Skip to content

Fix issue #50811 (NaN > NaN was true). - #50812

Merged
bors merged 1 commit into
rust-lang:masterfrom
kennytm:fix-50756-miri-bad-float-behavior
May 21, 2018
Merged

Fix issue #50811 (NaN > NaN was true).#50812
bors merged 1 commit into
rust-lang:masterfrom
kennytm:fix-50756-miri-bad-float-behavior

Conversation

@kennytm

@kennytmkennytm commented May 16, 2018•
edited
Loading

Copy link
Copy Markdown
Member

Fix#50811

Make sure the float comparison output is consistent with the expected behavior when NaN is involved.

Note: This PR is a BREAKING CHANGE. If you have used > or >= to compare floats, and make the result as the length of a fixed array type, like:

use std::f64::NAN;let x:[u8;(NAN > NAN)asusize] = [1];

then the code will no longer compile. Previously, all float comparison involving NaN will just return "Greater", i.e. NAN > NAN would wrongly return true during const evaluation. If you need to retain the old behavior (why), you may replace a > b with a != a || b != b || a > b.

@rust-highfive

rust-highfive commented May 16, 2018

Copy link
Copy Markdown
Contributor

r? @estebank

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

@rust-highfiverust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label May 16, 2018
@kennytm

kennytm commented May 16, 2018

Copy link
Copy Markdown
MemberAuthor

r? @oli-obk

Do you think we need to crater according to #50811 (comment)?

@kennytm
kennytmforce-pushed the fix-50756-miri-bad-float-behavior branch from 038a100 to cf3c554CompareMay 16, 2018 20:40
@rust-highfive

This comment has been minimized.

Sign in to view
@kennytm
kennytmforce-pushed the fix-50756-miri-bad-float-behavior branch from cf3c554 to 3a2bd0bCompareMay 16, 2018 21:02
@oli-obk

oli-obk commented May 16, 2018

Copy link
Copy Markdown
Contributor

Well... the clean way would be a future compat lint. And we definitely need some team to sign off on it.

That said: anyone depending on floating point corner cases for const eval things is somewhat asking for getting their code broken.

We could restore the original behaviour by making const eval decide on a way to do these edge cases depending on whether it is called in a true const environment or a promoted env, that would be horrible, but doable. Though I personally would like to see this PR merged as it is.

@kennytmkennytm added the relnotes Marks issues that should be documented in the release notes of the next release. label May 16, 2018
@Mark-SimulacrumMark-Simulacrum added the beta-nominated Nominated for backporting to the compiler in the beta channel. label May 17, 2018
@emilyalbiniemilyalbini mentioned this pull request May 17, 2018
Closed
@Mark-Simulacrum

Mark-Simulacrum commented May 17, 2018

Copy link
Copy Markdown
Member

cc @rust-lang/lang @rust-lang/compiler -- This PR is a candidate for stable backporting (1.26.1) as it fixes a bug in const eval; if you object, please note that here.

@kennytmkennytm added I-nominated T-lang Relevant to the language team T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 17, 2018
@kennytm
kennytmforce-pushed the fix-50756-miri-bad-float-behavior branch from 3a2bd0b to 50bc72dCompareMay 17, 2018 09:18
@nikomatsakisnikomatsakis mentioned this pull request May 17, 2018
Closed
@nikomatsakis

nikomatsakis commented May 17, 2018

Copy link
Copy Markdown
Contributor

Discussed in the @rust-lang/compiler meeting. General consensus was that NaN > Nan being true is definitely incorrect behavior and should be fixed. The only reason to hold up would be to warn users that may be affected, but we can't leave the semantics broken.

@emilyalbiniemilyalbini added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-team labels May 21, 2018
@emilyalbini

emilyalbini commented May 21, 2018

Copy link
Copy Markdown
Member

Ping from triage @oli-obk! How should this PR move forward?

