benchmarksPublic
benchmarksPublic
Commit 878fc46b
M
docs/vcs-conformance.md
@@ -70,8 +70,12 @@
7070
the same failure there is a regression). Rows mixing profiles are UNMEASURED;7171
- coverage: `coverage.<profile>.expected_skips` lists the skips a recorded7272
profile may have. Any other skip (a pass or xfail that turned into a skip) is a73-
coverage loss (UNMEASURED). Unrecorded profiles accept capability/gated skips74-
and say so; record the profile once its first full run is reviewed;73+
coverage loss (UNMEASURED). On an unrecorded profile only behaviours needing a74+
known filesystem-dependent capability (case sensitivity, Unicode75+
normalization, non-UTF-8 names) or gated behaviours may skip; any other skip76+
(e.g. a broken symlink/exec probe) is UNMEASURED. A row without `fs_profile`77+
is UNMEASURED, and a run with zero measured Oak behaviours is always78+
UNMEASURED. Record a profile once its first full run is reviewed;7579
- harness errors, reader-unmeasured legs and missing behaviours are UNMEASURED.7680
Git rows are the reference arm and never gate.7781
@@ -86,7 +90,8 @@
8690
binary; regressions make the verdict REGRESSED, unmeasured UNMEASURED, and8791
FIXED entries are printed as notes.8892
- `publish_gate.py` check 14: every published Oak row must carry binary identity89-
(`binary_sha256`, or `subject_details[subject].binary_sha256`); for each93+
(`binary_sha256`, or `subject_details[subject].binary_sha256`); the host's94+
`fs_profile` must have a recorded `coverage` entry; for each9095
published Oak binary the gate is evaluated on that binary's conformance rows9196
only (rows of another build can neither satisfy nor fail it) and must be PASS9297
with no FIXED entries.
M
scripts/oakbench/conformance_gate.py
Loading diff…
M
scripts/publish_gate.py
Loading diff…
M
tests/test_conformance_gate.py
Loading diff…