Split the single ext-snippet column into Alias / URL / Routing /
Last-updated, with the alias surfaced next to a colored type tag,
the URL ellipsized with a tooltip + open-in-new-tab, and the
ext:file.dat:tag snippet click-to-copy via ClipboardManager.
Switch Last-updated to a relative time ("2 hours ago") with the
absolute timestamp on hover, add a friendly empty state, and show
a result toast when "Update All" finishes with partial failures.
customGeoEmpty translated for all 13 locales.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The panel polls api.github.com on every page load. When the host has no
internet (DNS fails, GitHub blocked, etc.) jsonMsg's auto-WARN logging
floods the log with the same error every poll.
Bypass jsonMsg for getPanelUpdateInfo: log the error at Debug level and
return Success:false with the existing localized message so the frontend
popover behavior is unchanged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Backend:
- check HTTP status on every Cloudflare API call so error bodies don't
get parsed as success
- replace unchecked type assertions with comma-ok form (no more panics
when Cloudflare returns an error response)
- return real errors when license/id/token fields are missing instead
of swallowing the failure
- guard SetWarpLicense against an empty errors array
- 15s timeout on the shared http.Client
- build all request bodies and persisted state with json.Marshal
- bump API path to v0a4005 and CF-Client-Version to a-6.30-3596 to
match the current Cloudflare WARP client
Frontend (warp_modal.html):
- remove stray </a-form-item> closing tag
- declare config/peer with const and null-check before dereferencing
- guard addOutbound/resetOutbound against missing warpOutbound
- rename getResolved -> getReserved (the array it builds is "reserved")
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Try six IPv4 providers in turn, accept only HTTP 200 + IPv4-shaped body,
and prompt the user to enter their IP if every provider fails.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Raise module Go version to 1.26.3 and upgrade dependencies including github.com/valyala/fasthttp to v1.71.0 and google.golang.org/genproto/googleapis/rpc to a newer revision. go.sum was updated by module tooling to reflect these changes.
Move per-connection lifecycle out of the controller and into a new
service.WebSocketService. The controller is now HTTP-layer only:
authenticate, validate origin, upgrade, and hand the connection off.
- web/service/websocket.go (new): owns the read/write pumps, hub
registration, and connection lifetime. Pump constants are prefixed
(wsWriteWait, wsPongWait, wsPingPeriod, wsClientReadLimit) to avoid
collisions in the larger service package namespace.
- web/controller/websocket.go: trimmed to the upgrader, same-origin
check, auth gate, and hand-off to the service.
- web/web.go: wires controller.NewWebSocketController(service.NewWebSocketService(hub)).
The hub package (web/websocket) stays as low-level fan-out
infrastructure. Behavior is unchanged — this is a structural cleanup
to align with the rest of the codebase's controller/service split.
Also includes a small range-int modernization in login_limiter_test.go
that gopls flagged.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Two subtle race conditions in the browser WebSocket client:
1. Stale-event clobber. When connect() is called while the old socket is
in CLOSING state, the readyState guard falls through and a new socket
is assigned to this.ws. The old socket's queued close event then
nulls out this.ws, silently breaking send() until the next reconnect.
Same risk for delayed open/error/message handlers.
2. Reconnect-after-disconnect. clearTimeout() does not cancel a callback
that has already fired but whose macrotask has not yet run. If
disconnect() lands in that window, the queued reconnect callback
still calls #openSocket() and resurrects the connection.
Every event handler now bails out if this.ws no longer points at the
socket that fired the event, and the reconnect timer callback re-checks
shouldReconnect before opening a new socket.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* Implement CSRF protection and security hardening across the application
- Added CSRF token handling in axios requests and HTML templates.
- Introduced CSRF middleware to validate tokens for unsafe HTTP methods.
- Implemented login limiter to prevent brute-force attacks.
- Enhanced security headers in middleware for improved response security.
- Updated login notification to include safe metadata without passwords.
- Added tests for CSRF middleware and login limiter functionality.
* fix
Audit panel xhttp config against xray-core's runtime paths and split
fields per direction so each side carries only what it actually uses:
- Bidirectional (must match): host, path, mode, all xPadding*,
session*/seq*, uplinkData*/Key, scMaxEachPostBytes
- Server-only (inbound): noSSEHeader, scMaxBufferedPosts,
scStreamUpServerSecs, serverMaxHeaderBytes
- Client-only (outbound): uplinkHTTPMethod, uplinkChunkSize,
noGRPCHeader, scMinPostsIntervalMs, xmux
The inbound previously held client-only fields and the outbound was
missing every must-match field beyond host/path/mode — meaning a
panel-built outbound couldn't connect to an inbound with a custom
xPaddingKey/sessionKey/etc.
