Fork me on GitHub
#leiningen
<
2017-10-25
>
didiercrunch00:10:04

Hi there, I am trying to deploy https://github.com/roryk/clojupyter to clojar and it fails with timeout

didiercrunch00:10:39

does anyone has an idea what can cause this?

didiercrunch00:10:05

does clojars looks at the github's project's owner?

didiercrunch00:10:50

@shaun-mahood it does not work here

didiercrunch00:10:47

it is probably not in the default maven repo

danielcompton00:10:03

https://github.com/jai-imageio/jai-imageio-core is the more up-to-date version of it, but has some things pulled out for license compatibility

shaun-mahood00:10:35

Thanks for checking - I'm going to put it into http://deps.co :)

shaun-mahood00:10:51

I've heard more than once about how great it is that Maven central always works - is this just one of those weird edge cases (maybe licensing related?) where it doesn't?

danielcompton00:10:50

yeah I've never seen that happen before. It's a really old file, so could be some combination of bit-rot + licensing

shaun-mahood00:10:14

Ok that's kind of what I figured. Crystal reports can break anything :)

didiercrunch00:10:27

can someone try to deploy https://github.com/roryk/clojupyter to its own clojars namespace

didiercrunch00:10:08

for an unknown reason I cannot deploy it

didiercrunch00:10:12

it makes me mad

danielcompton01:10:33

weird, open an issue https://github.com/clojars/clojars-web/issues/ if it doesn't resolve itself

danielcompton01:10:55

@didiercrunch can you deploy other libraries?

didiercrunch01:10:33

it looks like it is an issue on my side

didiercrunch01:10:24

the above repo is almost the same thing than clojupyter

danielcompton01:10:58

When deploying a new lib we check if it exists on Maven Central, it's possible that check is failing here?

didiercrunch01:10:07

i do not see it

didiercrunch01:10:42

I try (defproject org.clojars.didiercrunch/clojupyter "0.1.0" and does not work more

didiercrunch01:10:58

org.clojars.didiercrunch/foobar is not working either

didiercrunch01:10:06

what can I do?

noisesmith01:10:10

use a different org name?

noisesmith01:10:24

wait n/m I think I misunderstood

didiercrunch01:10:50

a different org name does not work 😠

danielcompton01:10:42

I don't have time to help at the moment, but open an issue if you can't get this working or try #clojars

didiercrunch01:10:18

I am going to bed right now but I believe my jar is quite large (7M) and takes too much time to upload

didiercrunch01:10:57

I have a quite bad internet connection

didiercrunch01:10:13

is there a way to increase the timeout?

danielcompton03:10:17

And this is happening because it's making an uberjar like thing

didiercrunch11:10:11

apparently I'll need to go to the local coffeeshop to deploy the jar tonight

kennytilton19:10:06

Help, I am being overwhelmed by IntelliJ! 🙂 I just got a new box and installed IntelliJ and managed to again open my project the wrong way so it did not know I was using leiningen. A month ago some kind soul gave me the path to the option where I could fix that on the fly, but now I cannot find that. I re-opened and did it right the second time (I think--indentation is back to working) but I am stubborn and want to find that option (to modify a project to be under leiningen). The Google helpeth not. Any leads appreciated.

bbrinck21:10:42

In lein 2.7.1, I was (perhaps incorrectly?) calling (System/exit 1) to exit lein with exit status 1. This worked fine, but in 2.8.0, lein exits 0 (when when the subprocess fails)

bbrinck21:10:10

Is there a recommended way to make lein exit with a non-zero status?

noisesmith21:10:03

@bbrinck what’s the use case for running your code via lein and also caring about the exit status?

noisesmith21:10:27

by which I mean, for me leiningen is a build tool, and I have lein produce a jar for actual use

noisesmith21:10:49

@bbrinck just a hunch - if you add the trampoline arg, does the behavior get better? many behaviors about the jvm process are improved when trampolining, because without trampoline you are starting two vms, the lein vm (which the shell sees) and the second jvm that actually runs your code

noisesmith21:10:18

my guess is that jvm 1 (with lein) waits on jvm 2 (with your code) and nobody checks or cares what exit status jvm 2 delivers

noisesmith21:10:54

jvm 1, which returns to the shell, is returning 0 no matter what 2 does - though I bet this would be easy to patch if there were a good use case

bbrinck21:10:02

@noisesmith My use case is to fail builds is cljsbuild encounters warnings. On CI, we run cljsbuild to confirm CLJS code compiles. We can register a warning handler to cljsbuild and exit 1 if we get a warning

bbrinck21:10:22

let me try trampoline, that’s a good suggestion

noisesmith21:10:40

trampoline should be just fine for a cljsbuild task too

bbrinck21:10:08

@noisesmith sweet, trampoline worked.

bbrinck21:10:21

I would never have thought of that. Good suggestion

noisesmith21:10:50

see also any weird behavior around suspending the process, or closing / opening stdio

noisesmith21:10:06

trampoline just fixes the things by avoiding the weird process to process indirection

danielcompton21:10:56

@bbrinck we do something similar, this might be a bug in 2.8.0?

bbrinck21:10:09

I can confirm this worked in 2.7.1, but I wasn’t sure if this is a bug in 2.8.0 or if I was doing something unsupported in 2.7.1 🙂

bbrinck21:10:43

(happy to file an issue if it sounds like a bug)

danielcompton22:10:50

yeah I suspect it might be, as it would affect other stuff that runs inside the project that exits with a status. I haven't tested kibit with 2.8.0 but that might be a problem