meshbay.git/packaging/rpm, branch main

feat: decode and re-encode video on the GPU where there is one

2026-09-18T12:37:08Z

A 1080p film is decoded by whoever watches it and re-encoded by the node when no browser can decode the source. Both were on the CPU, and on an Atom or Celeron mini-PC neither reaches real time — which is what `transcode_incompatible_video` exists to refuse. This adds the mechanism that makes refusing it unnecessary. Node — `hwaccel.py`: VA-API on Linux, Quick Sync or NVENC on Windows, established by encoding 1080p and reading the file back with ffprobe. Nothing is accepted that does not produce the exact profile and level `stream_init` announces, since the client checks that string before it trusts a byte: an encoder that wrote another level would make the node's own codec string a lie, and ffmpeg takes `-level 4.1` and `-level 41` from h264_qsv without saying which it understood. Three modes per stream — hardware decode and encode, hardware encode alone, libx264 — demoted per source codec, because a GPU that decodes HEVC may have no decoder for MPEG-4 Part 2 and only asking it finds out. A mode that fails is detected on an empty stdout before `stream_init` goes out, so the viewer sees one working stream and never an error. Client — Chromium ships VA-API off on Linux. It is enabled where a render node and a driver are present, then verified through `navigator.mediaCapabilities`: a no moves to the next GL backend on the next launch and an exhausted list drops the switches, so a renamed feature cannot pass for a feature that is on and `--ignore-gpu-blocklist` cannot survive on a machine it did not help. Feature lists now merge rather than overwrite — `appendSwitch` replaces the value, and a second caller would have silently cancelled the mDNS switch aiortc depends on. Packaging — the drivers are weak dependencies on all four formats, so a machine without a GPU installs exactly as before. `dpkg -i` and `rpm -ivh` ignore weak deps; `packaging/README.md` now says so. Windows needs no driver: the bundled ffmpeg already carries h264_qsv and h264_nvenc, and a re-pin that dropped them would cost every low-power Windows node its hardware encoding silently. AMD on Windows (AMF) and macOS (VideoToolbox) are named gaps, not oversights: neither could be tried anywhere in this project, and both re-encode in software as before. Co-Authored-By: Claude Opus 5

Update spec contact to devel@meshbay.org

2026-09-14T13:20:48Z

feat(packaging): ship a node firewall profile for inbound WebRTC

2026-09-04T12:10:16Z

The packages carried a profile for LAN casting and none for the node's own peer traffic, on the reasoning that the node exposes only a loopback admin API. That is true of its administration surface and false of its transport. WebRTC binds an ephemeral UDP port per connection, so there is no fixed port to open, and a connection succeeds only if one side can initiate. Browsers publish their host candidate as an mDNS `.local` name, which aioice cannot resolve on any platform and discards — so the node can never call a browser back, and the browser must call the node. A node that refuses unsolicited inbound UDP is unreachable from every browser on its own LAN, leaving reflexive candidates, which fail whenever both peers share one public IP and the router will not hairpin. Hit twice in one session on two different hosts: a firewalld zone narrowed to mdns + 19550-19553/tcp, and a ufw host with default deny-incoming. Both presented as "the app cannot connect", neither as a firewall message. Passive like the cast profile: packaged, not activated. The guide says to scope it to a LAN zone or source, including the libvirt case, where traffic from a guest to its own hypervisor is not masqueraded and so must be scoped to the guest subnet rather than the LAN. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01DtfG7z6wHWj8RKHCvxQtY1

