2203 Commits
Author SHA1 Message Date
Markson Hon 5079fc75b1 Merge pull request #2033 from Zakkaus/leftovers-1 v2.5.4 2026-09-21 10:48:33 +08:00
Markson Hon 1f2326eeb3 Merge pull request #2032 from Zakkaus/bound-device-polish 2026-09-21 10:47:55 +08:00
Zakk c14e68e7b4 fix(core): register the tuic outbound in distro/all like anytls and juicity 2026-09-21 12:44:14 +10:00
Zakk a9abbe0fe6 feat: the Multiplex setting applies to VMess outbounds
Settings had a Multiplex switch and a concurrency field, validated and stored,
that no template code read. VMess outbounds get mux.enabled and the
concurrency when the switch is on, which is what the help text promised
("vmess only"). The core accepts the result (run -test).
2026-09-21 12:44:14 +10:00
Zakk bc33a2bc6f fix(gui): the DNS dialog opens the in-app docs; the node backend selector goes
The DNS help button linked v2fly's documentation of a different DNS
implementation; the docs button opens the transparent proxy page, whose DNS
section describes the core's module. The SS and Trojan forms offered a
"backend" of "system default" or "v2ray / xray", both the same core since
2.4.6; the link parameter v2raya-backend is still parsed and re-emitted.
2026-09-21 12:44:14 +10:00
Zakk ba24a94719 fix: names and descriptions after the bundled core
Strings that still named v2ray-core, "v2ray / xray" or a wiki: six locales
(the core is v2raya_core; the invalid-service message, the SS plugin help,
the transparent-proxy hint, the TUN example, the custom inbound tag hint),
the --v2ray-bin, --core-startup-timeout and --plugin-manager descriptions
and their copies in the OpenRC confd files (the plugin manager is called
with --stage=parse|configuration, --link and --v2raya-confdir, not with
life-cycle stages), two comments that forbade or excused the current design,
the package descriptions that listed SSR, the AUR Go minimum (go.mod says
1.26), the systemd Documentation= and the installer's shortcut that pointed
at the wiki and the docs site, and the README sentence on network requests,
which left out the weekly release check and the NTP query on VMess failure.
2026-09-21 12:44:14 +10:00
Zakk 0872d7db62 chore: remove what nothing uses any more
copyfile, the empty ports/ss.go stub, parseDnsAddr and DnsRouting (the DNS
config is generated for the core's module now), the Inter font and the
switch-menu icon of the old GUI, the docs-site dispatch workflow (the site is
retired; the docs are in the app), the unused v2ray_arch in the docker helper,
the root .gitignore rule for a web directory that moved, and the .dockerignore
patterns for binary and test names that no longer exist. The skipped
TestFillEmpty asserts what FillEmpty does instead of what it once did.
2026-09-21 12:44:14 +10:00
Zakk 35505baed5 fix: the bound-device option needs Linux 5.14 (SO_NETNS_COOKIE), and the docs say it is off by default 2026-09-21 12:28:42 +10:00
Zakk e84ab8a790 docs(gui): the bound-device option in the transparent proxy page; the geosite date label through i18n
The six locales' transparent-proxy pages explain --redirect-respect-bound-device
after the excluded prefixes; the flag table on the parameters page already lists
it. The "(geosite)" suffix on the local GFWList date is a locale string.
2026-09-21 12:23:51 +10:00
Zakk 01df00559b fix(service): the bound-device flag sits with the transparent-proxy options and names its kernel requirement
--redirect-respect-bound-device was the first field of Params, so it led
--help, --report config and the /etc/default/v2raya the packages generate
from it; it now follows --ipv6-support. Its description says what it does
in the user's terms (SO_BINDTODEVICE, NetworkManager's connectivity
checks) and what it needs (kernel 5.7, cgroup v2). When SO_NETNS_COOKIE
is missing the error says the kernel is too old instead of ENOPROTOOPT.
cilium/ebpf moves to the direct require block, as go mod tidy puts it.
