GH-35497: [C++] Use the latest tagged version of flatbuffers#38192
GH-35497: [C++] Use the latest tagged version of flatbuffers#38192pitrou merged 5 commits intoapache:mainfrom
Conversation
I used this hacky script to do it.
#!/bin/sh
SRC="${FLATBUFFERS_REPO}/include/flatbuffers"
DEST="${ARROW_REPO}/cpp/thirdparty/flatbuffers/include/flatbuffers/"
cp "${SRC}/allocator.h" "${DEST}"
cp "${SRC}/array.h" "${DEST}"
cp "${SRC}/base.h" "${DEST}"
cp "${SRC}/buffer.h" "${DEST}"
cp "${SRC}/buffer_ref.h" "${DEST}"
cp "${SRC}/default_allocator.h" "${DEST}"
cp "${SRC}/detached_buffer.h" "${DEST}"
cp "${SRC}/flatbuffer_builder.h" "${DEST}"
cp "${SRC}/flatbuffers.h" "${DEST}"
cp "${SRC}/stl_emulation.h" "${DEST}"
cp "${SRC}/string.h" "${DEST}"
cp "${SRC}/struct.h" "${DEST}"
cp "${SRC}/table.h" "${DEST}"
cp "${SRC}/vector.h" "${DEST}"
cp "${SRC}/vector_downward.h" "${DEST}"
cp "${SRC}/verifier.h" "${DEST}"
There was a problem hiding this comment.
If we switch to C++17-style arrow_vendored_private::flatbuffers it probably reduces the patch size and may make automation more doable as well.
There was a problem hiding this comment.
Is this change actually necessary or would the original "Check for std::string_view (in c++17)" snippet above work for us? We probably want to minimize the patch size.
There was a problem hiding this comment.
You're right. Since this version doesn't check for absl::string_view first, we don't need it.
|
@github-actions crossbow submit -g cpp -g python |
|
Revision: b019254 Submitted crossbow builds: ursacomputing/crossbow @ actions-2c6d2bba81 |
|
CI failures look unrelated, I'll merge. Thanks @felipecrv ! |
|
After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 01b42d5. There were 3 benchmark results indicating a performance regression:
The full Conbench report has more details. It also includes information about 1 possible false positive for unstable benchmarks that are known to sometimes produce them. |
|
@github-actions crossbow submit test-skyhook-integration |
|
Revision: b019254 Submitted crossbow builds: ursacomputing/crossbow @ actions-2eb8774cee
|
|
@felipecrv It seems that this broke Could you re-generate |
Sure. Let me create an issue and submit a PR soon. |
…pache#38192) ### Rationale for this change To use a more modern version of `flatc` in Arrow. ### What changes are included in this PR? 1) Re-generating the the C++ files with a `flatc` based on the latest tag on the `flatbuffer` repo. 2) Copy the minimal set of includes to our vendored folder 3) Manually re-apply the patches that have been applied to the previous headers to fix issues ### Are these changes tested? Tested by building everything correctly. * Closes: apache#35497 Authored-by: Felipe Oliveira Carvalho <felipekde@gmail.com> Signed-off-by: Antoine Pitrou <antoine@python.org>
…pache#38192) ### Rationale for this change To use a more modern version of `flatc` in Arrow. ### What changes are included in this PR? 1) Re-generating the the C++ files with a `flatc` based on the latest tag on the `flatbuffer` repo. 2) Copy the minimal set of includes to our vendored folder 3) Manually re-apply the patches that have been applied to the previous headers to fix issues ### Are these changes tested? Tested by building everything correctly. * Closes: apache#35497 Authored-by: Felipe Oliveira Carvalho <felipekde@gmail.com> Signed-off-by: Antoine Pitrou <antoine@python.org>
Rationale for this change
To use a more modern version of
flatcin Arrow.What changes are included in this PR?
flatcbased on the latest tag on theflatbufferrepo.Are these changes tested?
Tested by building everything correctly.