run 25
This commit is contained in:
Aiden
@@ -5362,3 +5362,195 @@ Best current follow-up:
|
||||
- test exact and host-shaped handling of `07 C0 2F 95 09 2E`
|
||||
- optionally compare whether exact `EC -> 7B` echo is timing-sensitive, since
|
||||
the host-shaped mirror path did not reach the same family
|
||||
|
||||
### HE18: Exact Echo Of `07 C0 2F 95 09 2E`
|
||||
|
||||
Recreate the `EC -> 7B` exact-echo branch, then immediately send the new
|
||||
`2F 95 09` family frame back exactly as seen.
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "07 80 7B 50 26 D0" --frame "07 C0 2F 95 09 2E" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-echo-exact-2f95092e.txt
|
||||
```
|
||||
|
||||
### HE19: Host-Shaped Mirror Of `07 C0 2F 95 09 2E`
|
||||
|
||||
Host-shaped checksum for `00 00 2F 95 09 ??` is `E9`.
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "07 80 7B 50 26 D0" --frame "00 00 2F 95 09 E9" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-echo-host-2f9509.txt
|
||||
```
|
||||
|
||||
What would count as a hit:
|
||||
|
||||
- any non-heartbeat response after the exact or host-shaped `2F 95 09` frame
|
||||
- a repeat of `07 C0 2F 95 09 2E`
|
||||
- any fresh family after the `2F` stage that suggests a real chained exchange
|
||||
- any sign that the `EC -> 7B -> 2F` path is the closest thing yet to a proper
|
||||
request/response ladder
|
||||
|
||||
### 2026-05-13 `2F` Mirror Result
|
||||
|
||||
Captures:
|
||||
|
||||
- `captures/rcp-heartbeat-echo-exact-2f95092e.txt`
|
||||
- `captures/rcp-heartbeat-echo-host-2f9509.txt`
|
||||
|
||||
Result summary:
|
||||
|
||||
- These runs did **not** extend the `EC -> 7B -> 2F` path into a stable next
|
||||
stage.
|
||||
- Instead, group 1 shifted the `EC` selector response itself into a new sibling
|
||||
family:
|
||||
- `07 80 FB 50 26 50`
|
||||
- That happened in both runs, before the `2F` follow-up frame was even sent.
|
||||
- After that:
|
||||
- exact `07 C0 2F 95 09 2E` produced only heartbeat
|
||||
- host-shaped `00 00 2F 95 09 E9` also produced only heartbeat
|
||||
- Group 2 was heartbeat-only in both tests.
|
||||
|
||||
Observed group-1 shape:
|
||||
|
||||
| Step | Exact-echo run | Host-mirror run |
|
||||
| --- | --- | --- |
|
||||
| `00 00 EC 40 30 C6` | `07 80 FB 50 26 50` x2 | `07 80 FB 50 26 50` x3 |
|
||||
| `07 80 7B 50 26 D0` | heartbeat only | heartbeat only |
|
||||
| `2F` follow-up | heartbeat only | heartbeat only |
|
||||
|
||||
Interpretation:
|
||||
|
||||
- We did **not** get a reproducible chained reply to the `2F` stage.
|
||||
- The more important finding is that the `EC` branch itself is context-sensitive
|
||||
and can emit at least two sibling downstream families:
|
||||
- `07 80 7B 50 26 D0`
|
||||
- `07 80 FB 50 26 50`
|
||||
- That makes the `EC` branch look more like a selector into a family space than
|
||||
a strict linear ladder.
|
||||
- `07 C0 2F 95 09 2E` is still real from the earlier run, but these follow-ups
|
||||
did not confirm it as the next stable step in an ongoing exchange.
|
||||
|
||||
Best current model:
|
||||
|
||||
- We are getting closer to a *structured* understanding of the RCP:
|
||||
certain host-side `Ex` values reliably push it into specific response
|
||||
families.
|
||||
- But we are **not yet** at a stable "conversation" where the panel is clearly
|
||||
accepting our last reply and moving to the next deterministic turn.
|
||||
- Right now the strongest evidence is for:
|
||||
- selector-like host entries (`E8`, `E9`, `EC`)
|
||||
- family-specific downstream responses (`7A`, `7B`, `FB`)
|
||||
- occasional exact-echo sensitivity on some branches
|
||||
- but not a fully reproducible multi-turn protocol ladder yet
|
||||
|
||||
### HE20: `EC` Timing / Context Split
|
||||
|
||||
Try to separate whether `EC` chooses `7B` vs `FB` because of timing around the
|
||||
selector step, or because of the deeper branch context.
