mongo-c-driver Validation Report¶
Measured 2026-10-15 (raw artifact), generated 2026-10-15 — SecantusDB 0.7.0b3 vs mongo-c-driver 57dba9c (vendor/mongo-c-driver/).
Run uv run python -m invoke validate-c to refresh. The official MongoDB C driver (libmongoc) is the lowest-level official client — and (with the Go and PHP-extension gauges) one of the strictest wire-protocol checks.
Summary¶
Suite |
Passed |
Failed |
Skipped |
Total |
Pass rate |
|---|---|---|---|---|---|
|
93 |
0 |
10 |
103 |
100.0% |
|
102 |
2 |
17 |
121 |
98.0% |
|
145 |
0 |
12 |
157 |
100.0% |
|
70 |
0 |
0 |
70 |
100.0% |
|
19 |
0 |
0 |
19 |
100.0% |
|
6 |
0 |
0 |
6 |
100.0% |
|
16 |
0 |
0 |
16 |
100.0% |
|
6 |
0 |
0 |
6 |
100.0% |
|
13 |
0 |
0 |
13 |
100.0% |
|
12 |
0 |
1 |
13 |
100.0% |
|
23 |
0 |
2 |
25 |
100.0% |
|
11 |
0 |
0 |
11 |
100.0% |
|
5 |
0 |
0 |
5 |
100.0% |
|
34 |
0 |
1 |
35 |
100.0% |
|
171 |
0 |
1 |
172 |
100.0% |
|
9 |
0 |
0 |
9 |
100.0% |
|
10 |
0 |
1 |
11 |
100.0% |
|
32 |
0 |
2 |
34 |
100.0% |
|
6 |
0 |
0 |
6 |
100.0% |
|
8 |
0 |
1 |
9 |
100.0% |
Overall |
791 |
2 |
48 |
841 |
99.7% |
Failures (2)¶
First 30 failed tests for triage:
/Client/ipv6/single
/Client/ipv6/single
How this is generated¶
invoke validate-c builds the vendored driver’s test-libmongoc binary once (CMake, ENABLE_TESTS=ON) and runs it TWICE, each against a fresh SecantusDB daemon on an ephemeral port with MONGOC_TEST_URI pointed at it, writing JSON results via -F: the curated -l prefixes against the default single-node replica set, then the few tests that assert standalone semantics (STANDALONE_ONLY) against a --standalone daemon. The two are merged into one result set. The in-scope prefixes, the skip-list of out-of-scope tests and the standalone list live in c_validation/include_paths.py.