'ccs config' starts the web-server on host 'localhost', which macOS
resolves to ::1, so a reuse probe limited to 127.0.0.1 missed the most
common already-running server and launch started a redundant second
instance. Each candidate port is now probed on 127.0.0.1 first and then
[::1]; an IPv6 hit writes the bracketed-literal baseUrl into bar.json,
which URLSession in the Swift app resolves correctly.
'ccs bar' always tried to start its own web-server. With a CCS server
already listening on 127.0.0.1:3000, get-port (probing the unspecified
address, which macOS allows to bind alongside a specific loopback
listener) reported 3000 as free, and the subsequent 127.0.0.1 bind
failed with EADDRINUSE, aborting launch.
Launch now probes candidate ports (bar.json port first, then the
default list) with a short-timeout GET /api/bar/summary and reuses the
first live CCS server it finds; only when none responds does it start a
new server. The get-port probe also passes host 127.0.0.1 so the
availability check matches the actual bind target. Probe failures are
treated as no-server-found and never break launch.
Closes#1500
The floating ccs-bar-latest release tag carries no version, so deriving the
app version from tag_name printed 'vccs-bar-latest' and pinned that literal
string. The version now comes from CFBundleShortVersionString in the
extracted bundle's Info.plist; an unreadable plist skips pinning and clears
any stale pin.
The post-install check compared app semver major against the CCS server
major, but the app is versioned independently of the CLI, so the mismatch
warning fired on every install. It is now a capability handshake against
GET /api/bar/summary: 200 confirms the server serves the bar API, 404 warns
the server predates CCS Bar, and anything else keeps the soft warning.
Install never hard-fails on the handshake.
Closes#1497
--help and -h are honored anywhere in the args so 'ccs bar install --help'
shows help instead of running the installer. Also wires the 'ccs help bar'
topic, shell-completion flag suggestions, and a more-help table entry.
The prior fix only corrected PROVIDERS_WITHOUT_EMAIL, but the mock still stubbed
validateNickname -> null, hasAccountNameConflict -> false, and registry paths -> '',
which leak to cliproxy-auth-routes (same bun worker) and broke nickname validation
+ registry ops (9 CI failures). Now every mock factory spreads the REAL
account-manager and overrides only the account-DATA reads + IO; afterAll restores
the real account-manager/account-safety modules. tier-lock 8/0, combined ordering
39/0, tsc clean.
bun's mock.module is global and sticky; this test mocked account-manager with
PROVIDERS_WITHOUT_EMAIL: [] at top level and in beforeEach, poisoning the module
registry for later files in the same bun worker. cliproxy-auth-routes imports
PROVIDERS_WITHOUT_EMAIL at load time, got [], and getStartAuthNicknameError's
guard returned null for all providers -> 9 flaky CI failures (ordering-dependent).
Use the real ['kiro','ghcp'] in every mock factory and re-register safe defaults
in afterAll. Verified: worst-case ordering 39/39, full fast bucket 2500/0.
defaultTailLines used Bun.spawn(['tail',...]), but the ccs server (and
'ccs bar launch') runs under node where Bun is undefined — it threw, the
per-path catch swallowed it, and Codex quota silently never appeared in real
deployments (it only worked under the bun-run dev harness). Replace with a
pure fs read: runtime- and OS-agnostic, no subprocess. Verified under node:
surfaces a live rollout's rate_limits (primary/secondary).
startServer was called without a host, defaulting to 'localhost' which
resolves to ::1 (IPv6) on macOS, while bar.json's baseUrl and the menu-bar
app use 127.0.0.1 — so the app could not reach its own server (connection
refused) and showed offline. Pass host 127.0.0.1 explicitly to match.
The menu only re-polled on open or after a mutation, so if the server briefly
dropped rows (e.g. a backend restart mid-session) the dropdown could stay stuck
showing stale/missing rows until reopened. Add a 60s background poll that
reconnects if the client was lost and reloads non-force (respecting server-side
caches, so it never hammers providers). As a bonus, alerts now evaluate every
interval instead of only on menu-open, making them genuinely proactive.
Add a 'Get CCS Bar' promo banner + card to the dashboard (mirroring the
OpenRouter promo pattern + design system) so users discover the macOS menu-bar
app, with a macOS-aware install CTA. Add a user-facing docs/ccs-bar.md covering
what it is, install via 'ccs bar install', launch, what it shows, uninstall,
and the loopback requirement -- closing the docs-sync gap for the new ccs bar
command.