2026-09-21 12:23:51 +10:00
Markson Hon 72465a8873 Merge pull request #1954 from Laceprndpm/fix/redirect-bound-device
feat(redirect): add opt-in BPF bypass for device-bound TCP sockets
v2.5.3
2026-09-21 08:29:10 +08:00
Markson Hon 849a21623b Merge pull request #2030 from RocketMaDev/fix-geosite
feat: display geosite date if gfwlist absent
2026-09-21 07:59:01 +08:00
Markson Hon 7a13577984 Merge pull request #2029 from Zakkaus/drawer-fold
feat(gui): the drawer folds to the rail
2026-09-21 07:56:07 +08:00
Markson Hon b9f97d7f4d Merge branch 'main' into drawer-fold
Signed-off-by: Markson Hon <50002150+MarksonHon@users.noreply.github.com>
2026-09-21 07:43:57 +08:00
Markson Hon 72c40647a8 Merge pull request #2027 from Zakkaus/eggs
feat(gui): two easter eggs on the logo: the M3 expressive shapes, and the mirrored layout
2026-09-21 07:42:24 +08:00
RocketDev 0104fa8405 feat: display geosite date if gfwlist absent
Currently, if system /usr/share/v2raya/geosite.dat is used, v2raya will
not post error, but will display "none" for local GFWList date.
Since v2rayA reads this compatible file, we should display a date for
it, so that user is not confused.

Assisted-by: GPT-5.6-terra
2026-09-21 03:29:57 +08:00
Zakk 70f2983a25 fix(gui): the logo fills 72% of its frame, so the rail's is 40 dp again and the drawer's 35 2026-09-21 02:54:43 +10:00
Zakk 0f3d97d14a fix(gui): the drawer's logo is 36 dp, not 28 2026-09-21 02:47:06 +10:00
Zakk 815da9f431 fix(gui): the fold buttons take the version's outline colour 2026-09-21 02:43:24 +10:00
Zakk a8aebd89d1 feat(gui): the fold and unfold buttons carry the burger with the arrow, pointing the way the drawer goes 2026-09-21 02:27:08 +10:00
Zakk 61da17fde2 feat(gui): the fold button sits by the version, the unfold button at the rail's foot 2026-09-21 02:23:33 +10:00
Zakk 6c481da971 feat(gui): the drawer folds to the rail
A menu button by the brand folds the standard drawer; the rail that
replaces it is the one the medium window already uses, with the same
menu button at the top to unfold. The choice is kept in localStorage.
2026-09-21 02:20:12 +10:00
Zakk ad1d3001bc feat(gui): the mirror comes with the sixth tap, when the frame is gone again 2026-09-21 02:04:30 +10:00
Zakk 833d5655c6 feat(gui): every fifth tap on the logo mirrors the layout, as the RTL locales see it 2026-09-21 02:01:58 +10:00
Zakk 94764684da feat(gui): the logo's frame takes the next M3 expressive shape on every tap 2026-09-21 01:35:27 +10:00
Markson Hon 5be2e88395 Merge pull request #2026 from Zakkaus/seed-picker 2026-09-20 23:25:01 +08:00
Markson Hon dd9c3248d1 Merge pull request #2024 from Zakkaus/unsaved-guard 2026-09-20 23:18:56 +08:00
Markson Hon 53f22a1344 Merge pull request #2023 from Zakkaus/docs-locales 2026-09-20 23:18:15 +08:00
Zakk 58e6faaf72 refactor(gui): the custom seed is picked with Vuetify's colour picker
The theme panel had its own hue slider and hex field for a custom seed.
v-color-picker in hex mode gives the canvas, the hue slider and the hex
input, so the slider, the hex parsing and their styles go.
2026-09-21 01:17:09 +10:00
Zakk c823aaf777 feat(gui): the browser asks before a reload drops unsaved settings or RoutingA edits
useUnsavedGuard registers beforeunload while a component with unsaved edits is mounted.
