arrow: [Benchmarks] All cpp microbenchmarks are failing with new GTest CMake Error
Describe the bug, including details regarding any error messages, version, and platform.
Since the PR https://github.com/apache/arrow/pull/34765 was merged, the cpp-micro
benchmarks are failing with the following error:
CMake Error at cmake_modules/ThirdpartyToolchain.cmake:2236 (message):
System GTest is built with a C++ standard lower than 17. Use bundled GTest
via passing in CMake flag
-DGTest_SOURCE="BUNDLED"
Here is an example Buildkite build.
This error message was explicitly added in that PR, so we’ll need to take action to correct it.
cc @HaochengLIU
Component(s)
Benchmarking, C++
About this issue
- Original URL
- State: closed
- Created a year ago
- Comments: 16 (16 by maintainers)
Commits related to this issue
- GH-34861: [C++] Move GTest std version check into SYSTEM if branch ### Rationale for this change Developers have reported that the current GTest std version check will break cpp microbenchmarks in BU... — committed to HaochengLIU/arrow by HaochengLIU a year ago
- GH-34861: [C++] Move GTest std version check into SYSTEM if branch (#34874) ### Rationale for this change Developers have reported that the current GTest std version check will break cpp microbenchma... — committed to apache/arrow by HaochengLIU a year ago
- GH-34861: [C++] Move GTest std version check into SYSTEM if branch (#34874) ### Rationale for this change Developers have reported that the current GTest std version check will break cpp microbenchma... — committed to Bit-Quill/arrow by HaochengLIU a year ago
- GH-34861: [C++] Move GTest std version check into SYSTEM if branch (#34874) ### Rationale for this change Developers have reported that the current GTest std version check will break cpp microbenchma... — committed to rtpsw/arrow by HaochengLIU a year ago
I did indeed mean to tag Jacob (@assignUser) who knows a bunch about our CI stuff (as well as has some experience with the benchmarking setup). But I totally see how that username looks funny there 😂
Thanks @kou ! I’ve confirmed that benchmarks are passing again.
I’ve implemented #34813. So this must be fixed too.
If this is still happen, we can reopen this.
haha, I was firstly shocked by how smart github is to map
assignUser
to issue assignee 😃😄 I’ll have a look later today!