fix(packaging): ship the example hub config, and stop leaving /etc/meshbay open

2026-09-02T09:22:23Z

Three defects, found while answering whether installing the .deb would land where the production server was just moved to by hand. - **The example config was never packaged.** `build-hub.sh` copied `packaging/conf/hub.toml.example` under `if [ -f ]`, and that path does not exist in this repo — so every package ever built shipped no example at all and said nothing about it. The postinst places no config either, on purpose (a shipped hub.toml is overwritten on upgrade; a shipped secret gets run in production), which left an installed hub with nothing to copy from. The file now exists, documents every key `config.py` reads including the captcha `allowed_hosts` the desktop client needs, and the copy is a hard failure rather than a silent skip. - **`/etc/meshbay` was created 0755.** It holds the hub's Ed25519 private key and its database password. The file modes protect the contents, but a world-listable config directory tells anyone with a shell what a hub keeps and where. Now 0750 root:meshbay, in both the deb postinst and the rpm scriptlet; the service reads it by group. - **The rpm would have failed to build on the new file.** `%files` claimed nothing under /etc, and rpmbuild refuses an installed file no line claims. It now declares the directory and the example, with explicit `%attr` and `%config` so an operator's edits become .rpmsave rather than vanishing. Package modes no longer follow the builder's umask either — the same source tree produced 775/664 on a machine with umask 002 and 755/644 with 022. `install -m` sets them. Verified by building: the deb now carries ./etc/meshbay/ at drwxr-x--- with hub.toml.example at 0644, and the embedded postinst tightens the directory as belt and braces rather than as the only thing making it right. The rpm path is unverified — no rpmbuild on this machine. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_014UtzVrzM7e2tG9fSpkR9ML

refactor(node): JSON-only control API, Node page absorbs the admin dashboard

2026-09-01T12:08:32Z

Remove the node daemon's server-rendered admin UI (GET / and /audit, the _render_* helpers and inline templates) and the `meshbay-node ui` CLI verb. The loopback control API stays; it is now JSON only, ruff-clean, and 453 lines (was 1074). Also drop three never-wired endpoints (/api/config, /api/chat/history, /ws/chat, plus broadcast_chat) and the pointless 18000/tcp firewall profiles. The desktop client's Node page (static/node-page.js) takes over what the dashboard showed, reorganised into six tabs (Overview, Groups, Roster, Peers, Audit, Settings): - Overview: version, node id, QUIC port, hub, index-cache maintenance - Roster: node-wide view with unpin - Peers and Audit: auto-load on open, no Load button - Audit: real usernames and group names (resolved from the roster and node.toml), Previous/Next pagination newest-first, Export CSV of every matching row - Settings: node settings, STUN, ICE, denylist, then Unlink from hub Backend: audit.get_entries gains `offset`; /api/audit and /api/peers resolve ids to names via a new _display_names helper; CSP tightened to default-src 'none' now that no HTML is served. draft-v6 sections 2.11 and 2.12 corrected -- the Node page uses the loopback API, not MNP. One capability is intentionally dropped: browser-based admin on a headless server. The CLI covers every operation there. See docs/refactor-node-ui.md. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_01MQCaZnde4Bjjdu84dhSuF5

feat(packaging): 4-package .deb/.rpm build system under /opt

2026-08-31T08:49:45Z

Shared venv architecture: meshbay-common owns the Python venv with all pip deps pre-installed; hub and node add only their code into it. Client is a standalone Electron app. No pip runs at install time. - Add build scripts (packaging/build/) for common, hub, node, client - Add orchestrator build-packages.sh with deb/rpm auto-detection - Add .deb control/postinst for all 4 packages - Add .rpm specs for all 4 packages (replaces python3-meshbay-common) - Add Gnome .desktop launcher and icon resizing - Add firewalld services (meshbay-cast, meshbay-node) and UFW profiles - Update systemd units to use /opt/meshbay-common/venv/bin/ paths - TMDB token baked into node package at build time via QE/node.env - Fix package-lock.json sync for protobufjs override Co-Authored-By: Claude Opus 4.6

chore: release 0.6.0

2026-08-23T13:15:47Z

The group UI's applications split, and the two missing-import bugs it surfaced and fixed along the way. MNP goes to 0.4: apps_enabled/apps_enabled_ack, and enabled_apps on the handshake ack, for the group-applications registry. Additive — a node that predates it is never sent the op, and a client that predates it never looks for the field. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_016SF6RKNBKg9qejmoMJ9ybA

