71 lines
3.2 KiB
Markdown
71 lines
3.2 KiB
Markdown
|
# ETSI ASN.1 modules
|
||
|
|
|
||
|
|
The ASN.1 schemas this project's UPER codecs are validated against. **Schemas only** — no parser
|
||
|
|
code. Nothing in the app or the firmware reads these files at build or run time; they exist so the
|
||
|
|
hand-written codecs in `app/src/main/java/.../domain/asn1/` can be checked against an independent
|
||
|
|
implementation.
|
||
|
|
|
||
|
|
## Why this is here rather than just documented
|
||
|
|
|
||
|
|
Three of this project's tests assert exact bytes:
|
||
|
|
|
||
|
|
- `CamEncodeGoldenTest` — the CAM this app transmits, byte for byte
|
||
|
|
- `DenmAirReceiveTest`, `SpatemUperCodecTest` — real captured frames with field values
|
||
|
|
|
||
|
|
Those expected values were produced by decoding with `asn1tools` compiled from these modules. Without
|
||
|
|
them the fixtures cannot be regenerated or re-verified, and a golden-byte test you cannot regenerate
|
||
|
|
is a test nobody can safely touch.
|
||
|
|
|
||
|
|
This matters because the project has shipped the same class of bug three times: a field encoded with
|
||
|
|
the wrong number of bits, which this codebase then read back with the *same* wrong number. Phone and
|
||
|
|
ESP32 agree perfectly with each other and with nothing else, so every internal round-trip test passes
|
||
|
|
while the frames on air are malformed (`CurvatureCalculationMode`, the GeoNetworking reserved bytes,
|
||
|
|
`yawRateConfidence`). Only a second, independent implementation catches that — which is what these
|
||
|
|
modules provide.
|
||
|
|
|
||
|
|
## Provenance
|
||
|
|
|
||
|
|
Taken from consider it GmbH's C-ITS-Parser:
|
||
|
|
|
||
|
|
- <https://github.com/consider-it/C-ITS-Parser>
|
||
|
|
- commit `f457426efc2486fac49a02fc9a1c8c7762d160e9`
|
||
|
|
- MIT licensed — see `LICENSE`, retained here as the licence requires
|
||
|
|
|
||
|
|
Only the seven `.asn` files below are copied, out of a 4.2 MB checkout. The upstream Rust parser is
|
||
|
|
not used by this project in any way. The schemas themselves are ETSI's standard definitions; the
|
||
|
|
upstream repo's contribution is assembling them into a compilable set.
|
||
|
|
|
||
|
|
| file | used for |
|
||
|
|
|---|---|
|
||
|
|
| `cam_1_4_1.asn` + `cdd_1_3_1_1.asn` | CAM encode/decode |
|
||
|
|
| `denm_1_3_1.asn` + `cdd_1_3_1_1.asn` | DENM decode |
|
||
|
|
| `spatem_2_2_1.asn`, `mapem_2_2_1.asn`, `dsrc_2_2_1.asn`, `cdd_2_2_1.asn` | SPATEM/MAPEM decode |
|
||
|
|
|
||
|
|
Note CAM/DENM use the release-1 common dictionary (`cdd_1_3_1_1`) while SPATEM/MAPEM use release 2
|
||
|
|
(`cdd_2_2_1`). Both are needed; they are not interchangeable.
|
||
|
|
|
||
|
|
These seven were verified sufficient on their own: copied into an empty directory, all three specs
|
||
|
|
compile and reproduce the committed golden bytes.
|
||
|
|
|
||
|
|
## Regenerating a fixture
|
||
|
|
|
||
|
|
Requires Python with `asn1tools` (verified with 0.167.0):
|
||
|
|
|
||
|
|
```python
|
||
|
|
import asn1tools
|
||
|
|
spec = asn1tools.compile_files(["asn1/cam_1_4_1.asn", "asn1/cdd_1_3_1_1.asn"], "uper")
|
||
|
|
spec.decode("CAM", raw_uper_bytes)
|
||
|
|
```
|
||
|
|
|
||
|
|
For SPATEM, compile `spatem_2_2_1.asn`, `mapem_2_2_1.asn`, `dsrc_2_2_1.asn`, `cdd_2_2_1.asn`
|
||
|
|
together and decode `"SPATEM"`.
|
||
|
|
|
||
|
|
**Never regenerate a golden fixture from this project's own encoder output** — that is precisely the
|
||
|
|
mistake these files exist to catch. Regenerate through `asn1tools`, or the test is worthless.
|
||
|
|
|
||
|
|
## Updating
|
||
|
|
|
||
|
|
Re-pin deliberately, not casually. If a newer ETSI release is adopted, copy the new modules, then
|
||
|
|
re-run the golden-byte tests and confirm any change in expected bytes is explained by the spec
|
||
|
|
change rather than by a decoder regression.
|