semantic-text-splitter: add build-semantic-text-splitter.yml for riscv64 wheels - #1520
Merged
Conversation
…v64 wheels Adds a riscv64 wheel build for semantic-text-splitter 0.32.0, a PyO3/maturin binding over the text-splitter crate. bindings/python/Cargo.toml's abi3 feature (pyo3/abi3-py310) is on by default, so one abi3 wheel serves cp310-cp314 and a separate build covers the free-threaded cp314t wheel that pyo3 produces automatically under Py_GIL_DISABLED.
luhenry
added a commit
that referenced
this pull request
Sep 8, 2026
maturin ships this project as a mixed-source package (semantic_text_splitter/ with an __init__.py wrapping the abi3 extension submodule), so the top-level module's __file__ ends in __init__.py, not .so. Check the wheel's RECORD via importlib.metadata.files() for a .so entry instead of the module's own __file__, matching what the abi3 build actually produced on riscv64 CI.
public PyPI ships no riscv64 wheel for tree-sitter-python 0.25.0, so pip built it from the sdist; that from-source build is missing the external scanner's symbols (ImportError: undefined symbol: tree_sitter_python_external_scanner_create). pypi.riseproject.dev already carries a correctly-built riscv64 wheel from this repo's own tree-sitter-python port, so route pip there via PIP_EXTRA_INDEX_URL.
test_hugging_face_from_file/test_markdown_hugging_face_from_file load a fixture via the relative path "tests/bert-base-cased.json", which only resolves from bindings/python (upstream's own CI sets that as the working directory). CIBW_TEST_SOURCES stages the full bindings/python/tests path under the test cwd, so invoking pytest from the staging root instead of bindings/python left those two tests unable to find the fixture file.
tree-sitter-python ships no cp314t wheel anywhere upstream (only cp310-abi3), and its PyPI sdist is missing src/scanner.c -- present in its own git tag, so this is a universal packaging gap, not riscv64-specific. Building it from sdist on the cp314t leg produces a real .so that dlopen rejects with "undefined symbol: tree_sitter_python_external_scanner_create". test_integration.py imports tree_sitter_python at module scope, so the whole suite is uncollectable on this interpreter regardless. Drop it from CIBW_TEST_REQUIRES on cp314t and skip pytest there; the compiled-extension and licence assertion (unaffected, checked before pytest runs) still proves the wheel itself is real.
luhenry
added a commit
that referenced
this pull request
Sep 8, 2026
luhenry
added a commit
that referenced
this pull request
Sep 8, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
semantic-text-splitter0.32.0Compiles benbrandt/text-splitter's Rust text-chunking library into the semantic-text-splitter PyO3 extension module. Upstream publishes no riscv64 wheel.
Mirrors upstream's
python.ymllinuxjob.Differs from upstream
PyO3/maturin-action- this repo's standard riscv64 build pathMatrix: cp310-abi3 (pyo3's abi3-py310 floor) + cp314t; no musllinux (rustup ships no riscv64 musl toolchain)
Testing
License: OK