mirror of
https://github.com/logos-messaging/nim-ffi.git
synced 2026-06-20 16:29:31 +00:00
19 lines
930 B
Plaintext
19 lines
930 B
Plaintext
# ThreadSanitizer suppressions for the FFI test suite.
|
|
#
|
|
# Format reference: https://github.com/google/sanitizers/wiki/ThreadSanitizerSuppressions
|
|
#
|
|
# Suppress the intentional thread leak from
|
|
# `destroyFFIContext does not hang when event loop is blocked`:
|
|
# when the FFI event loop is wedged by a synchronous handler, clearContext
|
|
# in ffi/ffi_context.nim deliberately skips joinThread on ThreadExitTimeout
|
|
# rather than hang the caller forever. The thread terminates cleanly, but
|
|
# was never joined — TSan reports that as a thread leak.
|
|
thread:*initContextResources*
|
|
|
|
# Same intentional-leak path: because the FFI thread is never joined,
|
|
# TSan cannot prove happens-before between its writes (inside Nim's
|
|
# threadProcWrapper / the embedded Thread[T] struct on the test's stack)
|
|
# and the next test's reuse of the same stack region. Reported as a
|
|
# data race; root cause is the documented leak above.
|
|
race:*threadProcWrapper*
|