Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Nov 08 01:01

    weihsiu on master

    Add slides of How to contribute… Add slides of Scala & Spark(1.6… Merge pull request #28 from jim… (compare)

  • Nov 08 01:01
    weihsiu closed #28
  • Nov 07 09:28
    jiminhsieh opened #28
  • Nov 07 03:48
    weihsiu commented #27
  • Nov 07 03:48

    weihsiu on master

    Move 2016's slides to folder 20… Move 2017's slides to folder 20… Move 2018's slides to folder 20… and 3 more (compare)

  • Nov 07 03:48
    weihsiu closed #27
  • Nov 07 02:48
    jiminhsieh opened #27
  • Sep 09 18:41

    pishen on master

    Add the slides of meetup #32 (compare)

  • Sep 08 15:01

    pishen on master

    Add the slides of meetup #31 (compare)

  • May 03 04:20

    pishen on master

    fix-number-length-for-order fix corresponding link Merge pull request #26 from Wei… (compare)

  • May 03 04:20
    pishen closed #26
  • May 03 04:17
    Wei-1 synchronize #26
  • May 03 04:16
    Wei-1 opened #26
  • Mar 29 07:03

    weihsiu on bengal

    (compare)

  • Mar 29 07:03

    weihsiu on master

    bengal initial commit Merge pull request #25 from Sca… (compare)

  • Mar 29 07:03
    weihsiu closed #25
  • Mar 29 07:03
    weihsiu opened #25
  • Mar 29 07:02

    weihsiu on bengal

    bengal initial commit (compare)

  • Feb 27 06:02

    pishen on master

    update ToC (compare)

  • Feb 27 05:58

    pishen on master

    add 2019-2-26 talk of wilbercaho Merge pull request #24 from cec… (compare)

jirijakes
@jirijakes
Or we can just drink whole night and then go directly to Matsuri :wine_glass:
Walter Chang
@weihsiu
Talk is cheap, jiri!
jirijakes
@jirijakes
Stream.fromEffect(drink).repeat(Schedule.doUntil("2020-05-15 08:30")).run
… here you have code.
Walter Chang
@weihsiu
just came across this, new encodings in scala that let you do universal and existential quantification of types. very cool! https://github.com/djspiewak/skolems
Yan
@yaroot
I need some help for finishing the api for my redis driver. the major issue is explicit pipelining and transaction, does it make sense to do make F explicit (eg, RedisIO a, get:: String -> RedisIO String, and then will be a run:: Pool -> RedisIO a -> IO a), does it make sense or it's better just having a unsafe pipeline api, then everything would be just IO
Walter Chang
@weihsiu
I like RedisIO a lot better. I can use it in my own code that's not IO-based and it's much cleaner than doing liftIO all over the place.
jirijakes
@jirijakes
Yeah, safe way with RedisIO and run should be better. And safer :-)
Yan
@yaroot
@weihsiu I'm thinking something like ConnectionIO from doobie, or ProgramT & friends from haskell. I'm sure IO (replace all IO in this converrsation with at least cats.effect.Effect) would be required like doobie's transactor
Walter Chang
@weihsiu
@yaroot r u using free?
Yan
@yaroot
is there a choice?
I haven't done anything yet
Yan
@yaroot
spent some time tinkering with cats.free, it doesn't work very well in this case
jirijakes
@jirijakes
Will you share more details, please? Curious…
Yan
@yaroot
applicative free might work, let me play with it a bit
Walter Chang
@weihsiu
beckoning to his pascal past perhaps!! lampepfl/dotty#7598
Walter Chang
@weihsiu
Hey, guys, I heard just heard that Wicked Wolfpack (the place we hold our Mokumoku) has got some new beers. So it's time for another one of our Scala Taiwan Drinkups. It will be this Saturday(11/30) afternoon at around 3pm. Please come by and join us. Wicked Wolfpack
No. 29, Section 7, Zhongshan North Road, Shilin District, Taipei City, 111
070 1001 7614
https://maps.app.goo.gl/b3i4DmCVZ9qHdMUa9
Yan
@yaroot
when kind-projector and scalac blows up _ >: (some other)β$13$(in type Λ$) with (some other)β$9$(in type Λ$) with β$9$(in type Λ$), even asInstanceOf won't save me
anyone know how to turn off the type checker for a single line :)
Walter Chang
@weihsiu
Yan
@yaroot
Thanks, I'll try it later. I got the error while composing 2 natural transformation which should result TupleK, the 2 nat functions already have several type params, then it got blow up in nat1 and nat2, and 1 type param got expended to the one I past above, even though it is invariant (covariant on my first try, same error)
intellij typechecks though, to my surprise, or maybe it just gives up trying
Walter Chang
@weihsiu
@yaroot vscode with metals ftw
Yan
@yaroot
let's say I don't like vscode, have my eye on onivim 2 though
Yan
@yaroot
crazy, nat2 and nat1, that's it
Yan
@yaroot
redis.set("foo", "bar") *> redis.get("foo"): RedisIO[Option[String]]
and when it runs
127.0.0.1:6379> MONITOR
OK
1574772082.589258 [0 127.0.0.1:30616] "SET" "foo" "bar"
1574772082.589276 [0 127.0.0.1:30616] "GET" "foo"
Yan
@yaroot
wow, openj9 uses significantly less memory than hotspot, without too much performance hit
Jimin Hsieh
@jiminhsieh
Yeah, openj9 is faster than hotspot JIT at startup. In the meantime it has smaller memory footprint.
But be careful that some of libaries are targeting for hotspot. You will get runtime exceptino when you are using openj9.
I met this issue at Dotty once.
lampepfl/dotty#4430
Jimin Hsieh
@jiminhsieh
Paolo G. Giarrusso helped me to report the issue since I am not so sure it’s a feature or a bug. 🤣
Yan
@yaroot
I always build with hotspot. There's 1 issue with a library that relies on hotspot quirks, other than that all went well.
Jimin Hsieh
@jiminhsieh
I just think about one thing is that I used OpenJ9 to compiling a Scala project once. The result is OpenJ9 will need more time than hotspot.(hotsport has better throughput) I remember that OpenJ9 will need 1.25 times of hotsopt compilation time. (The number probably is wrong)
Yan
@yaroot
luckily I got 2 very beefy CI machine so I would never worry about it. my major concern is however runtime cost, I've ported 1 app to haskell and it uses 1/3 of ram and cpu and it took less than a minute to build, versus minutes (compile/test/upload)
Vito Jeng
@vitojeng
一直以來都以為 akka stream java dsl 是用 java 寫的, 直到剛才真的去瞄了一下 source code...
才知道原來 被騙 誤會了那麼久...😅
Walter Chang
@weihsiu
@vitojeng hahaha
Walter Chang
@weihsiu
Just a friendly reminder that our year end drinkup is scheduled for tomorrow @ 3pm at Wicked Wolfpack. I heard from some unreliable source that they've got some new beers in. Looking forward to seeing you there!

