Native applied types #3033

odersky · 2017-08-29T08:50:56Z

This is ready for review now.

dottybot

Hello, and thank you for opening this PR! 🎉

All contributors have signed the CLA, thank you! ❤️

Commit Messages

We want to keep history, but for that to actually be useful we have
some rules on how to format our commit messages (relevant xkcd).

Please stick to these guidelines for commit messages:

Separate subject from body with a blank line

When fixing an issue, start your commit message with Fix #<ISSUE-NBR>:

Limit the subject line to 72 characters

Capitalize the subject line

Do not end the subject line with a period

Use the imperative mood in the subject line ("Added" instead of "Add")

Wrap the body at 80 characters

Use the body to explain what and why vs. how

adapted from https://chris.beams.io/posts/git-commit

Have an awesome day! ☀️

smarter · 2017-08-29T11:59:51Z

Github seems to be having some issues with this PR and it's preventing the CI from checking it out. Could you try pushing to it again? If that doesn't work we may have to close this PR and open a new one.

smarter · 2017-08-29T17:52:59Z

Ah, I think that what's going on is that our CI is configured to use the github branch corresponding to this PR merged on top of master. Since merging currently fails with conflicts, this branch doesn't exist. So I'm afraid you cannot test this on the CI until you rebase.

allanrenucci · 2017-08-31T12:28:26Z

@odersky The last CI failure is not legitimate. Github picked the wrong CI (I am testing a new CI instance with a newer version of Drone). If you push a new commit, it should pick up the right CI.

smarter · 2017-08-31T14:04:59Z

repl/src/dotty/tools/repl/UserFacingPrinter.scala

@@ -21,6 +21,7 @@ import dotc.printing.{ GlobalPrec, DotPrec, Printer, PlainPrinter }
 import dotc.typer.Implicits.SearchResult
 import dotc.typer.ImportInfo

+// TODO: Avoid code duplication between userfacing and refined printers


Most of the duplication is gone in #3032 so you should be able to drop this commit when rebasing.

smarter · 2017-08-31T19:12:19Z

The dotty-bin-tests CI failure should go away if you rebase (things changed in .drone.yml)

smarter · 2017-09-02T23:29:30Z

The exhaustiveness checker is outputting lots of false warnings right now:

object Test {
  def test(i: Option[Int]) = i match {
    case Some(a) =>
    case None =>
  }
}

-- [E028] Pattern Match Exhaustivity Warning: try/pm.scala:2:29 ----------------
2 |  def test(i: Option[Int]) = i match {
  |                             ^
  |                             match may not be exhaustive.
  |                             
  |                             It would fail on: Some[Int](_)

AppliedType will be used for all type applications, higher-kinded or not. That is, eventually, it will replace HKApply and also the encodings of type applications as refined types. This commit introduces AppliedType and adapts baseType computations to take it into account. AppliedType is not yet constructed anywhere, however.

So far, everything up to parents is adapted to new scheme. Parents is half done; needs more work once we change layour of ClassInfo.

Simplifies usage and removes some obscure code in Types

Systematically introduce handlers for AppliedTypes where there was a handler for HKApply.

Also, change asSeenFrom to use it for arguments

Avoid printing addresses and reduce junk.

We had a stackoverfloa in asSeenFrom even under the old scheme, since there was a bad change involving checkRealizableBounds.

liufengyun · 2017-09-05T10:35:50Z

That's very surprising since these points do not look any different from the non-bootstrapped points. I would expect some difference in one direction or another. Are we completely sure the new points are bootstrapped? :)

It's definitely the bootstrapped version.

It probably means that the generated code from Dotty is comparable to Scalac 2.12, or the optimisations in Dotty is insignificant compared to that of JVM.

smarter · 2017-09-05T22:07:49Z

test performance please d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349

dottybot · 2017-09-05T22:07:53Z

performance test scheduled: 1 job(s) in queue.

smarter · 2017-09-05T22:08:40Z

@liufengyun Oops, I guess liufengyun/bench#47 is not active yet ^ ?

liufengyun · 2017-09-05T22:14:49Z

@smarter : a colon is required after please. I just removed the job from the queue. You can try issue the command again.

smarter · 2017-09-05T22:16:29Z

test performance please: d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349

smarter · 2017-09-05T22:18:29Z

@liufengyun Hmm I think there's a bug in your script that I didn't catch while reviewing it: 0 is allowed in git hashes, but the bot replaced 3034f59 by 334f59 😄

liufengyun · 2017-09-05T22:25:37Z

@smarter : Thanks, I just noticed, fixed, deployed and removed the job.

smarter · 2017-09-05T22:26:15Z

test performance please: d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349

dottybot · 2017-09-05T22:27:07Z

performance test scheduled for d92e3e6,fb82b9,fc83c24,334f59,1e6572d,724538,393d91f,9af7324,922b379,a7ec349: 1 job(s) in queue.

smarter · 2017-09-05T22:27:38Z

Hmm, 0 is still missing in the hashes.

liufengyun · 2017-09-05T22:28:45Z

The polling service just restarted.

smarter · 2017-09-05T22:29:57Z

test performance please: d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349

dottybot · 2017-09-05T22:30:44Z

performance test scheduled for d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349: 1 job(s) in queue.

liufengyun · 2017-09-05T22:33:30Z

The daily benchmark is running from 12:00pm. Depending on how many PRs we merged today, each PR or commit will take ~40min. If everything goes well, we should be able to get the result in the morning.

dottybot · 2017-09-06T03:11:33Z

Performance test finished successfully:

Visit http://dotty-bench.epfl.ch/3033 to see the changes.

