Fork me on GitHub
#nrepl
<
2019-04-24
>
bozhidar11:04:42

@pez Did you manage to solve your problem? I saw the convo just now.

pez14:04:36

@bozhidar I am writing an issue on Leiningen right now. And will volunteer to prepare the PR is they think that removing the hard dependency on nrepl is a good idea.

pez14:04:41

But that is for the long run. In the shorter perspective I must get Calva to work with latest piggieback.

bozhidar14:04:36

So, what exactly is the problem? The latest piggieback didn’t really change anything (it was just updated to work with the current printing middleware) and seems to work fine for me.

pez14:04:27

There is a known (if I understoof @dominicm correctly) bug with out and err messages, that I think I can live with a while. But there is also an exception thrown. Probably because I do something stupid.

pez14:04:49

OK. I don't know for sure it stems from piggieback. But using nrepl 0.6.0, cider-nrepl 0.21.1 and piggieback 0.4.0 I get those two problems, and I don't get them fith 0.5.2, 0.20.0, and 0.3.10, respectively.

bozhidar14:04:03

The thing that @dpsutton discussed is a middleware in cider-nrepl, but I assume you’re referring to this.

bozhidar14:04:37

> OK. I don’t know for sure it stems from piggieback. But using nrepl 0.6.0, cider-nrepl 0.21.1 and piggieback 0.4.0 I get those two problems, and I don’t get them fith 0.5.2, 0.20.0, and 0.3.10, respectively.

pez14:04:44

Let me read that issue...

bozhidar14:04:56

Looking at the stacktraces you’ve posted they certainly seem like some version mismatch.

bozhidar15:04:22

E.g. it’s not possible to get an error about pr_values missing with cider-nrepl 0.21.

pez15:04:33

It doesn't look exactly like what I am experiencing.

bozhidar15:04:44

> Exception in thread “nRepl-session-8067c4a9-e320-4f8e-b84f-890c8cbf2097” java.lang.IllegalArgumentException: No implementation of method: :send of protocol: #’nrepl.transport/Transport found for class: nil

bozhidar15:04:54

Is that the stopper problem?

pez15:04:51

I am not sure if it is. I think I should be able to keep my repl client standing, and if so it is not a stopper.

pez15:04:30

But I haven't really tried fixing it on my end yet, Just very rececently realized the problem was there.

pez15:04:43

I have a pretty reproducable problem with those version combos I mentioned above. With the older set of versions, out and err messages arrive as I expect them to, but with the newer set, those messages are gone. What is pr_values, btw?

bozhidar15:04:58

It used to be the middleware that converted EDN values into strings that can go in the :value slot.

bozhidar15:04:38

It was replaced by a similar, but more flexible print middleware.

bozhidar15:04:21

> With the older set of versions, out and err messages arrive as I expect them to, but with the newer set, those messages are gone

bozhidar15:04:31

They behave now the same way they always did.

bozhidar15:04:29

I think you’ll find it very useful.

bozhidar15:04:40

Your error indicates that somehow the print transport is nil for you, but I can’t imagine how this can happen. Maybe you’re passing some weird params in the eval request?

pez15:04:27

Maybe I am. Will investigate deeper tonight, and let you know if I still think the problem is with the new versions. But thing is that even with the new versions the clj repl behaves as I expect, it is only the cljs repl that gives me grief.

bozhidar15:04:13

That’s pretty weird. This has been out for 3 months now and no one has reported similar issues.

pez15:04:24

I'll try to produce a repro. What will probably happen is that I find out what i am doing wrong.