Wicked Wolfpack

No. 29, Section 7, Zhongshan North Road, Shilin District, Taipei City, 111
070 1001 7614

https://g.co/kgs/1wKXFV

Eddie Cho
@EddieChoCho
Wow! Sounds cool! I would like to join the drinkup!
Walter Chang
@weihsiu
please come!
Yan
@yaroot
[info] Formatting 1 Scala sources...
[success] Total time: 0 s, completed Dec 1, 2019, 9:30:17 PM
so the new sbt-scalafmt uses the graal compiled binary?!
Walter Chang
@weihsiu
@yaroot how long did it take before the new release? i think as a sbt plugin, it is not possible to compile to a native graal image.
Yan
@yaroot
@weihsiu it could use the one already on maven central. also I'm seeing error logs like this
Thread[JVMCI CompilerThread3,9,system]: Compilation of scala.collection.immutable.Vector.appended(Object) failed: org.graalvm.compiler.java.BytecodeParser$BytecodeParserError: java.lang.NullPointerException
    at parsing scala.collection.immutable.VectorPointer.gotoPosWritable1(Vector.scala:1169)
    at jdk.internal.vm.compiler@11.0.5/org.graalvm.compiler.java.BytecodeParser.throwParserError(BytecodeParser.java:2400)
    at jdk.internal.vm.compiler@11.0.5/org.graalvm.compiler.java.BytecodeParser.iterateBytecodesForBlock(BytecodeParser.java:3140)
    at jdk.internal.vm.compiler@11.0.5/org.graalvm.compiler.java.BytecodeParser.processBlock(BytecodeParser.java:2944)
Walter Chang
@weihsiu
the annual Advent of Code has started. very good exercises to sharpen your scala skills and have fun at the same time. bonus point: do them all without using mutable state. https://adventofcode.com/2019
jirijakes
@jirijakes
Additionally, Chriss Penner who is releasing a book about optics, attempts to write AoC only with use of lenses: https://chrispenner.ca/tag/haskell
And Rúnar Bjarnasson writes AoC in Unison (new FP language, could be pretty cool!): https://gist.github.com/runarorama (another Unison AoC: https://github.com/benfradet/aoc-2019)
For inspiration, also Bartosz Milewski and Haskell: https://github.com/BartoszMilewski/AoC2019
Walter Chang
@weihsiu
I am writing it in dotty and rust ;)
jirijakes
@jirijakes
Whooo! I'd be curious about the dotty one… How does it look? Is it very different from predotty? I guess you don't really need much advanced features, right?
Walter Chang
@weihsiu
Not much different since the solutions are so simple. Just getting used to the new syntax is all.
Yan
@yaroot
I thought by using dotty you mean type level encoding
Walter Chang
@weihsiu
I will probably need Idris for that