Gate /api/bar/* behind the localhost-when-auth-disabled guard (single DRY
choke point) so native quota/tier/cost can't leak on a non-loopback bind with
auth disabled; add a guard test. Delete the dishonest maxRedirections test that
asserted the opposite of the production redirect hardening. Key per-account
today-cost on the local day (matching analytics) instead of UTC. Stop the
inner 429 retry in the Claude usage fetch so the outer cache + circuit breaker
honor Retry-After. Narrow the usage-transformer map type and fix stale
doc-comments; clarify the one-alert-per-reset-window quota rule; gitignore the
local demo scaffolding.
Move the bars/line choice out of Settings and into the Spend section header's
blank space as a small inline toggle, so the user flips the chart style right
where the chart is. Persistence unchanged.
Narrow the dropdown (380 to 360) and give it more vertical room (620 to 700).
Tighten the subscription cards (less spacing/padding) since they carry only a
few window rows. Make the spend chart taller (18 to 30) and let the user pick
its style in Settings: chunk bars (default) or a line graph with a subtle area
fill. Persist the choice in UserDefaults.
Move Settings out of the menu-bar popover into a standalone resizable AppKit
window so clicking it no longer steals focus and dismisses the bar (the .sheet
inside MenuBarExtra(.window) was the root cause). Force the actual NSWindow
appearance (aqua/darkAqua/system) on both the popover and the settings window
so Light/Dark visibly flips materials and semantic colors, not just custom
tokens. Replace the fragile popover quit dialog with an inline two-step
arm/confirm. Widen the dropdown and increase spacing/type for readability. Fill
the settings window responsively and make Done close it via the window
controller (dismiss() is a no-op in a hosted NSWindow).
Rename the footer 'Alerts' button to 'Settings' (gear icon) since it opens all
preferences including the appearance/theme picker, which was undiscoverable
behind an alerts label. Add a confirmation to the Quit button so a stray click
no longer closes the whole menu-bar app with no easy way back.
Add a theme token system (CCSBarCore/BarTheme.swift): a BarAppearance enum
(system/light/dark, default dark to preserve the current look), light+dark
token palettes, a colorScheme-driven resolver, and a SwiftUI EnvironmentKey.
The root forces .preferredColorScheme so Light renders light even under a dark
macOS, and paints an explicit light window surface. Thread the tokens (accent,
headroom bands, subscription, card surface, track) through every view so the
whole dropdown themes consistently, with chip text blending toward black on
light. Add an appearance picker in Preferences, persisted in UserDefaults.
Dark accent/subscription values are locked byte-identical and harness-guarded;
pool-account rows now share the muted band palette for cross-section
consistency.
Lift chip text toward white so small tier badges (max/pro) read clearly on the
dark surface. Rename the icon toggle from Color/Mono to 'Icon' (it only swaps
the menu-bar icon, not the bar theme) to remove the misunderstanding. Make the
Dashboard button actually work when the server is down: probe reachability and,
if not up, launch 'ccs config' (detached) which boots the web-server and opens
the dashboard itself; degrade gracefully when the ccs binary isn't found.
Replace the raw system green/yellow/orange/red headroom colors (garish on the
dark surface and colliding with the brand orange) with a muted, intuitive
green-amber-coral-red palette that leans coral for 'low' so a near-empty
window never reads as the accent. Drop the cool indigo card wash for a neutral
elevated surface so the warm bars no longer clash; subscription identity stays
on the section header and badge.
Replace the prose window lines with an aligned per-window bar list (5h /
weekly / Opus / Sonnet), colored by headroom, with the binding window
highlighted and a compact reset chip. Humanize durations beyond a day
(Nd Nh instead of raw hours) and only show the burn-rate pace clause when a
window is projected to exhaust before it resets, so a meaningless 'resets in
44h' / '~110h left' projection no longer appears.
Lead the dropdown with subscriptions: a dedicated card shows the binding
window as the hero (gauge, remaining %, reset countdown, and a burn-rate
'left at this pace' line), a secondary window, and the Opus/Sonnet split for
Max, with a stale marker for older Codex data. Demote spend/usage to a compact
strip and surface cross-tool headroom. Add pure burn-rate / time-to-exhaustion
and binding-window math in Core. Kill the menu scroll indicator for real
(NSScrollView hider) so content stays aligned. Spend-cap alerts default off
(pay-per-use, opt-in).
