Fork me on GitHub
#cljs-dev
<
2018-02-24
>
mfikes01:02:10

Our AppVeyor (Windows CI) tests are failing owing to an evident inability to securely download lein. I don't have any ideas on this one and may temporarily turn it off (in my GitHub ClojureScript fork) if we can't get them working again https://ci.appveyor.com/project/mfikes/clojurescript/build/1.0.412

dnolen14:02:58

to try out pREPL on master

dnolen14:02:13

this is pretty cool 🙂

dnolen15:02:58

and I verified that clojure.core.server/start-server just works with io-prepl

mfikes16:02:36

On master, the Nashorn script/test seems to get blocked for me in cljs.pprint-test. (I'm running Java 9.) I don't recall seeing this before. Hrm.

mfikes16:02:11

I may dig into it later, but here is my thread dump https://gist.github.com/mfikes/9a3f77dfce981f5359825d1990c7ad0b

dnolen17:02:19

@bhauman I fixed the watch + REPL issue on master - but note the original report wasn’t quite right - you can’t combine -m with -r I think you intended -c, I edited the issue to reflect that

bhauman17:02:59

wow!!! this is gettting really good

bhauman17:02:53

@dnolen and now to ask for more: what's the story on serving a default html??

bhauman17:02:13

for the browser repl

dnolen17:02:29

you would need to do that in browser REPL, I’m probably not going to spend time on that myself though

bhauman17:02:56

you'd be open to a potential solution for that?

bhauman17:02:20

the idea that the browser repl could just launch with no fiddling whatsoever could be fantastic for testing stuff out

dnolen17:02:21

I think it might be better to generate a default index.html if it doesn’t exist

bhauman17:02:38

yeah that sounds great as well

bhauman17:02:43

I was debating both

dnolen17:02:47

browser REPL could do this when it generates the client script

dnolen17:02:57

serving a thing the user can’t edit doesn’t make much sense to me

dnolen17:02:54

@bhauman I was skimming over rebel-readline - I think making that work over pREPL would be pretty easy?

bhauman17:02:05

yes it would

bhauman17:02:00

you are thinking about the service level that provides the readline with its extra functionality?

bhauman17:02:48

it would be a good experiment, but it's hard to beat the direct implementation I'm using

dnolen17:02:43

yeah I think you would only need this in the remote case? rebel-readline assumes ClojureScript is in-process?

bhauman17:02:42

yes, it would make great sense for a remote clojurescript repl

bhauman17:02:12

and it would be a great way to kick the tires

bhauman17:02:05

you might even be able to steal the clojure.main/repl loop and override eval, print et

bhauman17:02:26

nah bad idea

bhauman17:02:57

would just need to implement a client repl loop for a remote prepl and then wrap it in rebel readline and create a service for it

dnolen17:02:35

that’s right

bronsa19:02:51

@dnolen releasing 1.3.0-alpha1 with read+string

bronsa19:02:34

requires a source-logging-push-back-reader