Headers stay on the inbound for URL-share purposes only; xray's
listener ignores them at runtime, but they travel through the share
link's `extra` blob so the client picks them up.
Renames the URL helpers (applyXhttpPadding* -> applyXhttpExtra*) since
the blob now carries more than padding, and folds path/host/mode into
the helper so each link generator's xhttp branch is one line.
Adds two enforcement points for xray's "uplinkHTTPMethod=GET only in
packet-up" rule: the GET option is disabled when mode != packet-up,
and a watcher on the outbound modal auto-clears GET when the user
switches modes.
Hides the XMUX block behind an `enableXmux` switch on the outbound
form (mirrors the QUIC Params toggle) so the section doesn't clutter
the form by default; fromJson auto-flips it on for outbounds with
saved xmux config.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
selectedAuth was UI-only metadata (Xray never reads it) and entirely
redundant with the encryption string itself — the dropdown only
controlled which block from `xray vlessenc` to apply. Replace it with
two explicit buttons ("X25519" and "ML-KEM-768") so the user picks
the auth mode in one click instead of dropdown + Get-New-Keys.
- VLESSSettings drops the field from constructor, fromJson, and toJson;
legacy `selectedAuth` values still in DB will be silently shed on the
next save.
- getNewVlessEnc(authLabel) now takes the label as a parameter; clear
resets only decryption/encryption.
- Fallbacks visibility now keys on encryption === "none" (the same
thing the dropdown was effectively gating on).
- Info modal drops the redundant Authentication tag and colours the
encryption tag red when it's "none", green otherwise.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
testseed is only meaningful for the exact xtls-rprx-vision flow, but the
panel was emitting it for any non-empty flow (including the UDP variant)
and keeping it on the inbound after the flow was cleared via the client
modal. Tighten the gate end-to-end:
- VLESSSettings.toJson (inbound + outbound) now only emits testseed when
the flow is exactly xtls-rprx-vision and the array is 4 positive ints;
default state is empty so unmodified inbounds omit the field entirely.
- canEnableVisionSeed drops the udp443 variant per spec.
- Form adds a tooltip + theme-aware help text and an inline error when
the user partially fills the four inputs; submit is blocked in that
state. Reset clears to empty (= use server defaults).
- UpdateInboundClient strips a now-orphaned testseed when the spliced
client no longer leaves any XRV flow in the inbound.
- MigrationRequirements cleans up legacy rows where testseed lingered
after flow changes or was saved for non-XRV flows by older versions.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- DockerEntrypoint.sh: create jail.d/filter.d/action.d config files
before starting fail2ban so Docker containers no longer start with
0 active jails (fixes#4134)
- x-ui.sh create_iplimit_jails: lower maxretry from 2 to 1 so
fail2ban bans on the first log entry; with maxretry=2 and the
partitionLiveIps logic the second occurrence could arrive after the
32 s findtime window, silently preventing any ban (fixes#4163)
- x-ui.sh: fix datepattern (%%Y -> %Y) so fail2ban parses the Go
log timestamp correctly instead of looking for a literal %%Y string
- x-ui.sh / DockerEntrypoint.sh: fix date command in actionban /
actionunban echo (%%Y -> %Y) so the ban log records actual dates
- check_client_ip_job.go: replace log.SetOutput / log.SetFlags on
the global standard-library logger with a local log.New instance,
eliminating the dangling closed-file-handle between calls and
stopping unrelated stdlib log output from polluting 3xipl.log
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Switch net.IOCounters to per-interface mode and aggregate traffic while excluding loopback and common virtual/tunnel interfaces. Adds isVirtualInterface helper to filter interfaces by exact names and prefixes (docker, veth, tun, wg, tailscale, etc.), sums BytesSent/BytesRecv across valid interfaces, and assigns the totals to status.NetTraffic. Removes the previous warning branch when no counters were found and preserves NetIO rate calculations using lastStatus.
the panel rejected configurations like vless reality on tcp/443 and
hysteria2 on udp/443 even though those are independent sockets in
linux. the old checkPortExist looked only at port + listen.
inboundTransports now classifies each inbound by L4 transport:
hysteria/hysteria2/wireguard are udp; streamSettings.network=kcp is
udp; shadowsocks reads settings.network ("tcp"/"udp"/"tcp,udp");
mixed (socks/http) adds udp when settings.udp is true; everything
else is tcp. checkPortConflict pulls every row on the same port and
only flags a conflict when transport masks overlap. the listen-
overlap rule (specific addr vs any-addr on the same port) is kept.
inbounds.tag has a unique DB constraint and the controller derives
tags from port ("inbound-443"). without disambiguation a second
inbound on the same port would still hit a unique-constraint error.
generateInboundTag keeps the historical "inbound-<port>" shape when
the base tag is free, so existing routing rules survive the upgrade
unchanged, and appends "-tcp"/"-udp" only when the base is already
taken.
closes#4103.