Benchmarks is based on merge(s) with master

smarter · 2017-09-06T11:36:27Z

@liufengyun The results don't seem to contain the commits scheduled for benchmarking.

liufengyun · 2017-09-06T12:18:14Z

@smarter I'm investigating the problem, will keep you updated once it works.

liufengyun · 2017-09-06T15:31:36Z

test performance please: d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349

dottybot · 2017-09-06T15:31:48Z

performance test scheduled for d92e3e6 fb82b09 fc83c24 3034f59 1e6572d 7204538 393d91f 9af7324 922b379 a7ec349: 1 job(s) in queue.

liufengyun · 2017-09-06T19:52:11Z

test performance please: d92e3e6 fb82b09 fc83c24

dottybot · 2017-09-06T19:52:44Z

performance test scheduled for d92e3e6 fb82b09 fc83c24: 1 job(s) in queue.

dottybot · 2017-09-06T21:45:27Z

Performance test finished successfully:

Visit http://dotty-bench.epfl.ch/3033 to see the changes.

Benchmarks is based on merge(s) with master

smarter · 2017-09-06T21:52:24Z

Looks like fb82b09 was bad for performance, and fc83c24 made things better but still worse than before.

@liufengyun It would be nice if clicking on a point on the graph jumped to the commit diff on github instead of the corresponding PR, sorry for making another feature request :).

liufengyun · 2017-09-06T22:28:00Z

@smarter I'm happy to improve the bench infrastructure. Let's talk more about the feature tomorrow, it seems in some cases jump to PR is more friendly.

odersky · 2017-10-08T18:55:34Z

We already integrated the optimizations that are easily applicable, and the rest is complicated and does not seem to buy us much.

dottybot reviewed Aug 29, 2017

View reviewed changes

odersky force-pushed the try-applied-type branch 2 times, most recently from d2b91c1 to f001e4c Compare August 31, 2017 11:18

smarter reviewed Aug 31, 2017

View reviewed changes

odersky force-pushed the try-applied-type branch from 12d2a78 to 80f5018 Compare August 31, 2017 16:13

odersky force-pushed the try-applied-type branch 8 times, most recently from 1aa8181 to 6324124 Compare September 2, 2017 09:52

odersky added 12 commits September 4, 2017 18:35

Adapt operations in Types

5a34071

So far, everything up to parents is adapted to new scheme. Parents is half done; needs more work once we change layour of ClassInfo.

Move givenSelfType to ClassSymbol

2eeef2e

Simplifies usage and removes some obscure code in Types

Handle AppliedTypes

ab87816

Systematically introduce handlers for AppliedTypes where there was a handler for HKApply.

Introduce TypeArgRef

b779514

Also, change asSeenFrom to use it for arguments

Various fixes

87506ad

More fixes

102bfdc

Fix stray brace

c2c5bce

Fix typing of _* arguments

bdd72bd

Fix rebase breakage

7747ad1

Fix debug output to make it more stable

a273d37

Avoid printing addresses and reduce junk.

Fix Stackoverflow in asSeenFrom

1be4904

We had a stackoverfloa in asSeenFrom even under the old scheme, since there was a bad change involving checkRealizableBounds.

scala deleted a comment from dottybot Sep 5, 2017

smarter mentioned this pull request Sep 6, 2017

Type inference on type lambdas w.r.t tuples is confusing #3075

Closed

odersky mentioned this pull request Sep 30, 2017

Simplify substituters #3215

Merged

odersky closed this Oct 8, 2017

Native applied types #3033

Native applied types #3033

Uh oh!

Conversation

odersky commented Aug 29, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dottybot left a comment

Choose a reason for hiding this comment

Commit Messages

Uh oh!

smarter commented Aug 29, 2017

Uh oh!

smarter commented Aug 29, 2017

Uh oh!

allanrenucci commented Aug 31, 2017

Uh oh!

smarter Aug 31, 2017

Choose a reason for hiding this comment

Uh oh!

smarter commented Aug 31, 2017

Uh oh!

smarter commented Sep 2, 2017

Uh oh!

liufengyun commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

dottybot commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

liufengyun commented Sep 5, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

liufengyun commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

dottybot commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

liufengyun commented Sep 5, 2017

Uh oh!

smarter commented Sep 5, 2017

Uh oh!

dottybot commented Sep 5, 2017

Uh oh!

liufengyun commented Sep 5, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dottybot commented Sep 6, 2017

Uh oh!

smarter commented Sep 6, 2017

Uh oh!

liufengyun commented Sep 6, 2017

Uh oh!

liufengyun commented Sep 6, 2017

Uh oh!

dottybot commented Sep 6, 2017

Uh oh!

liufengyun commented Sep 6, 2017

Uh oh!

dottybot commented Sep 6, 2017

Uh oh!

dottybot commented Sep 6, 2017

Uh oh!

smarter commented Sep 6, 2017

Uh oh!

liufengyun commented Sep 6, 2017

Uh oh!

odersky commented Oct 8, 2017

Uh oh!

Uh oh!

odersky commented Aug 29, 2017 •

edited

Loading

liufengyun commented Sep 5, 2017 •

edited

Loading

smarter commented Sep 5, 2017 •

edited

Loading

liufengyun commented Sep 5, 2017 •

edited

Loading