traits with self-containing supertraits are not object safe #38603

arielb1 · 2016-12-25T17:57:14Z

This should be the last time I fix this function.

rust-highfive · 2016-12-25T17:57:18Z

(rust_highfive has picked a reviewer for you, use r? to override)

arielb1 · 2016-12-25T17:58:23Z

This fixes the ICE, but we still permit stuff such as

trait Q<T:?Sized> {}
trait Foo where u32: Q<Self> {}

fn main() {
    let f: Box<Foo> = ...;
}

That shouldn't cause an ICE because non-supertrait bounds are never used, but I still think I should future-compatibility-warning this.

eddyb · 2016-12-25T20:36:31Z

cc @rust-lang/compiler @brson We should probably gate this on Crater.

arielb1 · 2016-12-25T23:39:01Z

The current patch is pretty non-aggressive, prohibiting only cases that tend to ICE. I think we want to crater a more aggressive version of it (cc #38604).

arielb1 · 2016-12-31T00:42:07Z

now ready for crater

nikomatsakis · 2017-01-02T20:17:02Z

cc @brson -- can you run this through craterbomb?

brson · 2017-01-03T21:45:07Z

I'll run it through crater. Cargobomb isn't quite ready for this yet.

brson · 2017-01-04T22:50:32Z

Still cratering.

brson · 2017-01-06T01:10:39Z

Crater says https://gist.github.com/anonymous/3409c01561daa0670445c1ac9f967f6b

Some false positives but some legit errors.

nikomatsakis · 2017-01-09T22:58:54Z

arielb1 · 2017-01-10T10:08:58Z

I am not sure what's behind the rmt_serialize errors, but resolution finishes before my new code runs so it shouldn't be me.

All other errors look spurious.

nikomatsakis · 2017-01-10T21:58:31Z

@arielb1

I am not sure what's behind the rmt_serialize errors, but resolution finishes before my new code runs so it shouldn't be me.

That is odd. Do you have a local build? Can you give this a spin locally?

This should be the last time I fix this function. Fixes rust-lang#38404.

Fixes rust-lang#38604 needs a crater run

arielb1 · 2017-01-10T22:44:29Z

Looks like not our bug.

Compiling using the Cargo.lock in the repository works and uses the following versions:

   Compiling byteorder v0.3.13
   Compiling libc v0.2.8
   Compiling rustc-serialize v0.3.18
   Compiling rmp v0.7.3
   Compiling log v0.3.5
   Compiling rmp-serialize v0.7.0
   Compiling neovim-lib v0.1.1 (file:///home/ariel/Rust/neovim-lib)

While after a cargo update, or on crater, the following versions are used, triggering the error with both nightly and modified rustc:

   Compiling byteorder v1.0.0
   Compiling byteorder v0.4.2
   Compiling num-traits v0.1.36
   Compiling log v0.3.6
   Compiling rustc-serialize v0.3.22
   Compiling rmp v0.7.5
   Compiling rmp v0.8.1
   Compiling rmp-serialize v0.7.0

nikomatsakis · 2017-01-11T17:21:54Z

Looks like not our bug.

So you get the same results both with and without this PR?

Since this is being used as a library, the Cargo.lock that is found in the repo isn't meaningful anyway. Cargo.lock is only respected for the final target, iiuc.

arielb1 · 2017-01-14T15:29:51Z

@nikomatsakis

Exactly. When you use the Cargo.lock from source control, it works with both the old and new rustcs, and when you cargo update, it does not work.

nikomatsakis · 2017-01-17T22:27:02Z

@arielb1 I'm sort of confused by what you meant when you wrote "not our bug"

arielb1 · 2017-01-18T06:24:32Z

@nikomatsakis

That's it's broken, but not by our changes.

nikomatsakis · 2017-01-19T08:17:14Z

OK, I misread this comment at first:

When you use the Cargo.lock from source control, it works with both the old and new rustcs, and when you cargo update, it does not work.

Interesting.

nikomatsakis · 2017-01-19T08:17:52Z

@bors r+

bors · 2017-01-19T08:17:53Z

📌 Commit c85fb19 has been approved by nikomatsakis

bors · 2017-01-19T23:49:03Z

⌛ Testing commit c85fb19 with merge 7da72d5...

bors · 2017-01-20T01:18:53Z

💔 Test failed - status-appveyor

alexcrichton · 2017-01-20T01:19:41Z

@bors: retry * canceled to prioritize rollup

…

On Thu, Jan 19, 2017 at 5:18 PM, bors ***@***.***> wrote: 💔 Test failed - status-appveyor <https://ci.appveyor.com/project/rust-lang/rust/build/1.0.1613> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#38603 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAD95DENPdPNi8xHTyEm6EcO3u42gkknks5rUAt_gaJpZM4LVciT> .

bors · 2017-01-20T01:41:29Z

⌛ Testing commit c85fb19 with merge 8fa0690...