Expose per-window detail (5h / weekly / Opus / Sonnet) on native subscription
rows instead of collapsing to a single percentage, so the bar can show the
binding window, resets, and a burn-rate pace line. Fix the Codex collector to
scan recent session rollouts newest-first for the latest non-null rate_limits
(today's exec-mode session is null) and mark the result stale with its source
time, so Codex quota appears instead of silently vanishing.
Show Claude Code and Codex subscription rows with the Tier 1 quota gauge and
reset countdown, visually distinguished from CLIProxy provider accounts, and
let them drive the existing quota alerts. Formatting + harness coverage for the
native rows.
Surface the logged-in Claude Code and Codex subscription quota as first-class
summary rows so the existing gauge and alert engine render them with no new UI.
Claude Code: read the native token (~/.claude/.credentials.json, macOS Keychain
fallback) and reuse the existing Anthropic usage fetch+normalize via a new
token-fed entry point. The /oauth/usage endpoint is hostile to polling, so the
fetch is server-side only behind a 10-minute on-demand cache, in-flight
coalescing, Retry-After + exponential backoff with jitter, a 3-strike circuit
breaker with a 15-minute cooldown, and serve-stale-on-failure; logged-out or
unsupported subscriptions never spend a token call. Codex: zero-network read of
the latest rate_limits from local session rollouts, omitted when absent.
Native rows side-load bounded so a slow or failed fetch degrades to CLIProxy
rows, never an error.
Add real quota gauges with threshold bands and a reset countdown for
quota-capable accounts. Introduce a pure, deterministic alert rule engine
(CCSBarCore) covering quota-remaining, daily/monthly spend caps, reauth, and
cooldown, with per-rule dedupe keys that re-arm on period rollover or
clears-then-recurs so it never spams. Deliver via UNUserNotificationCenter
with an in-dropdown Alerts fallback when notifications are denied. Add a
preferences surface (per-rule toggles, caps, quota levels, glance mode)
persisted in UserDefaults, and a configurable menu-bar glance
(auto/today/month/lowest-quota/account-count) that still never shows a
lifetime dollar. Default spend caps raised to $500/day and $10000/month.
monthToDate (calendar 1st-of-month to now, local) on BarAnalytics in both
compute paths, kept distinct from the rolling last30d so a fresh month resets
toward zero. Feeds the monthly-spend alert and the month-spend glance without
a rolling-window false breach.
Mirror the richer analytics payload (quotaStatus, per-account last-active,
bySurface) in the Codable models. Add a per-surface usage section (Claude
Code, Codex, Droid, CLIProxy) with proportional bars beside the existing
spend grid, sparkline, and top models. Rebuild the menu-bar title chain so a
lifetime dollar can never sit in the always-on title (it read as live spend):
quota% then today spend then account count. Rebuild the accounts rows
(default badge, tri-state quota, honest no-data cost, health), pivot to an
honest idle hero when no recent data, hide the scroll indicator, and tighten
spacing.
Derive a tri-state quotaStatus (ok|unsupported|error) per account so a
provider with no quota API renders as 'no quota' with a healthy dot instead
of an alarming bare dash, and treat unsupported as healthy in deriveHealth.
Switch the analytics endpoint off the CLIProxy snapshot, which freezes
whenever the proxy restarts (usage is in-memory only), and onto the merged
daily/hourly usage the dashboard uses. Recent activity from Claude Code,
Codex, and Droid now shows, and a per-surface breakdown answers where usage
goes. Add lastActivityAt, daysSinceLastActivity, hasRecentData and a 30-day
series; per-account today_cost reads null (unknown) rather than a misleading 0.
Summary aggregator could block indefinitely: it ran the full system health
audit (a synchronous execSync) on every request and had no per-account or
request-level timeout. Now:
- per-account fetch is bounded; whole response is raced against a deadline
- health is derived per-account from each quota result (no blocking audit)
- stale-while-revalidate keeps the last value when a refresh is slow/fails
Adds GET /api/bar/analytics plus a pure, tested aggregator rolling up
today/7d/30d/all-time spend, a 7-day sparkline, and top models.
Tam Nhu Tran