morning.

morning

morning.

Some peeps are having a real bad day....

https://twitter.com/gitlabstatus/status/826591961444384768

Yeah... I was glad to read on the more detailed status page(s) that the repos were not affected, just Issues and Merge Requests, but still...

Morning, btw 🙂

It's nice when companies fuck up, are open about it, and because of that customers support them.

Moaning

"So in other words, out of 5 backup/replication techniques deployed none are working reliably or set up in the first place.” was just reading through that doc

#fuckcomputers

Morning @jonpither - How are things going with your move to Arch?

@paulspencerwilliams - I agree, transparency is the key at times like this; Gitlab will be fine because they are being clear about what the problem is and how they intend to fix it, so I expect that their customers will feel pretty much that they want to stick around.

Last time I was caught up in a Github snafu the comms were good but not this good...

...and even then the fact that there were comms made me a lot happier.

Hi @maleghast I'm loving it

when I moved from the Mac to Ubuntu, it was nice, but had some niggles, and I wasn't like 'wow, this is amazing'

but moving to Arch and I3, going much more minimal but using a proper window/tiling manager, it just fits the way I want to work

@jonpither - Glad to hear it! Once I am in a position to have a desktop / base somewhere I will be adopting Arch as my main work OS

I like numbered windows and sane tiles inside

Sounds very... comforting

😉

(I'm not taking the piss)

yeah. I won't lie to you, having @dominicm almost sitting next to me has made the transition very seamless

Also good to know - when I set up a workstation I will have my MBP to fall back on... 😉

but if you do try it, ping us away with Qs

Well, I am coming back to the UK, sooner than expected, and once we are settled and I've got some money coming in there will be a desktop machine and a desk and chair just for me for work and play. It's going to have a boot disk for Arch and (sadly) one for Win10 for gaming.

wow big news, you're coming back?

woohoo

I love working on my Mac, but it's going to be my "on the road" machine and my photography and media machine.

Yep, we will be coming back to the UK by way of a vacation in Japan in March.

cool, you've been missed 🙂

Aww, thanks - I have missed the Clojure community in general and you Juxt fellows and a certain Hugmonster General and a few others in particular.

wear an extra padded coat for the hug you'll get off @otfrom

No padding - definitely want to feel it 😉

I love those emoticons - they are splendid.

@maleghast If you buy a dedicated gfx card for gaming, you can instruct linux to not touch it, and do a direct pass through to a VM. Given that CPUs are not a particular bottleneck in gaming anymore (especially with virtualization being supported first class in the CPU) you may find that it's more than fast enough to game on Linux.

@dominicm - That is VERY interesting...

I had zero idea that that was even an option

Thanks!

https://www.reddit.com/r/pcmasterrace/comments/2z0evz/gpu_passthrough_or_how_to_play_any_game_at_near/ @maleghast

95% he claims.

@dominicm - 👀

Looks like you must dedicate an input to a monitor / have a second monitor. Only real drawback, not too bad really.

Yep, that looks very interesting. If nothing else it would also provide a Linux / Windows dev environment that could basically just run all the time, allowing me to do website testing on shudder IE without starting up a VM and all of that...

(minor benefit)

But most importantly it's Windows gaming without re-booting

🙌

Thx!

No problem!

morning

Bore da

(case (type "")
  java.lang.String
  "foo")

(let [x (type "")]
  (cond
    (= x java.lang.String)
    "foo"))

(case 'foo
  foo
  "foo")

Odd.

case doesn't evaluate test exprs

so you're trying to match String the class to java.lang.String the symbol

@bronsa yeah that's what I was thinking. Must admit that I'm finding some of the behaviours of case quite confusing.

well, case is optimized for constant time dispatching, evaluation test expressions would prevent that

@bronsa ah, didn't realise that was the decision. I thought it might have been a trade-off in order to support the (:a :b) syntax.

you can kind of cheat it by doing (case (type "") #=java.lang.String "foo" "bar")

but don't do that ;)

@bronsa Never seen #= before, where can I read about it?

I've cheated by swapping to condp for now

no idea if it's documented anywhere

it's read-eval

if you're familiar with common lisp, it's kind of like #.

except it's its shitty cousin that only supports a bunch of expressions, not true eval

Yeah, just tried playing with it & noticed that...

yeah, don't ever use it :) it's only in clojure for serialization purposes

@bronsa is this knowledge you've picked up doing tools.analyzer?

nah

well kinda

reading through the clojure compiler and impl in general

Explained a little bit of that in this: https://manning-content.s3.amazonaws.com/download/c/5f7f93e-be8a-4b4e-a679-8171b1e761ff/Clojure_Standard_Library_v1_MEAP_sample-function-2.pdf

but didn’t know about the @bronsa trick 🙂

that's all it is really, a show-off trick, I would never dare use it for real :)

case is a beautiful engineering piece BTW. It just takes some knowledge to use correctly

esp because it.. might break in spectacular ways under AOT

@bronsa, if I mention it I’ll do it with your name on the side 🙂

no worries, not going to

not sure I want to take responsibility for people knowing about it :P

it can break because case does a hashCode dispatch (for those types of objects), inlining at the bytecode level the hashcode of the class at compile time -- if on some other JDK the hashcode for that class is different, it's not going to match

Right, good to know. Thanks for clarifying.

hi hi

More n00b questions 🙂

(let [drop-services
        {:alsmp
         {:color :red
          :ips ["10.206.6.145"]
          :cmd "journalctl -xe --since -1min --no-pager -u dropwizard_alsmp-service -f"}
         :alsapi
         {:color :green
          :ips ["10.206.6.245"
                "10.206.6.42"]
          :cmd "journalctl -xe --since -1min --no-pager -u dropwizard_alsapi-service -f"}
         :dmdis
         {:color :blue
          :ips ["10.206.5.239"]
          :cmd "journalctl -xe --since -1min --no-pager -u dropwizard_dmdis-service -f"}
         :mips
         {:color :yellow
          :ips ["10.206.6.33"]
          :cmd "journalctl -xe --since -1min --no-pager -u dropwizard_mips-service -f"}}
        cm-infra
        {:all
         {:color :red
          :ips ["10.204.25.240"
                "10.204.26.33"
                "10.204.24.47"
                "10.204.29.23"

I'm wondering about splitting that out into a config file.

But it's like a dsl

So what do people tend to do?

That hash feeds into a function that basically multiplexes ssh connections so I can watch logs in realtime.

I don't think it is a DSL, it's just plain ol' data, in a specific form

yeah

I'm debating with myself to make a dsl.

versus just passing in the data

data > functions > macros

i.e

(journalctl dropwizard_alsmp ["10.206.6.145", "10.206.6.42"] :red)

But I'd find it odd to edit my "code" between runs rather than changing a config file.

hmmm

Anything I can read around this sort of thing?

Maybe this https://martinfowler.com/dsl.html

(with-awsprofile profile
  (with-env sit ;; instances with env tag set to sit
    (stream-dropwizard-service alsmp) ;; instances with service tag set to alsmp
    (stream-dropwizard-service alspi)
    (stream-dropwizard-service dmdis)))

I shall take my thinking offline where it belongs 🙂

@mattford config files, we tend to store maps as something called edn files

have a look at https://github.com/juxt/aero

This means you can have mulitple configs for different profile types

https://github.com/jasebell/onyx-kafka-demo/blob/master/resources/config.edn for an example

@mattford (and your thinking is perfectly fine here btw!)

@mattford also worth looking at https://github.com/weavejester/environ for reading environment variables via clojure.