Fork me on GitHub
#fulcro
<
2019-09-19
>
tony.kay01:09:47

contribute a port @grant ??? The old source won’t take much to adapt…just have not had time, and don’t use myself.

tony.kay01:09:10

glad to point out the parts to do…There should be adequate hooks for it to all work well

grant01:09:28

Cool, thanks Tony. I'll reach out if I can find some time to work on it.

tony.kay01:09:03

3/4ths of the time is prob documentation updates and testing

mitchelkuijpers15:09:36

We are porting to Fulcro 3 and we used to have our own version of the dynamic routers (because we ran into timing issues from time to time). But we also seem to have issues with Fulcro3 dynamic routing... Fulcro inspect makes me think there is some kind of timing issue:

tony.kay23:09:20

Could you look at the state machine in app state at each of those interactions? The ::uism/asm-id table holds router state machines by their fqname (as shown in tx log). Be interested in what current-state is.

tony.kay23:09:54

The timer for failure starts on deferred, and is supposed to be cancelled on target ready.

tony.kay23:09:40

state machine definition is:

tony.kay23:09:35

I wonder if it is in either the :initial or :routed state

tony.kay01:09:22

I stuck a try/catch in uism near the timer cancels..perhaps a handler threw an exception. Try Beta 18 SNAPSHOT

tony.kay01:09:42

Also enable debug logging in your console…there are a lot of UISM messages there

tony.kay02:09:06

bumped 18 SNAPSHOT…had a typo

mitchelkuijpers06:09:18

On thing that we noticed it that it only seems to happen when we a render body to the defrouter, so maybe that triggers some extra case

mitchelkuijpers07:09:07

This is with RC1 btw

currentoor16:09:49

@mitchelkuijpers what’s the situation? i.e. what do you expect to happen that’s not happing?

mitchelkuijpers16:09:52

It goes to loading an then to failed, while the load succeeds

mitchelkuijpers16:09:00

It happens randomly

mitchelkuijpers16:09:21

It should stop at :ready

tony.kay17:09:13

Hm. I thought we had gotten all of the timing issues fixed. The easy workaround is to use route-immediate, and use load markers in state instead. If you feel like trying to make a repro case I’d be glad to look at it.

mitchelkuijpers06:09:18

I'll use route-immediate for now, I need to get this project out. And then make a reproduction.

mitchelkuijpers06:09:37

This is at a user event

tony.kay17:09:29

are you initializing the routing system on start?

tony.kay17:09:03

I guess I should have asked: “is this at init startup?”

tony.kay17:09:10

or at user-driven event

tony.kay17:09:46

The state machine might need another “edge” for ready events…perhaps it was in the wrong state when you called route ready.

tony.kay17:09:45

Did you get a console warning/error about the event being “unexpected”?

mitchelkuijpers06:09:49

No I don't get any warning or error