bors · 2017-01-20T01:48:30Z

💔 Test failed - status-appveyor

alexcrichton · 2017-01-20T01:53:30Z

@bors: retry * network error

…

On Thu, Jan 19, 2017 at 5:48 PM, bors ***@***.***> wrote: 💔 Test failed - status-appveyor <https://ci.appveyor.com/project/rust-lang/rust/build/1.0.1615> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#38603 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAD95IYzwrk6ODaKeiwNaXS52qFQMT1Uks5rUBJwgaJpZM4LVciT> .

bors · 2017-01-20T02:35:13Z

⌛ Testing commit c85fb19 with merge 6be85b4...

bors · 2017-01-20T02:56:23Z

💔 Test failed - status-appveyor

alexcrichton · 2017-01-20T02:57:22Z

@bors: retry * canceled to prioritize rollup

…

On Thu, Jan 19, 2017 at 6:56 PM, bors ***@***.***> wrote: 💔 Test failed - status-appveyor <https://ci.appveyor.com/project/rust-lang/rust/build/1.0.1618> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#38603 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAD95JTCyLL3M3nF0MNYm7IdZTqSQTsNks5rUCJZgaJpZM4LVciT> .

bors · 2017-01-20T03:48:26Z

⌛ Testing commit c85fb19 with merge d38430b...

alexcrichton · 2017-01-20T03:49:24Z

@bors: retry

prioritizing rollup

bors · 2017-01-20T03:49:32Z

⌛ Testing commit c85fb19 with merge 5a6ac84...

alexcrichton · 2017-01-20T03:53:14Z

@bors: retry

bors · 2017-01-20T04:54:27Z

⌛ Testing commit c85fb19 with merge b53d374...

traits with self-containing supertraits are not object safe This should be the last time I fix this function. Fixes #38404.

bors · 2017-01-20T07:17:27Z

☀️ Test successful - status-appveyor, status-travis
Approved by: nikomatsakis
Pushing b53d374 to master...

Rollup of 28 pull requests - Successful merges: #38603, #38761, #38842, #38847, #38955, #38966, #39062, #39068, #39077, #39111, #39112, #39114, #39118, #39120, #39132, #39135, #39138, #39142, #39143, #39146, #39157, #39166, #39167, #39168, #39179, #39184, #39195, #39197 - Failed merges: #39060, #39145

rust-highfive assigned eddyb Dec 25, 2016

arielb1 force-pushed the supertrait-self-3 branch from 992789a to 5a9fa5d Compare December 25, 2016 18:11

arielb1 added I-needs-decision Issue: In need of a decision. and removed I-needs-decision Issue: In need of a decision. labels Dec 25, 2016

eddyb added the S-waiting-on-crater Status: Waiting on a crater run to be completed. label Dec 25, 2016

Ariel Ben-Yehuda and others added 2 commits January 11, 2017 00:02

traits with self-containing supertraits are not object safe

4f542f9

This should be the last time I fix this function. Fixes rust-lang#38404.

forbid all self-referencing predicates

c85fb19

Fixes rust-lang#38604 needs a crater run

arielb1 force-pushed the supertrait-self-3 branch from ec742c6 to c85fb19 Compare January 10, 2017 22:02

bors added a commit that referenced this pull request Jan 20, 2017

Auto merge of #38603 - arielb1:supertrait-self-3, r=nikomatsakis

b53d374

traits with self-containing supertraits are not object safe This should be the last time I fix this function. Fixes #38404.

bors merged commit c85fb19 into rust-lang:master Jan 20, 2017

traits with self-containing supertraits are not object safe #38603

traits with self-containing supertraits are not object safe #38603

Conversation

arielb1 commented Dec 25, 2016

rust-highfive commented Dec 25, 2016

arielb1 commented Dec 25, 2016 • edited Loading

eddyb commented Dec 25, 2016

arielb1 commented Dec 25, 2016

arielb1 commented Dec 31, 2016

nikomatsakis commented Jan 2, 2017

brson commented Jan 3, 2017

brson commented Jan 4, 2017

brson commented Jan 6, 2017

nikomatsakis commented Jan 9, 2017 • edited by arielb1 Loading

arielb1 commented Jan 10, 2017

nikomatsakis commented Jan 10, 2017

arielb1 commented Jan 10, 2017

nikomatsakis commented Jan 11, 2017

arielb1 commented Jan 14, 2017

nikomatsakis commented Jan 17, 2017

arielb1 commented Jan 18, 2017

nikomatsakis commented Jan 19, 2017

nikomatsakis commented Jan 19, 2017

bors commented Jan 19, 2017

bors commented Jan 19, 2017

bors commented Jan 20, 2017

alexcrichton commented Jan 20, 2017 via email

bors commented Jan 20, 2017

bors commented Jan 20, 2017

alexcrichton commented Jan 20, 2017 via email

bors commented Jan 20, 2017

bors commented Jan 20, 2017

alexcrichton commented Jan 20, 2017 via email

bors commented Jan 20, 2017

alexcrichton commented Jan 20, 2017

bors commented Jan 20, 2017

alexcrichton commented Jan 20, 2017

bors commented Jan 20, 2017

bors commented Jan 20, 2017

arielb1 commented Dec 25, 2016 •

edited

Loading

nikomatsakis commented Jan 9, 2017 •

edited by arielb1

Loading