The COVERAGE build type uses -O0 and gcov instrumentation, which slows
C++ query execution ~5-10x. This causes test_gil_released_during_query
to exceed its 0.5s timing threshold and fail with an inconclusive result.
- Set unique WORKING_DIRECTORY per test binary via cc_test()/cuda_test() to prevent filesystem path conflicts when running tests in parallel. Each test runs in ${CMAKE_BINARY_DIR}/test_tmp/${test_name}/.
- Enable parallel ctest execution in the unittest target with ProcessorCount-based --parallel flag (defaults to NPROC - 1).
- Set TEST_BINARY_DIR environment variable for crash recovery tests so they can locate helper binaries from isolated working directories.
- Update LocateDataGenerator() and LocateOptimizeGenerator() to search TEST_BINARY_DIR and TEST_BINARY_DIR/bin for helper executables.
* feat: linux ci with github runner
* fix: add CFLAGS with -march=native
* fix: fix linux x86 ci
* fix: with zen3 in x86 ci
* feat: nightly report with github-runner
* feat: refactor ci workflow
* feat: rename workflow
* feat: rename job
* feat: update ci badge
* feat: remove python coverage
* feat: set local_scheme with no-local-version
* fix: pre-install numpy for test pypi install zvec
* fix: add guess-next-dev
* feat: support cancel ci when has new pr
* feat: add job name
* feat: add job name for ci and fix concurrency group
rayx
dependabot[bot]
feihongxu0824
Cuiys
ZeFeng Yin
Salman Chishti