In GetXrayVersions, explicitly ignore the tag "26.5.3" and raise the minimum accepted Xray release from 26.3.10 to 26.4.25. This excludes a specific problematic release and updates the version parsing logic to only include >26 or 26.4.25+ releases.
Swap tr-table-rt and tr-table-lt on the size and totalGB elements in aClientTable.html so the size display and the total GB display are positioned correctly (size on the left, total on the right). This is a UI alignment fix with no functional logic changes.
Track and surface a subscription's enabled state from backend to frontend so the UI can show inactive subscriptions and use it in active-state logic.
Changes:
- sub/subService.go: track hasEnabledClient, set traffic.Enable, add Enabled to PageData and populate it in BuildPageData.
- sub/subController.go: include enabled in the page context.
- web/html/settings/panel/subscription/subpage.html: emit data-enabled attribute and render an "inactive" tag when disabled.
- web/assets/js/subscription.js: read data-enabled and include it in isActive() checks.
This ensures subscriptions with no enabled clients are marked inactive in the UI and excluded from being considered active.
Add a configurable option to restart Xray when clients are auto-disabled and persist disable actions.
Changes include:
- New setting restartXrayOnClientDisable (default true), getters/setters in SettingService, UI toggle in general settings, and translations for multiple locales.
- AddTraffic signature updated to return a third bool (clientsDisabled). disableInvalidClients now calls Xray API to remove users, marks client_traffics.enable=false, updates inbound.Settings JSON so clients appear disabled in stored settings, and returns appropriate counts/errors.
- XrayTrafficJob now checks the clientsDisabled flag and restarts Xray when the setting is enabled (with fallback to mark Xray as needing restart on failure).
- XrayService.GetXrayConfig call adjusted to ignore AddTraffic returns.
- Subscription generation (subService/subJson/subClash) no longer filters clients by their enable flag when matching subId.
- Minor fixes: check_client_ip_job now checks scanner.Err and improved API error handling/logging.
These changes ensure auto-disabled clients are propagated to Xray and the stored inbound settings, and provide an option to restart Xray automatically after auto-disable events.
#4142
Adjust QUIC parameter defaults and tighten form validation across inbound/outbound components.
- Set default brutalUp/brutalDown to 65537 and only include them in JSON when congestion is 'brutal' or 'force-brutal'.
- Change keepAlivePeriod defaults (inbound QUIC -> 5s, Hysteria stream -> 2s) and enforce minimums in the UI.
- Expose and serialize additional QUIC fields in outbound QuicParams: init/max stream windows, init/max connection windows, maxIdleTimeout, disablePathMTUDiscovery, maxIncomingStreams.
- Add UI min/placeholder constraints: stream/connection receive windows min=16384 and updated placeholders to show defaults, brutal fields min=65537, maxIncomingStreams min=8 (placeholders updated), keepAlive min adjusted.
- Add Wireguard and Hysteria entries to Protocols.
Touched files: web/assets/js/model/inbound.js, web/assets/js/model/outbound.js, web/html/form/outbound.html, web/html/form/stream/stream_finalmask.html.
Add a check for scanner.Err() after scanning log lines and return nil if an error occurred. This prevents further processing of potentially incomplete or invalid log entries when the scanner encountered an error.
when the admin adds a custom outbound (eg vless cascade to a second
server) and a routing rule sending all inbound traffic to it, that
catch-all gets evaluated before the existing api->api rule, so the
panel's internal stats inbound's traffic ends up on the cascade
outbound. the grpc stats query then can't see anything, GetTraffic
returns no inbound/user counters, and every client appears offline
with zero traffic even though the actual proxy path works fine.
before save, find the api rule and move it to the front of
routing.rules. if it's missing entirely, insert a default. other
rules keep their relative order.
closes#4113. probably also fixes the long-standing #2818 where the
documented workaround was "manually move the api rule to the top".
