mirrors/n-link
1
0
Fork 0
mirror of https://github.com/lights0123/n-link.git synced 2026-08-08 10:43:29 +00:00
Code Issues Projects Releases Wiki Activity

docs(plan): amend Plan A verification for unavailable oracle

Browse source 
n-link-cli ls / returns Busy then hangs (known CX II firmware issue,
owned by the user, out of scope here). Transfer-dependent gates are
downgraded to non-blocking; enumerate remains a full verification of the
IPC path since it reads USB descriptors only and does no libnspire
transfer.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Your Name 2026-07-19 09:23:10 -04:00
parent 8db940edc8
commit 94c62e810b
1 changed files with 36 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
docs/superpowers/plans/2026-07-19-desktop-tauri2-backend.md
View file

@ -22,6 +22,40 @@
**Oracle binary:** `/home/jjureta/Projects/n-link/nlink-cli/target/release/n-link-cli` **Oracle binary:** `/home/jjureta/Projects/n-link/nlink-cli/target/release/n-link-cli`
### AMENDMENT 2026-07-19: oracle unavailable
Verified on this machine: `n-link-cli ls /` returns `Busy` and then hangs
indefinitely (exit 124 at 45s, no output). Ruled out as causes: device-node
permissions (ACL grants the user rw), the udev rule (present), and device
disconnection (continuously enumerated).
This is the known TI-Nspire CX II firmware issue described in
`nlink-cli/README.md`. **The user has taken ownership of it and it is out of
scope for Plan A.**
Consequence for verification: any gate requiring a libnspire *data transfer*
cannot pass and is downgraded to non-blocking. Specifically:
| Gate | Status |
|---|---|
| Task 1 Step 5 (oracle baseline) | **Skipped** — cannot produce a listing |
| Task 7 Step 3, `open_device` / `list_dir` | **Non-blocking** — expected to hang or return `Busy` |
| Task 7 Step 4 (download + sha256 cross-check) | **Skipped entirely** |
What Plan A must still prove, unchanged and fully achievable:
1. A Tauri 2 window opens on webkit2gtk-4.1 — the actual goal of this port.
2. The backend compiles and all 13 commands register.
3. `enumerate` returns the CX II with correct camelCase keys. This reads only
USB descriptors via `rusb` and performs **no** libnspire transfer, so it is
unaffected by the `Busy` issue — and it exercises serde serialization, IPC,
and event wiring, which is the bulk of what this plan ports.
4. Hotplug `addDevice` / `removeDevice` events fire.
5. Headless `--help` works; deb and AppImage build, install, and render.
A transfer-path failure at Task 7 is therefore **expected** and must not be
treated as a port regression or a reason to halt.
--- ---
### Task 1: Install system dependencies and confirm the toolchain ### Task 1: Install system dependencies and confirm the toolchain
@ -1429,8 +1463,8 @@ Ubuntu 26.04 — the failure mode that motivated this port."
- [ ] A Tauri 2 window opens on Ubuntu 26.04 with webkit2gtk-4.1 - [ ] A Tauri 2 window opens on Ubuntu 26.04 with webkit2gtk-4.1
- [ ] All 13 commands are registered and callable from the webview - [ ] All 13 commands are registered and callable from the webview
- [ ] `enumerate` detects the CX II with correct camelCase payload keys - [ ] `enumerate` detects the CX II with correct camelCase payload keys
- [ ] `list_dir /` output matches `n-link-cli ls /` - [ ] ~~`list_dir /` output matches `n-link-cli ls /`~~ — **dropped**, see amendment
- [ ] A downloaded file's sha256 matches the same file fetched by `n-link-cli` - [ ] ~~Downloaded file sha256 matches `n-link-cli`~~ — **dropped**, see amendment
- [ ] Hotplug add/remove events fire - [ ] Hotplug add/remove events fire
- [ ] Headless CLI mode still works - [ ] Headless CLI mode still works
- [ ] `.deb` and `.AppImage` build; the release binary renders without a blank screen - [ ] `.deb` and `.AppImage` build; the release binary renders without a blank screen