-
Notifications
You must be signed in to change notification settings - Fork 267
Conversation
This PR will be mergeable as soon as we release scalding 0.17.1 with fix for |
Now, when scalding release is out we can merge this 2.12 upgrade for SB. What do you think @pankajroark @piyushnarang @johnynek ? |
Update: false alarm, tests are failing, digging into this now. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me once you have the build green.
@@ -188,7 +185,7 @@ def youngestForwardCompatible(subProj: String) = | |||
// Uncomment after release. | |||
// Some(subProj) | |||
// .filterNot(unreleasedModules.contains(_)) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
should we delete these commented lines?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
no, we need to uncomment after each version bump. Deleting seems like a good way to keep making mima mistakes like the scalding one.
looks like a legit failure: https://travis-ci.org/twitter/summingbird/jobs/252265982#L3657
This could be that we don't have the right chill version? It could be this exposes a bug in chill. |
@johnynek yes, it is legit =( Unfortunately this time it comes from scalacheck and it's same issue as we saw in chill, yes - too many lambdas in one class. I tried to workaround this by creating this lambdas in summingbird code but there are couple of things which makes everything more complicated - they declare everything as |
@ttim I'm happy to help get this fixed, either in this repo or ScalaCheck? What were you planning to do to work around this? Restructure the ScalaCheck code a bit? Create your own custom Gen instances? |
@non that's awesome! If you can also help with releasing new version it would be twice awesome =) In short term for SB I thought just to add code like this:
But if we can do release reasonably soon I will do PR to scalacheck as soon as possible (today/tomorrow). Plan is to add a test on that + divide |
@ttim My suggestion is that if the above work-around solves the issue for Summingbird we start there. I'm happy to try to expedite the suggested change to ScalaCheck, but unfortunately the arities files are traits right now, so restructuring them will break binary compatibility. This doesn't mean we can't do it, but might make a quick release more difficult. |
(Also, if you'd prefer help I can make a branch off of this one and try the approach out myself later tonight.) |
@non let's do this then. I'll merge my workaround. Meanwhile: this is a scala issue - https://issues.scala-lang.org/browse/SI-10232.
Causes:
|
@non created a bug in scalacheck repo typelevel/scalacheck#342, thank you very much for looking on this! |
build.sbt
Outdated
val hadoopVersion = "1.2.1" | ||
val junitVersion = "4.11" | ||
val log4jVersion = "1.2.16" | ||
val novocodeJunitVersion = "0.10" | ||
val scalaCheckVersion = "1.13.4" | ||
val scalatestVersion = "3.0.1" | ||
val scaldingVersion = "0.16.1-RC3" | ||
val scaldingVersion = "0.17.1" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can we roll back to 0.17.0
here or wait for 0.17.2
due to the recently discovered semver issue with 0.17.1
cc @piyushnarang
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
0.17.0
breaks SB tests so we can't roll back. I'm going to publish 0.17.2
today.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
comment on scalding versioning, but otherwise looks fine.
One other thing to note: it seems like the bug suggests that using (Thanks for giving me a repro -- I'll start seeing what I have to do to fix this issue in ScalaCheck.) |
@non I'm afraid |
I've got a fix on the ScalaCheck side that should let us publish an 0.13.6 release there that will resolve the issue without the workaround here. |
Scalacheck release isn't out therefore I think to merge this PR now, and I created issue to remove it later as soon as scalacheck will be released - #741. |
I think merging this now and removing the workaround later sounds reasonable, but I think @non is planning to publish the ScalaCheck release either this weekend or next week. |
Codecov Report
@@ Coverage Diff @@
## develop #724 +/- ##
===========================================
+ Coverage 72.21% 72.26% +0.04%
===========================================
Files 153 154 +1
Lines 3736 3742 +6
Branches 204 209 +5
===========================================
+ Hits 2698 2704 +6
Misses 1038 1038
Continue to review full report at Codecov.
|
@travisbrown-stripe I prefer to keep it merged and do not put any pressure on scalacheck =) I'll create removal PR immediately after merge thou. |
This PR bumps dependencies versions to most recent ones with Scala 2.12 support.
Algebird fixes based on #715.
This PR also removes support for Scala 2.10, because scalding doesn't support it.