|
||||
|
||||
#### HE20a: `EC` With Shorter Spacing
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.10 --read-after-frame 0.25 --read-after-group 0.8 --log captures/rcp-heartbeat-ec-short-spacing.txt
|
||||
```
|
||||
|
||||
#### HE20b: `EC` With Longer Spacing
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.35 --read-after-frame 0.35 --read-after-group 0.8 --log captures/rcp-heartbeat-ec-long-spacing.txt
|
||||
```
|
||||
|
||||
#### HE20c: `EC` Without Leading `A0`
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 EC 40 30 C6" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-ec-no-a0.txt
|
||||
```
|
||||
|
||||
What would count as a hit:
|
||||
|
||||
- one setup consistently yielding `07 80 7B 50 26 D0`
|
||||
- another setup consistently yielding `07 80 FB 50 26 50`
|
||||
- evidence that leading `A0` is part of the selector context rather than just a
|
||||
neutral primer
|
||||
|
||||
### HE21: Exact Echo Of `07 80 FB 50 26 50`
|
||||
|
||||
If `FB` is a real sibling branch, exact echoing it may show whether it behaves
|
||||
more like `7B` or like the inert `7A` families.
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "07 80 FB 50 26 50" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-echo-exact-fb502650.txt
|
||||
```
|
||||
|
||||
### HE22: Host-Shaped Mirror Of `07 80 FB 50 26 50`
|
||||
|
||||
Host-shaped checksum for `00 00 FB 50 26 ??` is `D7`.
|
||||
|
||||
```powershell
|
||||
python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 EC 40 30 C6" --frame "00 00 FB 50 26 D7" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-echo-host-fb5026.txt
|
||||
```
|
||||
|
||||
What would count as a hit:
|
||||
|
||||
- any non-heartbeat response after exact or host-shaped `FB`
|
||||
- crossover from `FB` into `2F`, `7B`, or another fresh family
|
||||
- evidence that `FB` is a meaningful downstream branch rather than just a
|
||||
selector-side variant
|
||||
|
||||
### 2026-05-13 `EC` Timing And `FB` Result
|
||||
|
||||
Captures:
|
||||
|
||||
- `captures/rcp-heartbeat-ec-short-spacing.txt`
|
||||
- `captures/rcp-heartbeat-ec-long-spacing.txt`
|
||||
- `captures/rcp-heartbeat-ec-no-a0.txt`
|
||||
- `captures/rcp-heartbeat-echo-exact-fb502650.txt`
|
||||
- `captures/rcp-heartbeat-echo-host-fb5026.txt`
|
||||
|
||||
#### HE20: Timing / Context Split
|
||||
|
||||
Observed outcomes:
|
||||
|
||||
| Setup | Group-1 result |
|
||||
| --- | --- |
|
||||
| short spacing (`0.10 s`) | `07 80 7B 50 26 D0` |
|
||||
| long spacing (`0.35 s`) | `07 80 7B 50 26 D0` |
|
||||
| no leading `A0` | `07 80 C0 40 30 6D` |
|
||||
|
||||
Interpretation:
|
||||
|
||||
- The `A0` lead-in matters a lot for the `EC` branch.
|
||||
- With `A0` present, both short and long spacing still favored
|
||||
`07 80 7B 50 26 D0`.
|
||||
- Without `A0`, `EC` collapsed back into the known heartbeat-family transient
|
||||
`07 80 C0 40 30 6D`.
|
||||
- So the best current read is that `A0` is part of the selector context for the
|
||||
`EC -> 7B/FB` family space, not just a neutral primer.
|
||||
- Timing still may matter for `7B` vs `FB`, but this batch says context matters
|
||||
more strongly than spacing in the tested range.
|
||||
|
||||
#### HE21 / HE22: `FB` Echo Handling
|
||||
|
||||
Observed outcomes:
|
||||
|
||||
| Setup | Group-1 result |
|
||||
| --- | --- |
|
||||
| exact `07 80 FB 50 26 50` | `EC` again produced `07 80 7B 50 26 D0`; exact `FB` echo produced no new family |
|
||||
| host-shaped `00 00 FB 50 26 D7` | `EC` produced `07 80 7B 50 26 D0`; host-shaped `FB` mirror produced no new family |
|
||||
|
||||
Interpretation:
|
||||
|
||||
- `FB` did not behave like a meaningful next-turn reply target.
|
||||
- Both exact and host-shaped `FB` handling fell flat after the `EC` selector
|
||||
produced a `7B` response in these runs.
|
||||
- That weakens the idea that `FB` is a stable downstream branch command. It now
|
||||
looks more like a sibling family observation that can appear on the `EC`
|
||||
branch, but not something the panel predictably wants answered.
|
||||
|
||||
Best current `EC` model:
|
||||
|
||||
- `A0 + EC` can open a selector-like family space.
|
||||
- In that family space:
|
||||
- `7B` is the most stable downstream response so far
|
||||
- `FB` is real, but less stable and not yet actionable
|
||||
- without `A0`, `EC` falls back toward heartbeat-family behavior
|
||||
- This is closer to a controlled state map than where we started, but it is
|
||||
still not a stable multi-turn "conversation" ladder.
|
||||
|
||||
Reference in New Issue
Block a user