Uh oh!
There was an error while loading. Please reload this page.
Remove pointer comparison from slice equality - #80209
Closed
erikdesjardins wants to merge 3 commits into
Closed
Conversation
ContributorAuthor
re #71735 (comment): I'm not sure how to test |
Mark-Simulacrum
commented
Dec 20, 2020
Member
Can we use an older version of hyper perhaps? |
ContributorAuthor
Hmm, even circa May 2016, Using
So basically what you'd expect, reverting this makes string comparison marginally slower, but comparing the enum directly is ~the same. Although to be honest I'm not really happy with either my benchmark or what #33892 (comment) describes. Yes, they involve other code, but the only thing that changes is one critical line--basically a microbenchmark with extra steps. But I'm not sure what a good, noncontrived benchmark for this would be. |
Mark-Simulacrum
commented
Dec 23, 2020
Member
I think the codegen test here should probably be changed to test for something we expect to be present, rather than the lack of pointer equality - I am not sure what the IR with this PR looks like, but maybe we can assert that there is a single memcmp generated or something along those lines? I am fine with not trying to test hyper further; I still believe that the check here for pointer equality is unlikely to be impactful in practice (and we know it actively hurts some cases). @bors rollup=never though. |
erikdesjardins
commented
Dec 23, 2020
ContributorAuthor
Changed the test, let me know if you have a better pattern. The old IR is: definezeroexti1@is_zero_slice([4 x i8]* noaliasreadonlyalign1dereferenceable(4) %data) unnamed_addr #0 {
start:
%0 = icmpeq [4 x i8]* %data, getelementptrinbounds (<{ [4 x i8] }>, <{ [4 x i8] }>* @alloc2, i640, i320)
bri1%0, label %"_ZN4core5array103_$LT$impl$u20$core..cmp..PartialEq$LT$$u5b$B$u3b$$u20$N$u5d$$GT$$u20$for$u20$$u5b$A$u3b$$u20$N$u5d$$GT$2eq17h702fd33f1073d678E.exit", label%bb9.i.i.i
bb9.i.i.i: ; preds = %start%1 = getelementptr [4 x i8], [4 x i8]* %data, i640, i640%bcmp.i.i.i = tailcalli32@bcmp(i8*nonnulldereferenceable(4) %1, i8*nonnulldereferenceable(4) getelementptrinbounds (<{ [4 x i8] }>, <{ [4 x i8] }>* @alloc2, i640, i320, i640), i644) #2%2 = icmpeqi32%bcmp.i.i.i, 0brlabel %"_ZN4core5array103_$LT$impl$u20$core..cmp..PartialEq$LT$$u5b$B$u3b$$u20$N$u5d$$GT$$u20$for$u20$$u5b$A$u3b$$u20$N$u5d$$GT$2eq17h702fd33f1073d678E.exit""_ZN4core5array103_$LT$impl$u20$core..cmp..PartialEq$LT$$u5b$B$u3b$$u20$N$u5d$$GT$$u20$for$u20$$u5b$A$u3b$$u20$N$u5d$$GT$2eq17h702fd33f1073d678E.exit": ; preds = %start, %bb9.i.i.i%.0.i.i.i = phii1 [ %2, %bb9.i.i.i ], [ true, %start ]
reti1%.0.i.i.i
}and the new IR is: definezeroexti1@is_zero_slice([4 x i8]* noaliasnocapturereadonlyalign1dereferenceable(4) %data) unnamed_addr #0 {
start:
%0 = getelementptr [4 x i8], [4 x i8]* %data, i640, i640%bcmp.i.i.i = tailcalli32@bcmp(i8*nonnulldereferenceable(4) %0, i8*nonnulldereferenceable(4) getelementptrinbounds (<{ [4 x i8] }>, <{ [4 x i8] }>* @alloc2, i640, i320, i640), i644) #2%1 = icmpeqi32%bcmp.i.i.i, 0reti1%1
} |
Mark-Simulacrum
commented
Dec 23, 2020
Member
Yeah, I think the test is good. We can relax it if it turns out to be too sensitive. @bors r+ rollup=never squash |
bors
commented
Dec 23, 2020
Collaborator
📌 Commit f29b6b8 has been approved by |
bors
commented
Dec 24, 2020
Collaborator
⌛ Testing commit f29b6b8 with merge 280255c7b4b67d559361bdd77d0f1f9c22b073b2... |
rust-log-analyzer
commented
Dec 24, 2020
Collaborator
The job Click to see the possible cause of the failure (guessed by this bot) |
bors
commented
Dec 24, 2020
Collaborator
💔 Test failed - checks-actions |
Mark-Simulacrum
commented
Dec 24, 2020
Member
@bors r+ squash rollup=never |
bors
commented
Dec 24, 2020
Collaborator
📌 Commit fae3ede has been approved by |
bors
commented
Dec 24, 2020
Collaborator
⌛ Testing commit fae3ede with merge 93bac3bc6043448eb0b1de155765fbe6318656ec... |
rust-log-analyzer
commented
Dec 24, 2020
Collaborator
The job Click to see the possible cause of the failure (guessed by this bot) |
bors
commented
Dec 24, 2020
Collaborator
💔 Test failed - checks-actions |
Mark-Simulacrum
commented
Dec 24, 2020
Member
@bors retry |
bors
commented
Dec 25, 2020
Collaborator
⌛ Testing commit fae3ede with merge 80cb6a64ae252b9aa8682fecabd621506106a115... |
rust-log-analyzer
commented
Dec 25, 2020
Collaborator
The job Click to see the possible cause of the failure (guessed by this bot) |
bors
commented
Dec 25, 2020
Collaborator
💔 Test failed - checks-actions |
erikdesjardins
commented
Dec 26, 2020
ContributorAuthor
looks like another spurious failure |
Mark-Simulacrum
commented
Dec 26, 2020
Member
@bors retry |
bors added a commit
to rust-lang-ci/rust
that referenced
this pull request
Dec 26, 2020
Remove pointer comparison from slice equality This resurrects rust-lang#71735. Fixesrust-lang#71602, helps with rust-lang#80140. r? `@Mark-Simulacrum`
bors
commented
Dec 26, 2020
Collaborator
bors
commented
Dec 26, 2020
Collaborator
☀️ Test successful - checks-actions |
tesuji
commented
Dec 27, 2020
Contributor
Why was the PR not auto-closed? Is this a github bug or bors bug? |
ContributorAuthor
It also removes authorship (733cb54), which is unusual; I would expect the resulting commit from this rebase command to have the same author as the first commit from the PR. |
Mark-Simulacrum
commented
Dec 29, 2020
Member
Yeah, those both seem like bors bugs. I will try to investigate later today. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This resurrects #71735.
Fixes#71602, helps with #80140.
r? @Mark-Simulacrum