From 7c9a4658bcb39329a80279e8456bbac58ec9e94f Mon Sep 17 00:00:00 2001
From: Shuhei Kadowaki <40514306+aviatesk@users.noreply.github.com>
Date: Wed, 17 Jul 2024 17:39:27 +0900
Subject: [PATCH] =?UTF-8?q?LinearAlgebra:=20use=20`=E2=89=88`=20instead=20?=
 =?UTF-8?q?of=20`=3D=3D`=20for=20`tr`=20tests=20in=20symmetric.jl=20(#5514?=
 =?UTF-8?q?3)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

After investigating JuliaLang/julia#54090, I found that the issue was
not caused by the effects of `checksquare`, but by the use of the
`@simd` macro within `tr(::Matrix)`:

https://github.com/JuliaLang/julia/blob/0945b9d7740855c82a09fed42fbf6bc561e02c77/stdlib/LinearAlgebra/src/dense.jl#L373-L380

While simply removing the `@simd` macro was considered, the strict
left-to-right summation without `@simd` otherwise is not necessarily
more accurate, so I concluded that the problem lies in the test code,
which tests the (strict) equality of two different `tr` execution
results. I have modified the test code to use `≈` instead of `==`.

- fixes #54090
---
 stdlib/LinearAlgebra/test/symmetric.jl | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/stdlib/LinearAlgebra/test/symmetric.jl b/stdlib/LinearAlgebra/test/symmetric.jl
index 414b4ad5dd722..89e9ca0d6a51d 100644
--- a/stdlib/LinearAlgebra/test/symmetric.jl
+++ b/stdlib/LinearAlgebra/test/symmetric.jl
@@ -223,8 +223,8 @@ end
 
         @testset "linalg unary ops" begin
             @testset "tr" begin
-                @test tr(asym) == tr(Symmetric(asym))
-                @test tr(aherm) == tr(Hermitian(aherm))
+                @test tr(asym) ≈ tr(Symmetric(asym))
+                @test tr(aherm) ≈ tr(Hermitian(aherm))
             end
 
             @testset "isposdef[!]" begin