Fork me on GitHub
#cider
<
2018-10-26
>
bozhidar05:10:16

@pri OK, so I think it’d be really best for you to file a ticket where we track this. The chat is a poor medium for this.

alexyakushev07:10:31

@bozhidar My clojure-mode version is 20181005.452, and the Sesman issue is still there. Alright, I'll try to dig a little more to report the precise issue.

alexyakushev07:10:34

@dominicm @bozhidar Regarding boot repl, there are two things I've hit so far. 1. boot repl correctly loads data readers, cider.tasks/nrepl-server doesn't seem to. Hence, for example, the debugger doesn't work (missing dbg reader). 2. I'm used to "patching" the default repl task with custom logic, e.g. compiling Java before starting the REPL. I guess, this one is on me, and is easily fixed by patching the cider.tasks/nrepl-server task as well, but I guess this still makes it a breaking behavior enough to mention in the changelog.

bozhidar07:10:39

> Alright, I’ll try to dig a little more to report the precise issue

bozhidar07:10:44

That’d be appreciated.

bozhidar07:10:53

> 1. boot repl correctly loads data readers, cider.tasks/nrepl-server doesn’t seem to. Hence, for example, the debugger doesn’t work (missing dbg reader).

bozhidar07:10:34

I assume that’s not hard to fix, right? I’ve never used boot, this was all @richiardiandrea’s show so far and I hear he moved away from boot as well. 🙂 I guess we should at least open a ticket to track this.

richiardiandrea15:10:20

Yep not using boot anymore but can look into that, I am not officially surprised about how many things the boot repl task is doing 😃

bozhidar07:10:41

> 2. I’m used to “patching” the default repl task with custom logic, e.g. compiling Java before starting the REPL. I guess, this one is on me, and is easily fixed by patching the cider.tasks/nrepl-server task as well, but I guess this still makes it a breaking behavior enough to mention in the changelog. (edited)

bozhidar07:10:03

Yeah, we should totally mention this in the changelog and in the docs.

alexyakushev07:10:30

Re boot #1. I think the fix should be simple. I'll also take a look.

alexyakushev07:10:29

If you need help regarding the Boot stuff, feel free to ping me in PRs more liberally, can't promise many contributions but would at least alpha-test.

alexyakushev07:10:34

We use Boot exclusively.

bozhidar08:10:08

Thanks! That’s really good to know!

dominicm08:10:49

@alexyakushev what do you mean patching? Normally I just have a task like javac, and then I compose the tasks.

alexyakushev08:10:34

Either that, or boot.core/replace-task!

alexyakushev08:10:12

In any case, we relied on CIDER calling boot repl specifically.

alexyakushev08:10:40

It's no big problem to switch to another one, but the change came as a surprise.

dominicm08:10:35

I see. That's not a strategy I've really seen. I don't think relying on cider calling a particular function is part of the API contract.

bozhidar08:10:49

@alexyakushev At any rate - I’ve updated the changelog and I plan to tweak a bit the docs related to jack-in later.

bozhidar09:10:46

Btw, I haven’t forgotten about the compliment bump as well, but I was hoping to cut a new cider-nrepl release soon and the snapshot version of compliment would be an issue for this.

bozhidar09:10:29

Likely the test failures are just some Travis flakiness. I’m getting more and more frustrated with Travis, but I don’t have time to move the build to Circle.

alexyakushev10:10:42

That PR fails with JDK 10 quite reliably, but I can't explain that

alexyakushev10:10:53

Works with JDK11, but other cljs stuff fails then

alexyakushev10:10:53

It is a snapshot only because it wasn't tested globally yet, but it doesn't change anything besides JDK9+ handling.

alexyakushev10:10:32

If we could merge the snapshot and use for ~1 week, we would follow with the releases of compliment and new cider-nrepl

bozhidar12:10:02

@alexyakushev Can you only remove the Java 10 change or change this to 11?

bozhidar12:10:13

I’m fine with your proposal in general.

bozhidar12:10:27

Or just remove the build matrix changes.

alexyakushev12:10:04

@bozhidar I didn't quite get what you mean

bozhidar12:10:27

@alexyakushev There’s a commit changing Java 9 to 10.

bozhidar12:10:44

(because Java 10 is dead I don’t see the point on testing on it)

bozhidar12:10:06

The build on travis is super slow even now. 🙂

alexyakushev13:10:10

> Can you only remove the Java 10 change or change this to 11? Then some other CIDER cljs stuff fails.

alexyakushev13:10:13

Are you OK with that?

alexyakushev13:10:35

Anyway, did exactly that

bozhidar17:10:36

@alexyakushev I’ve pushed a new snapshot. The cljs failures are unrelated. I hope we’ll address those eventually.