Fork me on GitHub
#clojure
<
2018-01-21
>
misha00:01:03

I split it to avoid quoting arg to macro, in the first place, plus to keep form with syntax quote/unquote as small as possible. but now all of it is irrelevant

kenrestivo08:01:07

and also org.slf4j/log4j-over-slf4j and org.slf4j/slf4j-simple IIRC

mike_ananev11:01:35

Is it interesting for clojure world?

bitti11:01:31

Regarding my question from yesterday I already figured out that twitter-api is using http.async.client which is using com.ning/async-http-client which is using io.netty/netty which is using com.ning.http.client.providers.netty.NettyAsyncHttpProvider to make async calls (when io.netty is in the classpath) which can prevent the application from closing when the NettyAsyncHttpProvider is not closed properly as per https://stackoverflow.com/questions/30534600/how-to-close-asynchttpclient-with-netty-for-an-asynchronous-http-request.

bitti11:01:45

Now I'm not sure to which of these library owners I should report a bug to

bitti12:01:04

ok, now I finally see that there is already an open issue: https://github.com/adamwynne/twitter-api/issues/74

laujensen13:01:10

Hey! Does anyone have to recent performance benchmarks on webstacks? Im seeing a delay in requests of about 100msecs that I cant explain. The hoops jumped through are nginx -> immutant/jboss -> ring -> moustache -> handler. The handler is 3msecs, the total request 100msecs before it starts streaming

rauh19:01:32

How do I get void.class in clojure?

kwladyka22:01:12

I know that is so evil question here, but… While I love Clojure, I am looking solution which don’t use so much memory to use on machines with very low memory. Which solution (language / framework / whatever) could i like then?

sundarj22:01:01

never used it, so take that with much salt

ghadi22:01:27

How much memory @kwladyka?

kwladyka22:01:57

not sure yet, but i am thinking about software for GPUs mining

kwladyka22:01:16

probably in docker

ghadi22:01:01

More than the memory, GPU programming is totally different as a paradigm

kwladyka22:01:38

yeah, but for now i am thinking about software to automate all things, like monitoring, temperature protection, bios updates. Like plug & play. So it is more about software around GPU mining, then directly software to mining.

kwladyka22:01:22

but while we are in topic do you recommend some languages for it?

carocad11:01:00

@kwladyka check https://github.com/uncomplicate for GPU-related Clojure libraries. Those are the best to the best of my knowledge

carocad11:01:53

as for low memory devices there is: https://github.com/nakkaya/ferret but I have never tried it 😄

kwladyka22:01:58

Can you add a few word why rust / lua?

sundarj22:01:04

from what little i know, rust is like human-friendlier C

kwladyka22:01:38

heh sounds good whatever it means in practice 😉

jgh22:01:42

rust is more like a safer c++

kwladyka22:01:54

Is it like Clojure -> Java? Rust -> C++?

jgh22:01:59

“human friendly” is perhaps debatable

sundarj22:01:10

haha fair enough

jgh22:01:25

no i wouldnt say that, it’s just another systems language that fills a similar niche to c++

sundarj22:01:43

i defer to your knowledge 🙂

jgh22:01:13

go might be a decent option, but it depends on your latency requirements etc since go is garbage collected (similar to jvm)

jgh22:01:40

and also doesnt run well on 32-bit arm systems (not sure if that’s where you’re going with this low memory requirement)

kwladyka22:01:56

i don’t care about 32-bit in that case

genmeblog22:01:27

lua is used in embedded systems

jgh22:01:56

oh you’re looking to do gpu stuff

kwladyka22:01:23

Any other language, than rust / lua worth attention? golang?

jgh22:01:28

c or c++ will give you the best support, other languages will have varying levels of support.

jgh22:01:39

since you’ll be working with cuda or opencl or something like that

kwladyka22:01:29

Hmm not really sure how it work in practice. On first fire i would like to detect overheat, monitoring if everything works fine and if not notify, update GPUs bios, etc. I don’t think i have to use C++ for this things around GPU mining

kwladyka22:01:06

But Java / Clojure in my intuition don’t look good, because of memory consumption.

kwladyka22:01:29

but maybe i am totally wrong

jgh22:01:40

oh you’re just doing monitoring

jgh22:01:05

yeah im sure you could do that with like python or lua or whatever.

kwladyka22:01:05

more automation, than monitoring

jgh22:01:43

python probably the most straightforward since it has a big ecosystem and is well supported

kwladyka23:01:39

yeah, maybe it is naive, but as Clojure fun i was looking something which can give me more pleasure, than Python 😉

jgh23:01:14

heh…go would probably do the trick too 😉

kwladyka23:01:29

thx, for discussion. It is time to go sleep. Goodnight 🙂

twashing23:01:41

@ymilky Was hoping to ask you a quick and simple question around #franzy. I have kafka (and zookeeper) running in a docker-compose setup.

twashing23:01:43

If I can produce and consume using kafka tools (confluentinc/cp-kafka:3.3.1), what leap do I need to make, to subscribe (or manually consume) from the same topic, using franzy?

twashing23:01:57

Using the examples here, messages that are definitely on a topic, are not being consumed, when using franzy. https://github.com/ymilky/franzy#subscription-based-consumer