Use len instead of size_hint where appropiate

[tbu-]

This makes it clearer that we're not just looking for a lower bound but
rather know that the iterator is an ExactSizeIterator.

[rust-highfive]

r? @alexcrichton

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

[nagisa]

Doesn’t len have an assertion which size_hint doesn’t? I guess I’m being too paranoid, but there certainly must be some extra overhead which warrants some benchmarks… e.g. how does the performance of iterating though CharIndices change?

[tbu-]

If that were the case, we should rather modify len than use size_hint in places where len would be more appropiate.

But it looks like LLVM can see right through it:

use std::slice;

pub fn slice_len(iter: &slice::Iter<u8>) -> usize {
    iter.len()
}

pub fn slice_len2(iter: &slice::Iter<u8>) -> usize {
    iter.size_hint().0
}

produces the same bytecode

; ModuleID = 'a.0.rs'
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

%"2.std::slice::Iter<u8>" = type { i8*, i8*, %"2.std::marker::PhantomData<&'static u8>" }
%"2.std::marker::PhantomData<&'static u8>" = type {}

; Function Attrs: norecurse readonly uwtable
define i64 @_ZN1a9slice_len17hc2890fcf572cc394E(%"2.std::slice::Iter<u8>"* noalias nocapture readonly dereferenceable(16)) unnamed_addr #0 {
entry-block:
  %1 = getelementptr inbounds %"2.std::slice::Iter<u8>", %"2.std::slice::Iter<u8>"* %0, i64 0, i32 1
  %2 = bitcast i8** %1 to i64*
  %3 = load i64, i64* %2, align 8, !alias.scope !0, !noalias !5
  %4 = bitcast %"2.std::slice::Iter<u8>"* %0 to i64*
  %5 = load i64, i64* %4, align 8, !alias.scope !0, !noalias !5
  %6 = sub i64 %3, %5
  ret i64 %6
}

; Function Attrs: norecurse nounwind readonly uwtable
define i64 @_ZN1a10slice_len217h415029022c569338E(%"2.std::slice::Iter<u8>"* noalias nocapture readonly dereferenceable(16)) unnamed_addr #1 {
entry-block:
  %1 = getelementptr inbounds %"2.std::slice::Iter<u8>", %"2.std::slice::Iter<u8>"* %0, i64 0, i32 1
  %2 = bitcast i8** %1 to i64*
  %3 = load i64, i64* %2, align 8, !alias.scope !7, !noalias !10
  %4 = bitcast %"2.std::slice::Iter<u8>"* %0 to i64*
  %5 = load i64, i64* %4, align 8, !alias.scope !7, !noalias !10
  %6 = sub i64 %3, %5
  ret i64 %6
}

attributes #0 = { norecurse readonly uwtable }
attributes #1 = { norecurse nounwind readonly uwtable }

!0 = !{!1, !3}
!1 = distinct !{!1, !2, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE: argument 1"}
!2 = distinct !{!2, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE"}
!3 = distinct !{!3, !4, !"_ZN4core4iter6traits17ExactSizeIterator3len17hb31ac8a36a76a083E: argument 0"}
!4 = distinct !{!4, !"_ZN4core4iter6traits17ExactSizeIterator3len17hb31ac8a36a76a083E"}
!5 = !{!6}
!6 = distinct !{!6, !2, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE: argument 0"}
!7 = !{!8}
!8 = distinct !{!8, !9, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE: argument 1"}
!9 = distinct !{!9, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE"}
!10 = !{!11}
!11 = distinct !{!11, !9, !"_ZN79_$LT$std..slice..Iter$LT$$u27$a$C$$u20$T$GT$$u20$as$u20$std..iter..Iterator$GT$9size_hint17h39fecfdf3007b42aE: argument 0"}

[alexcrichton]

Yes most of these look like it's trivial for LLVM to see through as the size_hint method is just returning (a, Some(a)) which is easy to cause the assertion that (Some(low) == hi) to get optimized away. Looks fine to me.

@bors: r+ 8ff5c43

[bors]

⌛ Testing commit 8ff5c43 with merge ad07411...

[bors]

💔 Test failed - auto-mac-64-opt

[alexcrichton]

@bors: retry

On Thu, Jun 23, 2016 at 4:00 PM, bors [email protected] wrote:

💔 Test failed - auto-mac-64-opt
http://buildbot.rust-lang.org/builders/auto-mac-64-opt/builds/9566

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#34425 (comment), or mute
the thread
https://github.com/notifications/unsubscribe/AAD95LMBFxGwiGVYuRHs0pT2m2nXRbgNks5qOxACgaJpZM4I8qoV
.

[bors]

⌛ Testing commit 8ff5c43 with merge 90943a5...

[alexcrichton]

@bors: retry force clean

• restarted buildbot

[bors]

⌛ Testing commit 8ff5c43 with merge 4b89deb...

[bors]

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-cargotest, auto-linux-64-cross-freebsd, auto-linux-64-debug-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-opt-mir, auto-linux-64-opt-rustbuild, auto-linux-64-x-android-t, auto-linux-cross-opt, auto-linux-musl-64-opt, auto-mac-32-opt, auto-mac-64-opt, auto-mac-64-opt-rustbuild, auto-mac-cross-ios-opt, auto-win-gnu-32-opt, auto-win-gnu-32-opt-rustbuild, auto-win-gnu-64-opt, auto-win-msvc-32-opt, auto-win-msvc-64-cargotest, auto-win-msvc-64-opt, auto-win-msvc-64-opt-mir, auto-win-msvc-64-opt-rustbuild