Skip to content

publish for Scala 2.12.0-RC1? #174

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
SethTisue opened this issue Sep 20, 2016 · 18 comments
Closed

publish for Scala 2.12.0-RC1? #174

SethTisue opened this issue Sep 20, 2016 · 18 comments
Assignees

Comments

@SethTisue
Copy link
Contributor

perhaps useful, Adriaan's changes to make it work in the Scala 2.12 community build:

@kevinoliver
Copy link
Contributor

Thanks @SethTisue. I'll see about getting those patches incorporated.

kevinoliver referenced this issue in adriaanm/util Sep 20, 2016
Resolves name clash of private[this] var with protected def
@kevinoliver
Copy link
Contributor

Ok so adriaanm@17625f4 got sorted out in 90f2180

And a patch for adriaanm@2bb041a is on its way.

Thanks again for the heads up.

@kevinoliver
Copy link
Contributor

Cool. The fix for adriaanm@2bb041a just got merged internally and will get pushed to the develop branch this Monday. Thanks again.

finaglehelper pushed a commit that referenced this issue Sep 26, 2016
Problem / Solution

#174 noted that
`c.t.concurrent.ConcurrentMultiMap` does not work with Scala 2.12 due
to extending `Tuple2`. This uses the suggested workaround of extending
`Product2`.

RB_ID=873604
@SethTisue
Copy link
Contributor Author

the Scala 2.12 community build now tracks your develop branch, as of scala/community-build#307

you might still like to publish artifacts for 2.12.0-RC1?

@kevinoliver
Copy link
Contributor

thanks!

@mosesn, looks like we should move to 2.12.0-RC1

@mosesn
Copy link
Contributor

mosesn commented Sep 27, 2016

👍 yeah, I'm hoping to get to it next week. tried a couple weeks ago but some of our deps weren't publishing for it yet.

@mosesn
Copy link
Contributor

mosesn commented Oct 4, 2016

@SethTisue jsyk, right now we're blocked on upgrading scalatest to 3.x internally, since it looks like they've stopped publishing 2.x versions for 2.12. If the migration is painful, we might lobby scalatest to temporarily start publishing 2.x versions again.

@SethTisue
Copy link
Contributor Author

we might lobby scalatest

that might be easier now that Scala 2.12.0 JARs are on Maven Central

@stuhood
Copy link

stuhood commented Oct 31, 2016

We've leaned in on the scalatest 3.x upgrade, and are nearly finished there. Possibly this week.

@jvican
Copy link

jvican commented Nov 9, 2016

Looking forward to the release of twitter/util for 2.12.0.

@kevinoliver
Copy link
Contributor

@jcrossley and others have been working on this and it looks like it should be ready real soon now.

@SethTisue
Copy link
Contributor Author

over at scala/community-build#399 I'm seeing

[twitter-util] [info] ConfigTest:
[twitter-util] [info] Config
[twitter-util] [info] - should computed should delay evaluation
[twitter-util] [info] - should subclass can override indepedent var for use in dependent var
[twitter-util] [info]   should missingValues
[twitter-util] [info]   - should must return empty Seq when no values are missing
[twitter-util] [info]   - should must find top-level missing values
[twitter-util] [info]   - should must find top-level and nested missing values
[twitter-util] [info]   - should must find nested missing values in optional sub-configs *** FAILED ***
[twitter-util] [info]     List("$anonfun$new$16.w", "baz.w") did not equal List("baz.w") (ConfigTest.scala:84)
[twitter-util] [error] Failed tests:
[twitter-util] [error]  com.twitter.util.ConfigTest

@mosesn
Copy link
Contributor

mosesn commented Nov 17, 2016

@SethTisue we haven't changed anything in Config in a very long time, so I'm a little suspicious.

@SethTisue
Copy link
Contributor Author

SethTisue commented Nov 17, 2016

well, this is on a recently nightly of Scala 2.12.1. offhand, it looks like it might have to do with toString on lambdas giving a different answer in 2.12 versus 2.11?

it's certainly possible there's been some regression on our end, but I'd need your help to determine that; I can't afford to go independently investigating in a totally unfamiliar codebase

@SethTisue
Copy link
Contributor Author

the specific nightly used was 2.12.1-371bc2c-nightly. if you want to try it yourself locally, http://stackoverflow.com/q/40622878/86485

@vkostyukov
Copy link
Contributor

@SethTisue @mosesn I'll look into that. Let me try to reproduce that locally first.

@kevinoliver
Copy link
Contributor

ok i can reproduce this locally. added this to Build.scala:

  resolvers += "nightlies" at
    "https://scala-ci.typesafe.com/artifactory/scala-release-temp/"
  scalaVersion := "2.12.1-371bc2c-nightly"
  scalaBinaryVersion := "2.12"  // or "2.11"

And then run:

$ ./sbt ++2.12.0 util-core/test
<snip>
[error] Failed: Total 1065, Failed 1, Errors 0, Passed 1064
[error] Failed tests:
[error] 	com.twitter.util.ConfigTest
[error] (util-core/test:test) sbt.TestsFailedException: Tests unsuccessful

@kevinoliver
Copy link
Contributor

@SethTisue i merged a workaround and hopefully that fixes it.

i didn't spend long digging into the difference in the generated code between 2.11, as this Config.missingValues code is deprecated and not well maintained. hoping we can remove it in the not too distant future.

finaglehelper pushed a commit that referenced this issue Dec 5, 2016
Problem

As reported in #174 new versions
of Scala 2.12 changed the behavior of what Config.missingValues sees.
Config is deprecated, long in the tooth, a bit finnicky, and generally
unsupported by us.

Solution

Make some patches to the tests to loosen the constraints and also
document that the behavior is not perfect in 2.12.

RB_ID=895199
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Development

No branches or pull requests

6 participants