Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jul 09 15:35
    soronpo commented #186
  • Jul 09 15:35
    soronpo commented #186
  • Jul 09 15:24
    kevinwright commented #186
  • Jul 09 15:23
    kevinwright commented #186
  • Jul 09 14:58
    som-snytt commented #186
  • Jul 09 13:54
    milessabin commented #186
  • Jul 09 13:03
    smarter commented #186
  • Jul 09 13:02
    smarter commented #186
  • Jul 09 12:26
    kevinwright commented #186
  • Jul 09 12:26
    kevinwright commented #186
  • Jul 03 20:38
    som-snytt commented #186
  • Jul 03 20:01
    clhodapp commented #186
  • Jul 03 17:34
    SethTisue commented #186
  • Jul 03 16:09
    sellout commented #186
  • Jul 03 08:49
    odersky commented #186
  • Jul 03 04:45
    gvolpe commented #186
  • Jul 02 17:37
    zainab-ali commented #186
  • Jul 01 18:01
    smarter commented #186
  • Jul 01 16:03
    tpolecat commented #186
  • Jul 01 14:49
    tindzk commented #186
Frank S. Thomas
@fthomas
@Odomontois the Scala versions are defined by sbt-travisci via .travis.yml
Bjørn Madsen
@aeons
hey, I get this error:
no-symbol does not have a type constructor (this may indicate scalac cannot find fundamental classes)
is that something that has a “well, duh” answer or should I work on getting a minimal example?
Paul Phillips
@paulp
@aeons The parenthetical is one of those ad hoc error message adjustments made because that's the way things explode when it can't find either scala-reflect or scala-compiler on the classpath, can't remember which. But if nothing precedes that error, that's probably the reason.
Or even scala-library. Something something classpath.
Bjørn Madsen
@aeons
okay
that is super weird because it started appearing suddenly after adding some code, but not changing any build configuration
so, changing to scalac and back to typelevel scala made it go away...
Paul Phillips
@paulp
Did you try a full clean before that
Bjørn Madsen
@aeons
yes, several times
and also manually deleting target/ and project/target/
Paul Phillips
@paulp
The failure happens during compilation?
Bjørn Madsen
@aeons
yes
it says it’s at a line like
case class Service[F[_]: Monad](dependency: Dependecy[F])
with call site: value dependency in package…
Paul Phillips
@paulp
@Odomontois it means the typelevel changes caused the "is this implicit better than that implicit" order relation to become intransitive.
Edmund Noble
@edmundnoble
The SI-2712 fix?
Paul Phillips
@paulp
No, the -Yinduction-heuristics
Edmund Noble
@edmundnoble
Oh shit. I didn't think it changed behavior.
Paul Phillips
@paulp
I'm sure it wasn't intended to.
Edmund Noble
@edmundnoble
I feel like we never understood implicit resolution enough to change it :/
Paul Phillips
@paulp
Simply inspecting values in scalac is easily enough to change behavior.
Edmund Noble
@edmundnoble
Then replace "implicit resolution" with "scalac"
Paul Phillips
@paulp
@aeons it's been too long for me, so I can only say a rigorous inspection of what's on your classpath is likely to lead to the answer
Bjørn Madsen
@aeons
okay thanks, I will try that if it pops up again :)
Miles Sabin
@milessabin
@Odomontois do you have a small reproduction of that IllegalArgumentException?
Odomontois
@Odomontois
@milessabin I'v very sorry, but like the last time, I can't reproduce it without circe and akka-http implicits. It somehow works on all simplified examples.
Miles Sabin
@milessabin
@Odomontois if you can that'd be very helpful. If you can create a ticket on typelevel/scala with that stacktrace that'd be useful too :-)
Odomontois
@Odomontois
@milessabin Ok. I'll do my best
Miles Sabin
@milessabin
Many thanks :-)
Andy Scott
@andyscott

SBT seems happy with

scalaOrganization := "org.typelevel",
scalaVersion      := "2.12.1",

but very unhappy with

scalaOrganization := "org.typelevel",
scalaVersion      := "2.12.2-bin-typelevel-4”,

(unable to resolve scala compiler dependency, etc).

What’s the silly mistake I’m making?
Matthew Scharley
@mscharley
pretty sure there's a separate option for the bin suffix, but I forget what now sorry
Seth Tisue
@SethTisue
@andyscott what sbt version?
I forget when sbt’s knowledge of the -bin- thing was added, but it wasn’t toooo long ago
Andy Scott
@andyscott
sbt.version=0.13.15
Miles Sabin
@milessabin
@andyscott there must be some more context for your problem. See my tls-demo for instance: https://github.com/milessabin/tls-demo/blob/master/build.sbt
Dale Wijnand
@dwijnand
Use of coursier for instance.
Or I should say a version of coursier that predates recent typelevel fixes
Torsten Schmits
@tek
is there a way to manually change a tree's owner in a macro?
Torsten Schmits
@tek
nvm, seems that's not the problem
Hamish Dickson
@hamishdickson
heya, would you expect this exhaustiveness warning with -Yliteral-types enabled?
scala> val t: true = true
t: true = true

scala> t match {
     | case true => 1
     | }
<console>:13: warning: match may not be exhaustive.
It would fail on the following input: false
       t match {
       ^
res0: Int = 1
Alexander Konovalov
@alexknvl
I would not, but Scala compiler is a mystery.
;)
Rob Norris
@tpolecat
Would it make sense for a GADT match on a singleton type to be able to materialize the value on the RHS?
It can't now but it seems like it would be ok to do so.
Well I guess Nothing would mess it up.
Edmund Noble
@edmundnoble
Yeah that's the subtle difference between case O => O and case o: O.type =>. Equals is used for one, eq for the other.
I don't think Nothing would mess it up nearly as much as an overriden .equals.
I think :confused: I'm not sure and not at my computer.