Some validation combination will keep figwheel from reloading

[sveri]

Hi,

Today I had a funny evening tracking down some really strange behavior.
I setup a minimal example to reproduce that. Please checkout https://github.com/sveri/prismaticschemabug

The code in question is in https://github.com/sveri/prismaticschemabug/blob/master/src/schemabug/core.cljs

To reproduce start lein figwheelfrom a command line from the project. After compilation open the provided index.html from the resources folder. It will load the compiled javascript and print "a".
Now, change the (println "a") to (println "b").
The expected behavior is that figwheel reloads the changes, but nothing happens.

To fix it. comment line 22 and uncomment line 17.
Reload the page. It will print "b" now as expected.
Now change the println expression to print something else and figwheel will reload the changes as expected.

There is no error or stacktrace available. I have no idea what is going on here. The reason I provided the whole validation expression is, if I remove some of the other (s/one...) expression the reloading will work too.

Hopefully some of you can figure out what is wrong there, might also be figwheels fault, I have no clue.

Thanks for looking at it,
Sven

[w01fe]

That's ... really weird. That String is just embedded in the errors as-is, so I'm not sure how Schema could be causing this -- might even be something weird with in JS-land. That said, we'll look into it when we get a chance. Thanks for the report!

[sveri]

I wish you good luck with it. This looks like the kind of stuff that somehow solves itself a year and some version updates later to just reintroduce itself again 2 years and more upgrades later if it is not urgent enough to work on.

[w01fe]

I looked into it briefly, and couldn't figure out any root causes.

One thing I did notice is that if I comment out everything after the "print" and load the browser, then I can uncomment the stuff below and change "a" to "b" and back and it all works as expected. This leads me to believe it might be something to do with figwheel, although it could definitely be something weird with schema that triggers it.

Given I don't know much about figwheel, and don't see any error messages, I'm not sure how to make progress on this from here.

[sveri]

It seems like I cannot easily reproduce it just for clojure. I tried to contact the figwheel maintainer via slack. I'd say we wait if he has an opinion for this, if that is ok for you.

[w01fe]

Sounds good, thanks!

On Sat, Dec 12, 2015 at 6:37 PM, sveri [email protected] wrote:

It seems like I cannot easily reproduce it just for clojure. I tried to
contact the figwheel maintainer via slack. I'd say we wait if he has an
opinion for this, if that is ok for you.

—
Reply to this email directly or view it on GitHub
#315 (comment).

[w01fe]

Did you ever hear back from the figwheel maintainer?

[sveri]

Nope, unluckily not.