feat(client): the platform seam, and an Electron shell that has never been run

2026-08-18T07:42:34Z

Stage D, and the honest half of it. D1 — the seam (done, and verified) ---------------------------------- `static/platform.js`. `HUB` becomes `platform.hubBase()` and the transport is built with the same base, so one address has one source. In a browser it returns '' and every path stays relative to the origin that served the page — the acceptance criterion for this split was "the browser SPA behaves identically", and it does. `platform.js` joins `_ASSETS`, or a change to it would not move the content hash and a cached browser would never ask for it. D2 — the shell (written, never launched) ----------------------------------------- **There is no npm on this machine. Electron was never installed and `packages/meshbay-client/` has not been run once.** That is stated here rather than discovered later. What is there: a main process serving the packaged interface over a privileged `app://` scheme (`secure` and `standard` are not cosmetic — without them the service worker refuses to register and streamed downloads break silently), a preload exposing an enumerated bridge that never passes a filesystem path, a window with `sandbox`, `contextIsolation` and no node integration, navigation away from the package refused, and a CSP where the hub is reachable over connect-src and is not a script source. The hub address arrives as a process argument because `platform.hubBase()` runs before anything can await. `test_desktop_shell.py` pins each of those by reading the source — the treatment `test_downloads.py` already gives the three browser save paths. It catches a property being removed and proves nothing about the application running. Two were checked by breaking them. The interface is *copied* into the package by `build/sync-ui.js` from the hub's static directory, and `ui/` is gitignored: a silent fork is the only real way to end up maintaining the interface twice. D3 — partial ------------ The bridge, and the part worth having now: safeStorage's backend is reported rather than assumed. On Linux it falls back to a fixed key when no keyring is running, silently — someone who believes the OS is holding their keys is told when it is not. The native key lifecycle belongs with D4 and needs a running application to mean anything. D8 — partial, and a real defect found -------------------------------------- `meshbay-node.spec` installed the SYSTEM template — the one carrying `User=%i` — into `%{_userunitdir}`. A user unit already runs as its owner and cannot carry `User=`; systemd refuses the file, so the packaged unit could never have started. Nothing noticed because nobody had built and installed the RPM. Two units now: the template to `%{_unitdir}`, and a new `meshbay-node-user.service` that a person enables themselves without a password — which is what lets the desktop client install a node without asking for one. It carries ExecReload, so `meshbay-node reload` does not have to stop a service somebody is streaming from, and documents the drop-in for a drive outside the home, RequiresMountsFor included. 798 tests pass; e2e.py still passes end to end. Nothing here was built or launched: no npm, no rpmbuild. Co-Authored-By: Claude Opus 5

chore: release 0.5.0

2026-08-16T13:29:30Z

Two rounds of features and one long hunt. The hub gained leaving a group, a cap of ten live public groups per owner, and the rule that a group is listed only once a node has announced it — with `prune-groups` to collect the ones that never got one. Presence rides on the group list, from the registry the hub already keeps for signaling. The web client speaks ten languages, splits Profile from Settings, and reads chat the way it is written: newest first, paging backwards. The rest was one symptom — "close the viewer, the next video hangs" — with three independent causes underneath, none of which the test suite or e2e.py could see. A background task the loop only weakly referenced, collected while it held a transcode slot. A connection-state handler that forgot a peer without stopping it. And `await proc.wait()` deadlocking on ffmpeg's own unread output, which no amount of SIGKILL resolves. Found by instrumenting the node and reading the log, after two confident fixes that addressed real but different bugs. MNP goes to 0.2: PING/PONG and backward chat paging, both additive. Co-Authored-By: Claude Opus 5

chore: release 0.4.0

2026-08-15T22:22:01Z

The web client speaks ten languages instead of one: French, Spanish, Brazilian Portuguese, Simplified Chinese, Japanese, German, Italian, Dutch and Polish, all formal, with `hub`, `node` and `GEK` deliberately left in English so the interface still matches the CLI and the docs. Catalogues are fetched per language rather than shipped together, plural forms go through Intl.PluralRules because Polish needs four of them, and locale matching keeps the region so pt-BR and zh-CN resolve to the files written for them. Splitting one module into a loader and ten catalogues gave the SPA a version dependency it did not have before, and the hub was serving static assets with no explicit freshness at all. A browser that cached half a deploy either rendered every string as its own key or, in the other direction, failed to link the module graph and showed nothing. Static responses now carry no-cache, which costs one conditional request and answers 304 with no body. The node and common packages carry no functional change; they move with the version because the three are released together. Co-Authored-By: Claude Opus 5