[DYOD] Add generic implementation of Parallel Inplace Merge Sort #2605
Conversation
|
Given this list (https://en.cppreference.com/w/cpp/compiler_support), GCC 10 should at least support |
|
In our intermediate review PR, we got it to get past the |
We want to use features that gcc 9 doesn't support. Commands to reproduce: ```sh sed -i 's/gcc-9/gcc-10/g' Jenkinsfile Dockerfile install_dependencies.sh sed -i 's/g++-9/g++-10/g' Jenkinsfile Dockerfile install_dependencies.sh sed -i 's/gcc9/gcc10/g' Jenkinsfile Dockerfile install_dependencies.sh sed -i 's/g++9/g++10/g' Jenkinsfile Dockerfile install_dependencies.sh ```
|
Okay, works with gcc-10, but somehow clang-11 gives the error The full file starts working with clang 13 and breaks for different reasons in clang 11 and 12 on godbolt: https://godbolt.org/z/qdvq61WWs |
Command to reproduce: ```bash sed -i -e 's/clang++-11/clang++-13/g' -e 's/clang-11/clang-13/g' -e 's/clang++11/clang++13/g' -e 's/clang11/clang13/g' Jenkinsfile Dockerfile install_dependencies.sh ```
836e9c4 to
c6e821c
Compare
These changes are extracted into a separate PR from our project work on the window function operator, as discussed with supervisors. The main PR will be opened later :)
I broke down the changes into three clean commits, so that you can keep them if you want, but feel free to squash anyways if you feel like it.
Since this uses
span(and other C++20 features) again, gcc-9 doesn't work. Should I also include the commit updating it here, or open another PR, or do something completely different?