Disable SysinfoTest.NominalCPUFrequency in absl_sysinfo_test
This fails occasionally on aarch64, and upstream reports that it is not meaningful except for Google internal users. See: NominalCPUFrequency Test from SysInfoTest Suite Fails on M1 Mac https://github.com/abseil/abseil-cpp/issues/1053#issuecomment-961432444
This commit is contained in:
parent
0da16bcfd8
commit
4b8d306dee
44
abseil-cpp-20211102.0-disable-nominalcpufrequency.patch
Normal file
44
abseil-cpp-20211102.0-disable-nominalcpufrequency.patch
Normal file
@ -0,0 +1,44 @@
|
||||
From d984ddf1a64ce9fae36395b423cf6e52afc9a07a Mon Sep 17 00:00:00 2001
|
||||
From: "Benjamin A. Beasley" <code@musicinmybrain.net>
|
||||
Date: Wed, 16 Mar 2022 14:33:41 -0400
|
||||
Subject: [PATCH] Disable SysinfoTest.NominalCPUFrequency
|
||||
|
||||
SysinfoTest.NominalCPUFrequency in absl_sysinfo_test fails occasionally
|
||||
on aarch64, but see:
|
||||
|
||||
NominalCPUFrequency Test from SysInfoTest Suite Fails on M1 Mac
|
||||
https://github.com/abseil/abseil-cpp/issues/1053#issuecomment-961432444
|
||||
|
||||
in which an upstream author opines:
|
||||
|
||||
If the only problem you are trying to solve is a failing test, this is safe
|
||||
to ignore since this code is never called. I should consider stripping this
|
||||
test out of the open source release. NominalCPUFrequency is only called in
|
||||
code private to Google and we do have tests on the platforms we use it on.
|
||||
|
||||
We therefore disable it on all architectures, since any future failures
|
||||
will also not be meaningful.
|
||||
|
||||
Note also that this test is removed upstream in commit
|
||||
732b5580f089101ce4b8cdff55bb6461c59a6720 (internal commit
|
||||
7e8da4f14afd25d11713eee6b743ba31605332bf).
|
||||
---
|
||||
absl/base/internal/sysinfo_test.cc | 2 +-
|
||||
1 file changed, 1 insertion(+), 1 deletion(-)
|
||||
|
||||
diff --git a/absl/base/internal/sysinfo_test.cc b/absl/base/internal/sysinfo_test.cc
|
||||
index 5f9e45f..1a944f0 100644
|
||||
--- a/absl/base/internal/sysinfo_test.cc
|
||||
+++ b/absl/base/internal/sysinfo_test.cc
|
||||
@@ -44,7 +44,7 @@ TEST(SysinfoTest, NumCPUs) {
|
||||
// frequency, while others do not. Since we can't predict a priori what a given
|
||||
// machine is going to do, just disable this test on POWER on Linux.
|
||||
#if !(defined(__linux) && (defined(__ppc64__) || defined(__PPC64__)))
|
||||
-TEST(SysinfoTest, NominalCPUFrequency) {
|
||||
+TEST(SysinfoTest, DISABLED_NominalCPUFrequency) {
|
||||
// Linux only exposes the CPU frequency on certain architectures, and
|
||||
// Emscripten doesn't expose it at all.
|
||||
#if defined(__linux__) && \
|
||||
--
|
||||
2.35.1
|
||||
|
@ -14,6 +14,26 @@ Source0: https://github.com/abseil/abseil-cpp/archive/%{version}/%{name}-
|
||||
# released version of GTest. Not submitted upstream, as this is a workaround
|
||||
# rather than a fix. https://github.com/abseil/abseil-cpp/issues/1063
|
||||
Patch0: abseil-cpp-20211102.0-gtest-unreleased-features.patch
|
||||
# SysinfoTest.NominalCPUFrequency in absl_sysinfo_test fails occasionally
|
||||
# on aarch64, but see:
|
||||
#
|
||||
# NominalCPUFrequency Test from SysInfoTest Suite Fails on M1 Mac
|
||||
# https://github.com/abseil/abseil-cpp/issues/1053#issuecomment-961432444
|
||||
#
|
||||
# in which an upstream author opines:
|
||||
#
|
||||
# If the only problem you are trying to solve is a failing test, this is safe
|
||||
# to ignore since this code is never called. I should consider stripping this
|
||||
# test out of the open source release. NominalCPUFrequency is only called in
|
||||
# code private to Google and we do have tests on the platforms we use it on.
|
||||
#
|
||||
# We therefore disable it on all architectures, since any future failures
|
||||
# will also not be meaningful.
|
||||
#
|
||||
# Note also that this test is removed upstream in commit
|
||||
# 732b5580f089101ce4b8cdff55bb6461c59a6720 (internal commit
|
||||
# 7e8da4f14afd25d11713eee6b743ba31605332bf).
|
||||
Patch1: abseil-cpp-20211102.0-disable-nominalcpufrequency.patch
|
||||
|
||||
BuildRequires: cmake
|
||||
# The default make backend would work just as well; ninja is observably faster
|
||||
@ -98,6 +118,8 @@ find . -type f -name '*.cc' \
|
||||
%changelog
|
||||
* Tue Mar 15 2022 Benjamin A. Beasley <code@musicinmybrain.net> - 20211102.0-2
|
||||
- Disable LTO on s390x to work around test failure
|
||||
- Skip SysinfoTest.NominalCPUFrequency on all architectures; it fails
|
||||
occasionally on aarch64, and upstream says we should not care
|
||||
|
||||
* Fri Feb 18 2022 Benjamin A. Beasley <code@musicinmybrain.net> - 20211102.0-1
|
||||
- Update to 20211102.0 (close RHBZ#2019691)
|
||||
|
Loading…
Reference in New Issue
Block a user