fulcro.ui.form-state] FORM NOT NORMALIZED: [:COMPONENT/by-id :foo]

JAtkins22:01:10

Initial guess: your problem is you aren't using the FS join in your query (`fs/form-config-join`).

njj22:01:26

Yeah that’s usually the case, but this form has that

JAtkins22:01:43

If that's not the issue more info is required. Lambda idents aren't very special - they work basically the same as normal idents

njj22:01:57

and its also doing the whole post mutation in fs/add-form-config* etc..

JAtkins22:01:20

What's going on with your state? given the message I would assume the form state is present, but just not normalized (it's a reach, I know 🙂) ... in that case your root issue would be the fs/add-form-config*? Are you using the ident? (e.g. (fs/add-form-config* s Class (comp/get-ident Class {})))

JAtkins22:01:02

could be the wrong path though. given that the FS data is there, you probably already are using the ident properly

tony.kay22:01:04

@mss Try 1.0.3 snapshot of awkay/workspaces. I just merged in the fix from mainline nubank

mss22:01:33

seems to be working for me!

tony.kay22:01:41

great, thanks

tony.kay22:01:25

1.0.3 released to clojars

tony.kay22:01:12

if it works I’ll push a formal release