Support Scala 2.13

[kerinin]

I tried monkey-patching scala_repositories:

scala_repositories(
  scala_version_shas = (
    "2.13.0",
    {
      "scala_compiler": "71d3be89400375bb014ec7eb0b367b91a8f7c7c2354c4f2d9efe28f1af8f960c",
      "scala_library": "bd3b2fa8b922295ccf1537aba1850d455f82003c10484df509d29ff177cc1edc",
      "scala_reflect": "aef72378abbc8b05cfdc7144fc0e55904dd87bf46ce2e973ae667dc3ed2fcee8"
    }
  ),
  scala_extra_jars = {
    "2.13": {
      "scalatest": {
        "version": "3.0.5",
        "sha256": "b416b5bcef6720da469a8d8a5726e457fc2d1cd5d316e1bc283aa75a2ae005e5",
      },
      "scalactic": {
        "version": "3.0.5",
        "sha256": "57e25b4fd969b1758fe042595112c874dfea99dca5cc48eebe07ac38772a0c41",
      },
      "scala_xml": {
        "version": "1.0.5",
        "sha256": "035015366f54f403d076d95f4529ce9eeaf544064dbc17c2d10e4f5908ef4256",
      },
      "scala_parser_combinators": {
        "version": "1.0.4",
        "sha256": "282c78d064d3e8f09b3663190d9494b85e0bb7d96b0da05994fe994384d96111",
      },
    },
  },
)

But it looks like there are incompatibilities in rules_scala itself:

$ bazel build ...                                                                                                                                                                                                                                                                                                                    Saturday August 10 08:38:25 AM
INFO: Analyzed 10 targets (2 packages loaded, 417 targets configured).
INFO: Found 10 targets...
ERROR: /private/var/tmp/_bazel_ryanmichael/17a3ffa16963a088f6b5ad97d780663a/external/io_bazel_rules_scala/src/java/io/bazel/rulesscala/scalac/BUILD:10:1: Building external/io_bazel_rules_scala/src/java/io/bazel/rulesscala/scalac/scalac.jar (4 source files) failed (Exit 1)
external/io_bazel_rules_scala/src/java/io/bazel/rulesscala/scalac/ScalacProcessor.java:242: error: cannot find symbol
      reporter.printSummary();
              ^
  symbol:   method printSummary()
  location: variable reporter of type ConsoleReporter
INFO: Elapsed time: 1.967s, Critical Path: 0.53s
INFO: 0 processes.
FAILED: Build did NOT complete successfully

[johnynek]

I imagine the call to nsc will need to change. Happy to help you work on a PR here. We can have a separate source file for 2.13 and change as needed, then in the toolchain we can point to a different executable.

[ittaiz]

I think we’re also using deprecated methods there with respect to the reporter. I tried handling it but it broke console output completely...

…

On Sat, 10 Aug 2019 at 20:57 P. Oscar Boykin ***@***.***> wrote: I imagine the call to nsc will need to change. Happy to help you work on a PR here. We can have a separate source file for 2.13 and change as needed, then in the toolchain we can point to a different executable. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#809>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAKQQF2EG5QQHBKQDFLVG5TQD36PRANCNFSM4IKZMKYQ> .

[k-bespalov]

any news..?

[bjchambers]

I spent some time tinkering with this, and figured I'd share the branch and what I did get working / not working:

1. I had to update test_version significantly since it was broken in multiple ways. It needed to load http_archive, skylib, stop using jvm_flags, and register the scala & scala_proto toolchains.
2. I had to update a bunch of dependencies since the current versions don't support 2.13. The most problematic was fastparser, which is used by scalapb, which doesn't have a version supporting all of 2.11, 2.12 and 2.13.
3. Once that was passed, there are a variety of flags and code in use that weren't compatible with 2.13. I used a MethodHandle in ScalacProcessor to avoid forking the code.

There are a variety of things left that don't work in 2.13 (when running test_version and commenting out 2.11 and 2.12), and I'm not sure all the changes work with the earlier versions. If anyone wants to take a look the branch is here:
https://github.com/bjchambers/rules_scala-1/tree/scala-2.13

I'm not sure whether I'll have any more time or not to look at it, but if I do I'll update that branch.

[mmlac]

What is the plan on Scala 2.13 (and then soon Scala 3.0) support?

Is the plan to only look at this after moving to toolchains and depend on #940 and #962 or is language support an issue to be addressed separately as the others mainly deal with the goal of supporting multiple Scala versions at the same time?

[smparkes]

I would not wait on #962. I'm contemplating closing that. It's sufficient but not necessary.

[ittaiz]

The plan is to have the ability to use a different vector of scala repositories to work with the repo.

Whoever needs this is welcome to contribute it (like Oscar mentioned).

We (Wix) will probably get to it in the next few quarters for our own needs.

One thing (which I’m not 100% sure where we should discuss this) is whether we should drop support for 2.11 to ease using both 2.12 and 2.13
I think that if someone works on this and finds big problems I’d be open towards dropping 2.11 support (given community agreement).

[Shastick]

Hi there, anything new on this issue?

Would anyone have some pointers on what needs to be done and where to start? (I'm relatively new to Bazel so it would be some very iterative trial and error at best, but I might be able to scrape together some time.)

[ittaiz]

Nothing new. You can take a look at the test which a different workspace which shows how 2.11 can be configured. That shows the entry points to users. From that you can track your way in. The main question is how many changes are required to our scalac invoked and such.

On Tue, 1 Sep 2020 at 13:39 Shastick ***@***.***> wrote: Hi there, anything new on this issue? Would anyone have some pointers on what needs to be done and where to start? (I'm relatively new to Bazel so it would be some very iterative trial and error at best, but I *might* be able to scrape together some time.) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#809 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKQQF73IM2GAVQOHKNT4XTSDTFPXANCNFSM4IKZMKYQ> . --

*Ittai Zeidman* 40 Hanamal street, Tel Aviv, Israel <http://www.wix.com>

[liucijus]

Support added via #1118