libprotobuf v4.23.0

[h-vetinari]

Based on #155, will be rebased once that's in.

Closes #157

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[h-vetinari]

So this fails with:

[342/391] Building CXX object CMakeFiles/tests.dir/src/google/protobuf/unittest.pb.cc.o
FAILED: CMakeFiles/tests.dir/src/google/protobuf/unittest.pb.cc.o 
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/_build_env/bin/x86_64-conda-linux-gnu-c++ -DGOOGLE_PROTOBUF_CMAKE_BUILD -DGTEST_LINKED_AS_SHARED_LIBRARY=1 -DHAVE_ZLIB -DPROTOBUF_USE_DLLS -I/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/build-shared -I/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src -I/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/third_party/utf8_range -fvisibility-inlines-hidden -fmessage-length=0 -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt  -ffunction-sections -pipe -isystem /home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla/include -fdebug-prefix-map=/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work=/usr/local/src/conda/libprotobuf-4.23.0 -fdebug-prefix-map=/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla=/usr/local/src/conda-prefix -DNDEBUG -O3 -DNDEBUG -std=gnu++17 -fPIE -DGOOGLE_PROTOBUF_TEST_PLUGIN_PATH=\"/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/build-shared/test_plugin\" -MD -MT CMakeFiles/tests.dir/src/google/protobuf/unittest.pb.cc.o -MF CMakeFiles/tests.dir/src/google/protobuf/unittest.pb.cc.o.d -o CMakeFiles/tests.dir/src/google/protobuf/unittest.pb.cc.o -c /home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc: In constructor 'constexpr protobuf_unittest::TestCord::TestCord(google::protobuf::internal::ConstantInitialized)':
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:3084:73: error: temporary of non-literal type 'absl::lts_20230125::Cord' in a constant expression
 3084 |       TestCord::Impl_::_default_optional_bytes_cord_default_func_{})}} {}
      |                                                                         ^
In file included from /home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/io/coded_stream.h:134,
                 from /home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.h:24,
                 from /home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:4:
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla/include/absl/strings/cord.h:150:7: note: 'absl::lts_20230125::Cord' is not literal because:
  150 | class Cord {
      |       ^~~~
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla/include/absl/strings/cord.h:150:7: note:   'absl::lts_20230125::Cord' has a non-trivial destructor
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc: At global scope:
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:3094:67: error: 'constinit' variable 'protobuf_unittest::_TestCord_default_instance_' does not have a constant initializer
 3094 |     PROTOBUF_ATTRIBUTE_INIT_PRIORITY1 TestCordDefaultTypeInternal _TestCord_default_instance_;
      |                                                                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:3094:67:   in 'constexpr' expansion of 'protobuf_unittest::TestCordDefaultTypeInternal()'
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:3086:54: error: 'constexpr protobuf_unittest::TestCord::TestCord(google::protobuf::internal::ConstantInitialized)' called in a constant expression
 3086 |   PROTOBUF_CONSTEXPR TestCordDefaultTypeInternal() : _instance(::_pbi::ConstantInitialized{}) {}
      |                                                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/conda/feedstock_root/build_artifacts/libprotobuf-split_1683984171729/work/src/google/protobuf/unittest.pb.cc:3078:20: note: 'constexpr protobuf_unittest::TestCord::TestCord(google::protobuf::internal::ConstantInitialized)' declared here
 3078 | PROTOBUF_CONSTEXPR TestCord::TestCord(
      |                    ^~~~~~~~

Though we're using abseil 20230125.2 and not 20230125.3 like protobuf, the difference is only one commit, and we've backported that one.

So while we've got abseil in the error log, I doubt that it has something to do with the unvendoring.

[h-vetinari]

Very unusually, linux-64 builds are the only ones that are failing, while everything else passes. So it must be related to the test targets (which we're not building on linux-aarch&ppc), and it's specific to GCC (i.e. doesn't appear with Clang & MSVC).

Any insights you might have on this would be much appreciated @coryan! :)

[coryan]

Very unusually, linux-64 builds are the only ones that are failing, while everything else passes. So it must be related to the test targets (which we're not building on linux-aarch&ppc), and it's specific to GCC (i.e. doesn't appear with Clang & MSVC).

I think you are running into the problems fixed by protocolbuffers/protobuf@ae7ec4e. It only affects GCC because only GCC defaults to C++17 at the moment (that is a guess).

That patch will be part of 23.1:

protocolbuffers/protobuf@v23.0...23.x

[h-vetinari]

I think you are running into the problems fixed by protocolbuffers/protobuf@ae7ec4e.

That looks bang on, thanks a lot!

It only affects GCC because only GCC defaults to C++17 at the moment (that is a guess).

I switched to C++17 globally in #144 already, but as the commit message mentions it seems a GCC issue

[github-actions]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!