Fork me on GitHub
#cider
<
2018-05-04
>
bozhidar15:05:34

@gastove I’m certain that the latest master should work properly. The bug was in orchard, btw. Maybe the auto-deployment failed or something like this.

dpsutton16:05:03

@bhauman just responded on your ticket but (setq cider-font-lock-dynamically nil) and see if that speeds up your buffer

gastove17:05:22

@bozhidar okay, got it, thanks 🙂 I'll either install latest master or watch for the next cider release 👍

bozhidar21:05:05

@gastove I just checked on clojars and I see the version of orchard there is the right one. I wonder if some reason your snapshot deps are out of date.

gastove21:05:33

I've got... at least three, maybe four versions of the cider-nrepl jar

gastove21:05:13

I don't have any jars with names at all like orchard, but I don't know if I should

gastove21:05:29

(At least, not in my m2 repository)

bozhidar22:05:18

Ah, now I understand what happened. As orchard is not a normal dependency for cider-nrepl, but an inlined one, I also need to cut a new snapshot release of cider-nrepl manually.

bozhidar22:05:26

Mystery solved!

bozhidar22:05:52

> @bhauman just responded on your ticket but (setq cider-font-lock-dynamically nil) and see if that speeds up your buffer

bozhidar22:05:09

@dpsutton is probably right, there’s also some font-locking for reader conditionals that can also cause such problems. I wonder why the problem becomes worse with time, though. My guess is that we’re not invalidating some ns data (which we use for font-locking) and it grows huge.

bozhidar22:05:19

@gastove Should be OK now, just pushed a new snapshot.

dpsutton22:05:56

i was in a 1800 line file and it was terrible. i turned it off and it was night and day. i think this is a buffer size issue not cljc issue

bozhidar22:05:04

Hmm, I wonder why, though. The generated font-locking rules should be just as effective as the static ones, and most people haven’t complained about this.

bozhidar22:05:49

Anyways, I’m certain that part of the code can be optimized. It almost hasn’t changed since @malabarba “retired”.

dpsutton22:05:46

for sure. i'd be interested in tracking down what's going on. also figuring out imortant locals are. for instance, if you shadow a name with a let binding, does cider-find-var understand that new context

dpsutton22:05:00

not sure how important it is to judge what kind of fix is required / allowed

bozhidar22:05:47

Currently CIDER has not idea about locals outside the debugger, but many people have asked for this to be improved.

bozhidar22:05:26

At the very least when you do find-var it should tell you that something is a local and not try to resolve it as a var.

mikerod23:05:21

Sounds like a job for tools.analyzer

dpsutton23:05:23

be cool to play around with the graal vm native compiled stuff