Fork me on GitHub
#aws
<
2018-12-08
>
kulminaator11:12:01

following the steps on aws-api , i guess with this was not supposed to happen

=> (def s3-client (aws/client {:api :s3}))
2018-12-08 13:02:49.534:INFO::async-dispatch-1: Logging initialized @58389ms
WARNING: An illegal reflective access operation has occurred
WARNING: Illegal reflective access by org.eclipse.jetty.util.BufferUtil (file:/home/martin/.m2/repository/org/eclipse/jetty/jetty-util/9.3.7.v20160115/jetty-util-9.3.7.v20160115.jar) to field java.nio.MappedByteBuffer.fd

kulminaator11:12:51

:thinking_face: that also looks like a rather old jetty version it's trying to use under the hood

kulminaator11:12:37

trying to use the newer jetty utils from 9.3.x or 9.4.x series are broken in different ways

Alex Miller (Clojure team)14:12:37

without using any external deps, you should be getting 9.3.7.v20160115

Alex Miller (Clojure team)14:12:49

do you have other deps in use?

Alex Miller (Clojure team)14:12:19

I guess that’s what you’re seeing above

Alex Miller (Clojure team)14:12:09

With Java 11, I’m not seeing that warning, would be curious what you’re on

Alex Miller (Clojure team)14:12:47

Although the reflector resolution can vary between JDKs and runs so it’s possible I’m just luckier. It’s important to note though that this is a warning, not an error and the code should work just fine.

Alex Miller (Clojure team)14:12:59

I think it’d be helpful to leave a github issue at https://github.com/cognitect-labs/aws-api/issues - certainly worth doing whatever is needed to update the http-client deps

kulminaator16:12:10

i will try to reproduce and give the details there

kulminaator16:12:39

reproduced on a clean sheet , https://github.com/cognitect-labs/aws-api/issues/19 added details here. not in a hurry with this for now, i will add a middle layer into my actual application so i could switch forth and back between aws-api and aws s3 sdk for a while 🙂

kulminaator16:12:20

just in case you're wondering what i'll use the s3 for -> i have to make copies of s3 buckets in a point of time ... buckets that reach a few terabytes and have millions of objects inside. multithreading of clojure comes in handy as a single s3 client on a single thread will never be able to finish stuff like that, even if stuff runs on aws ec2 machines "next to the s3" in the same region. latency is killing me there.