The five modules underneath this backend that were never about AI
sessions -- the pinned CA and leaf, QR enrollment and the bearer token,
wg0 binding and the certificate's SANs, owner-only files, and the RON
house rules -- were written twice, once here and once in dev-updater,
and had drifted. They now come from the submodule, as a path dependency
so both projects stay locked to one commit.
What stayed is what makes this project itself: the routes, the drivers,
the config schema, and the auth middleware, which is generic over this
server's state. Sharing a transport is worth doing; sharing an API would
mean inventing a vocabulary neither project wants.
Four dependencies go with the code -- rcgen, qrcode, subtle and if-addrs
are no longer named here at all -- and the three that remain are now
described by what still uses them rather than by what used to.
Verified by running it, not only by building: a fresh server generates
its CA, prints an `aiapp://enroll` QR with the scheme now passed as a
parameter, covers 127.0.0.1, 10.0.2.2 and wg0's 10.66.0.1 in the leaf,
answers an enrolled token and returns 401 without one, and writes
config.ron in the house rules with every file owner-only. 36 tests pass,
clippy is silent, rustfmt is clean.
The second half of the llama.cpp work: a session can now name a downloaded
model and talk to it. `llama-server` is spawned through the same transport
as any other driver, polled until the model is loaded, then driven over its
OpenAI-compatible streaming endpoint and translated into the same events
the Claude driver emits -- so the transcript, the SSE stream and the phone
need to know nothing new.
**The conversation is rebuilt from the transcript, not held in the driver.**
llama-server is stateless between requests, so the whole history goes with
every one, and the obvious place to keep it is a Vec in the driver. That
fails the requirement: memory in a driver is invisible to a second device
and gone on restart, and this app is meant to work across devices. Reading
it back also means the model is prompted with exactly what the phone was
shown -- including a reply that was interrupted half way, which is in the
transcript because the deltas were already emitted.
That leaves the Claude driver as the odd one out rather than this one: the
CLI's memory of a conversation is a cache in front of the same transcript,
not a second truth. Said so at the top of llama.rs, because it is the sort
of inconsistency that gets "fixed" in the wrong direction.
Session settings arrive as a driver-interpreted `params` map rather than
new typed fields, so the shared schema does not grow one dialect's
vocabulary. Context size, gpu layers and threads become server flags;
temperature and the rest ride on each request, so changing them need not
reload a model.
**Also fixes an orphan this feature would have created.** Drivers set
kill_on_drop, which covers a session being deleted -- but nothing drops on
the way out of a SIGTERM, so signalling the server left its children
running. For the Claude CLI that is untidy; for a llama-server holding a
model it is gigabytes belonging to nobody. The server now stops its
sessions on SIGTERM and SIGINT. Found by killing a test server and noticing
two 600 MB processes still resident.
Remote llama sessions are refused rather than half-working: the model is
reached over HTTP, and forwarding that port to an ssh host is the "reach
this port" operation the transport does not have yet.
Verified end to end against a real model: downloaded Qwen3-0.6B Q8_0
through the app's own download route, spawned a session on it, and held a
two-turn conversation -- "my favourite colour is teal" then "what is my
favourite colour?", answered "teal", which is the transcript replay doing
its job. Token counts arrive. An earlier attempt with the IQ2_XXS quant
produced fluent nonsense, which turned out to be the quantisation rather
than the pipeline: llama-cli produces the same from that file directly.
Four unit tests cover the fold and the path guard.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017xn8nHw1tw1R6PtiY1eEtw
The code was hand-formatted -- close to rustfmt's output but not it, mostly
in keeping chains and call arguments on one line where the formatter would
break them. That is a per-line decision every future change has to make
again, and reproducing it would mean a config whose only job is to preserve
how the code already looks.
So this is `cargo fmt` at its defaults, with no rustfmt.toml, which is
where the sibling dev-updater checkout already sits: it is clean at the
defaults today, so the two repos now agree on layout without either of them
configuring it.
Formatting only -- no behaviour, no renames, nothing reordered. Verified
after: cargo test (35 pass), cargo clippy --all-targets clean, cargo fmt
--check clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017xn8nHw1tw1R6PtiY1eEtw
A doc comment is not compiled, so nothing catches one that has outlived
what it described. `cargo doc` does catch a subset, and it was failing:
three unresolved intra-doc links in config.rs, all mine from today -- two
naming `parse`/`render` from outside the module that defines them, and one
`format` that is ambiguous with the macro. Rustdoc now runs clean with
broken_intra_doc_links denied.
auth.rs's module doc pointed at a test called `token_is_never_logged`,
which no longer exists: it was folded into the combined gating+logging test
because tracing caches callsite interest process-wide. The doc now
describes the test that is there and says why it is one test.
AGENTS.md contradicted itself twice, both checked against a running server
rather than by reading. Its curl example passed `--cacert certs/ca.pem`, a
path relative to the checkout, while the bullet above it correctly says the
CA is generated under $XDG_CONFIG_HOME -- so the documented command fails
before it connects; with the real path it returns `[]`. And it said wg0
"doesn't exist on this machine yet" as the reason for --bind 127.0.0.1,
which the section below it already contradicts: wg0 is up at 10.66.0.1, and
the actual reason is that the emulator dials 10.0.2.2 and cannot reach it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017xn8nHw1tw1R6PtiY1eEtw
The same move, for the same reason: this file is written and read by hand,
and JSON has no comments to say why a host is configured the way it is.
Both house rules come across with it, in config.rs's `format` module and
nowhere else -- a file is the *body* of the config, so no outer parentheses
and nothing indented for them, and `Some` is implicit, which is what makes
`skip_serializing_if` on every optional field load-bearing rather than
tidiness.
The switch is outright: there is no reader for the old format. That is
invisible everywhere except here, because this file holds the enrolled
token hashes -- starting empty leaves the phone unable to talk to the
server and looks, from the phone, like the config having been lost. So a
config.json left beside the new file is named in the log and left alone,
rather than read or deleted.
One wart, documented at DriverKind: the kebab-case spelling is the string
the phone compares against, so it stays, and the file pays for it with
`kind: r#claude-cli` -- a hyphen is not a RON identifier. Renaming the
variant would change what an already-installed build is talking to.
Verified: cargo test, cargo clippy --all-targets, and a real start against
a scratch state directory -- a hand-typed config with comments and a bare
`port: 2222` loads, and what the server writes back sits at column 0 with
no Some(...) in it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017xn8nHw1tw1R6PtiY1eEtw
The whole pipe behind one Driver trait and a common event model:
spawn/list/delete sessions, message + question answering, append-only
JSONL transcripts whose sequence numbers are the phone's resume cursor
(surviving backend restarts), bearer-token middleware wrapping every
route including the fallback, wg0-only binding that fails closed, and
first-run token enrollment via a terminal QR.
Verified: cargo test (10), clippy clean, and curl end-to-end over pinned
TLS -- auth rejection, spawn, streamed SSE replay/resume, /question
round trip, restart continuing seq numbers, delete removing everything.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017xn8nHw1tw1R6PtiY1eEtw
iris
