The same change the Mac build got in 1.10.0: the bundle carries the shell
alone; the admin page's catalog installs modules into %APPDATA%\ProdDash.
README.txt gains the step.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The macOS app and the Windows bundle shipped every module in the repo, on
by default, and a shell update silently added whatever new module had
landed on main — Band Lineup appeared on AB that way. A module was also
only as fresh as the last shell update. Now a packaged install (no .git —
the same test the updater uses to choose the archive over git pull)
carries no modules at all: every one arrives from the admin page's catalog
into the data directory, and the shell's own update never writes modules/.
A git checkout keeps running the modules in its tree.
Existing installs move over by themselves: at the first start, the bundled
modules that are switched on are copied into the data directory as
installed copies (a compiled helper keeps its mode), an uninstalled one is
dropped, and the app folder's modules/ is removed. Settings are keyed by
module id, so nothing changes for the operator.
Around it: the dashboard's empty state says "No modules yet — open Admin"
when the registry is empty; the admin page says "Nothing installed yet",
puts a checkbox on every compatible catalog entry and an Install selected
button (POST /api/admin/modules/install {ids}) that fetches them in one go.
The Mac build stops copying modules/ and no longer compiles or signs Swift
helpers — those ship prebuilt inside a module, per the guide.
tools/unbundled.test.js drives a packaged copy of the shell through the
move, an archive update that must not bring modules back, and a multi-install.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
launcher/macos/build.sh takes ARCH=x86_64 for Intel Macs — the Swift target
and the embedded Node follow (nodejs.org calls that one darwin-x64). The
default stays arm64.
launcher/windows/build.sh is new: the same payload plus Node for Windows,
checksum-verified against nodejs.org, and ProdDash.bat, which keeps the
server running and starts it again on exit code 75 — the admin page's update
restart — the way the macOS launcher does. No Windows launcher app yet.
The README's Run it section now points at the Releases page and the
ProdDash-releases repo that holds the files, one line each for Apple
silicon, Intel and Windows.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Two things the audit found, both about the archive path every shipped .app
uses.
**Settings could be wiped.** The update kept config/modules.json and
config/layouts/ by name, but not config/proddash.json — which holds the admin
PIN, the theme and the forge pointer. On an install whose data directory IS
the app's own config/ (an unwritable Application Support, or PRODDASH_DATA_DIR
aimed inside the app), an update wrote the repo's shipped defaults over it:
the theme went, and admin silently unlocked. Nothing inside the data
directory is written now, wherever it happens to sit, and the two legacy
names stay listed for an install whose state never moved out.
**A bad update was unrecoverable.** No rollback, and a fresh .app would not
help: seeding only replaces an *older* copy, and a broken one carries a
higher version — so the fix was deleting a folder nothing told you about.
Three changes, in the order they catch things:
- The download must parse before anything is swapped in. `node --check`
against the new server.js, and a failure is reported with the syntax error
and leaves the running copy untouched.
- Past that, the copy being replaced is set aside as `app.previous`, and
dropped once the new version has stayed up ninety seconds — so that folder
existing means "the last update has not proven itself yet", and never
anything older.
- The launcher, which already gives up after four stops in two minutes, now
looks for that folder before it does and puts the old copy back. Once: if
the old one will not run either, the update was not the problem.
The launcher half only reaches a machine with a new .app; the server half
travels by the normal update.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Installing is dragging ProdDash to Applications. The app carries the
dashboard and the Node that runs it, so nothing has to be installed first and
nothing can go missing underneath it.
ProdDash.app/Contents/
MacOS/ProdDash the launcher
Resources/node/bin/node Node 24.21.0 LTS, fetched and checksum-verified
Resources/app/ server.js, public/, modules/
Everything mutable stays where it already was, in ~/Library/Application
Support/ProdDash — and that now includes app/, the copy of ProdDash that
actually runs. The copy is not ceremony: ProdDash updates itself in place,
and nothing may write inside a signed .app. So the bundle holds a read-only
baseline and the data directory holds the running copy, which means both
routes to a new version keep working and don't fight — installing a newer
.app replaces the copy, while a copy that has already updated itself past the
bundle is left alone.
build.sh now assembles all of it: fetches Node against nodejs.org's published
checksums (cached in .cache/), embeds the server, compiles each module's
native helper so the Mac running ProdDash needs no compiler, signs nested
code before the bundle, and builds the disk image. --dev skips the payload
for launcher work, which still runs against a checkout.
Signing is set up for a Developer ID with the hardened runtime and
notarisation (--notarize), and falls back to ad-hoc when no such certificate
is installed. Entitlements are split by what each binary actually does: JIT
for node, audio input for the launcher and for module helpers that read one.
Built and run end to end: first-run install, bundled node, dashboard and
admin serving, re-seed in both directions (newer app replaces an older copy;
an updated copy survives), disk image mounts and its app verifies.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A pass for less prose and fewer boxes.
- Header shows the version and node; the port is already in the addresses
below it, and the module list belonged in the log.
- Settings: "Network" over the port (no explanation under it), "Startup"
instead of "Starting up", and no node chooser — the path stays visible,
since which node runs is the first thing worth knowing when it won't.
- Permissions: one line per permission — a status dot, its name, and the one
thing you can do about it. The standing explanation is gone and the
modules' own reasons moved to each row's tooltip. Local network results
still appear under their row after a check, with a short line and a link
only when nothing answered at all. Re-check sits centred at the bottom,
outside the group.
The app-level "notify me" switch goes with it: macOS's own per-app
notification setting already does that job, so notifications now follow the
permission alone, asked for by the row's Enable button rather than on first
run.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A native Swift app (launcher/macos, ~2 MB, no dependencies) that starts,
stops and watches over the server from the menu bar — and holds the macOS
permissions its modules need.
The permissions are the point as much as the buttons are. macOS grants
privacy permissions to the application that owns a process, and a child
inherits its parent's identity, so with the server started from here node —
and the Timers module's ltc-capture — belong to this app bundle and the
microphone and local-network prompts are asked, and remembered, as ProdDash.
Started over SSH or from a launchd job there is nobody to ask: the permission
is silently absent rather than refused, which is the "endless zeroes" failure
ltc-listener.js warns about.
The shell was already built for a supervisor: with PRODDASH_LAUNCHER=1 it
exits 75 instead of respawning itself when the admin page applies an update,
which the launcher reads as a restart request rather than a fault. Crashes
get a backed-off restart (1s, 2s, 5s, 10s) and four in two minutes stops the
cycle. The server is spawned into its own process group, so stopping it also
stops anything it started — an orphaned ltc-capture would otherwise hold an
audio device indefinitely — and signals, a logout and Quit all stop it first.
- Menu bar: ProdDash's own 2x2 tile mark, its accent tile lit green while
serving, amber mid-move, red when something needs a person. Start, stop,
restart, open the dashboard or admin, open at login, quit.
- Window: status and both addresses, plus Log (also written to
~/Library/Logs/ProdDash, kept a week), Settings, and Permissions.
- The port is written to the data directory's proddash.json — ProdDash's own
file — so the port set here holds however it is next started. Something
already on the port is reported, and offered as a takeover when it answers
like a ProdDash started another way.
- Modules declare what they need in their manifests ("permissions": [...],
documented in docs/MODULE-GUIDE.md and ignored by the shell); the launcher
lists exactly that, with each module's reason, asks for the microphone,
and — since macOS reports nothing about local network access — offers a
check that browses for services and tries every configured upstream.
Timers, ProdCom and ProPresenter modules declare their permissions (patch
bumps). Build with launcher/macos/build.sh; output is gitignored.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
