LearningCircuit 653707a556 fix(encoding): add encoding="utf-8" to bare open() / read_text / write_text in examples and scripts (#4118) ...

Cleanup follow-up to #3797. The check-open-encoding hook was originally scoped
with exclude: ^(tests/|examples/|scripts/) because those directories had ~45
pre-existing bare open() calls and addressing them was out of scope for the
core Windows bug fix.

This commit:
  * adds encoding="utf-8" to 45 read/write call sites under examples/ and
    scripts/ — JSON benchmark results, config-doc generators, workflow
    status pages, and the datetime-timezone pre-commit hook
  * narrows the hook exclude to ^tests/ only, so future regressions in
    examples/scripts/ are blocked at commit time

Windows users running the benchmark scripts and config-doc generator would
previously hit silent failures or UnicodeDecodeErrors on non-ASCII content
under cp1252. The package itself was already protected by #3797.

2026-05-18 21:45:04 +02:00

..

api_usage

…

benchmarks

fix(encoding): add encoding="utf-8" to bare open() / read_text / write_text in examples and scripts (#4118)

2026-05-18 21:45:04 +02:00

elasticsearch

…

llm_integration

…

optimization

fix(encoding): add encoding="utf-8" to bare open() / read_text / write_text in examples and scripts (#4118)

2026-05-18 21:45:04 +02:00

2008-finicial-crisis.md

…

detailed_report_how_to_improve_retrieval_augmented_generation_in_p.md

…

example_browsecomp.py

…

fusion-energy-research-developments.md

…

medical_topic.md

…

run_benchmark.py

…

show_env_vars.py

…

traditional-spice-trade-routes.md

…