@oli-obk

oli-obk commented May 21, 2018

Copy link
Copy Markdown
Contributor

@bors r+

@bors

bors commented May 21, 2018

Copy link
Copy Markdown
Collaborator

📌 Commit 50bc72d has been approved by oli-obk

@borsbors removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label May 21, 2018
@borsbors added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label May 21, 2018
@Mark-Simulacrum

Mark-Simulacrum commented May 21, 2018

Copy link
Copy Markdown
Member

@bors p=1

@bors

bors commented May 21, 2018

Copy link
Copy Markdown
Collaborator

⌛ Testing commit 50bc72d with merge cb20f68...

bors added a commit that referenced this pull request May 21, 2018
@bors
Auto merge of#50812- kennytm:fix-50756-miri-bad-float-behavior, r=o…
cb20f68
…li-obk
Fix issue #50811 (`NaN > NaN` was true).
Fix#50811
Make sure the float comparison output is consistent with the expected behavior when NaN is involved.
----
Note: This PR is a **BREAKING CHANGE**. If you have used `>` or `>=` to compare floats, and make the result as the length of a fixed array type, like:
```rust
use std::f64::NAN;
let x: [u8; (NAN > NAN) as usize] = [1];
```
then the code will no longer compile. Previously, all float comparison involving NaN will just return "Greater", i.e. `NAN > NAN` would wrongly return `true` during const evaluation. If you need to retain the old behavior (why), you may replace `a > b` with `a != a || b != b || a > b`.
@bors

bors commented May 21, 2018

Copy link
Copy Markdown
Collaborator

☀️ Test successful - status-appveyor, status-travis
Approved by: oli-obk
Pushing cb20f68 to master...

@bors
bors merged commit 50bc72d into rust-lang:masterMay 21, 2018
@kennytm
kennytm deleted the fix-50756-miri-bad-float-behavior branch May 22, 2018 05:46
@nikomatsakisnikomatsakis added the beta-accepted Accepted for backporting to the compiler in the beta channel. label May 24, 2018
@emilyalbiniemilyalbini mentioned this pull request May 24, 2018
Merged
@Mark-SimulacrumMark-Simulacrum mentioned this pull request May 24, 2018
Merged
bors added a commit that referenced this pull request May 25, 2018
@bors
Auto merge of#51045- Mark-Simulacrum:stable-point, r=alexcrichton
827013a
Stable point release (1.26.1)
This includes all items on [the wishlist](#50756), plus #50694, which backported cleanly.
The target date is May 29th, Tuesday next week.
cc @rust-lang/compiler@oli-obk@eddyb -- I backported #50812, but it wasn't a clean patch so review would be appreciated.
bors added a commit that referenced this pull request Jun 1, 2018
@bors
Auto merge of#51037- pietroalbini:beta-backports, r=oli-obk
13abb91
[beta] Process backports
Merged and approved:
* #50812: Fix issue #50811 (`NaN > NaN` was true).
* #50827: Update LLVM to `56c931901cfb85cd6f7ed44c7d7520a8de1edf97`
* #50879: Fix naming conventions for new lints
* #51011: rustdoc: hide macro export statements from docs
* #51051: prohibit turbofish in `impl Trait` methods
* #51052: restore emplacement syntax (obsolete)
* #51146: typeck: Do not pass the field check on field error
* #51235: remove notion of Implicit derefs from mem-cat
r? @ghost
@emilyalbiniemilyalbini removed the beta-nominated Nominated for backporting to the compiler in the beta channel. label Jun 2, 2018
bors added a commit that referenced this pull request Jun 3, 2018
@bors
Auto merge of#51037- pietroalbini:beta-backports, r=pietroalbini
eb01d0f
[beta] Process backports
Merged and approved:
* #50812: Fix issue #50811 (`NaN > NaN` was true).
* #50879: Fix naming conventions for new lints
* #51011: rustdoc: hide macro export statements from docs
* #51051: prohibit turbofish in impl Trait methods
* #51052: restore emplacement syntax (obsolete)
* #51146: typeck: Do not pass the field check on field error
* #51235: remove notion of Implicit derefs from mem-cat
r? @ghost
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

@oli-obkoli-obk

Labels

beta-acceptedAccepted for backporting to the compiler in the beta channel.relnotesMarks issues that should be documented in the release notes of the next release.S-waiting-on-borsStatus: Waiting on bors to run and complete tests. Bors will change the label on completion.T-compilerRelevant to the compiler team, which will review and decide on the PR/issue.T-langRelevant to the language team

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

NaN > NaN returns true

8 participants

@kennytm@rust-highfive@oli-obk@Mark-Simulacrum@nikomatsakis@emilyalbini@bors@estebank