The settings model keeps the snapshot it loaded or saved, so the page knows when the
form differs; the RoutingA editor already compares its text with the original.
2026-09-21 01:06:19 +10:00
Zakk 1624cd680f fix(gui): inline code in the docs is an inline-block, so Firefox does not overlap it with RTL text 2026-09-21 00:57:06 +10:00
Zakk ccb3219eb2 docs(gui): the in-app documentation in Persian, Korean, Portuguese and Russian
Six pages each under gui/src/docs/{fa,ko,pt,ru}/, named after the i18n locale keys so
docs/index.ts picks them up. Headings, code blocks and inline code are identical to the
English pages; bold UI labels are the locale's own strings; log messages stay in English.

Inline code in the docs body is now an LTR isolate that does not wrap, and code blocks
are LTR, so paths and flags keep their order in the Persian pages.
2026-09-21 00:50:32 +10:00
Markson Hon f6f04c08d1 Merge pull request #2022 from Zakkaus/stale-issues
ci: close issues with no activity for 60 days
2026-09-20 22:17:22 +08:00
Zakk ca089c3391 ci: close issues with no activity for 60 days
A daily job lists open issues not updated for 60 days, except those labeled
notice or enhancement, and closes up to 100 of them as not planned with a
comment; there are 220 such issues today, the oldest from 2024.
2026-09-21 00:08:30 +10:00
Markson Hon d71e29a998 Merge pull request #2021 from Zakkaus/dns-bootstrap-rules-2
fix(dns): the service, the module's bootstrap and the hijacked resolv.conf follow the DNS rules' direct upstreams
v2.5.2
2026-09-20 21:31:24 +08:00
Zakk 74277851bb fix(dns): the service, the module's bootstrap and the hijacked resolv.conf follow the DNS rules' direct upstreams
Node hostnames, upstream hostnames and the service's own lookups (subscriptions,
latency tests) were resolved by hard-coded public resolvers (223.5.5.5, 119.29.29.29,
...) whatever the DNS rules said; the second nameserver written into the hijacked
/etc/resolv.conf was hard-coded too. directDnsServers() now collects the plain
udp/tcp IP upstreams of the rules with the direct outbound, and

- resolv.PreferredServers hands them to the service's resolver before its public list
- the module's bootstrap_dns is the system's resolvers plus those upstreams
- the upstream for node hostnames is the first of them
- HijackResolv writes the first port-53 one as the second nameserver, and so does the
  restore when no backup exists

getSystemDnsServers reads the backup while resolv.conf is hijacked, so a restart
no longer hands the module its own listener as bootstrap.
2026-09-20 23:23:08 +10:00
Markson Hon 900e7e26b6 Merge pull request #2018 from Zakkaus/dns-upstream-scheme
feat(dns): DNS over HTTPS upstreams, direct and through the proxy; DoT keeps its server name; DoQ refused on save
2026-09-20 19:58:24 +08:00
Markson Hon 005d9999ac Merge pull request #2017 from Zakkaus/routinga-templates-menu
feat(gui): RoutingA templates in their own menu, seventeen of them; the row and dialog say what RoutingA is
2026-09-20 19:58:02 +08:00
Markson Hon 90b0087229 Merge pull request #2015 from Zakkaus/openrc-polish
fix(install): OpenRC runs v2rayA under supervise-daemon; a new user's ~/.config is created; the ignored flag goes
2026-09-20 19:57:37 +08:00
Zakk cc60098633 feat(dns): DNS over HTTPS upstreams, direct and through the proxy; DNS over TLS keeps its server name
The DNS module spoke UDP, TCP and DNS over TLS. An https:// upstream is
now an RFC 8484 POST through an http.Client the upstream keeps between
queries; the TLS handshake is done in the dialer so the server name is
the upstream's hostname while the address dialled is what bootstrap
resolved for it — the module cannot ask itself for its own upstream's
address. A direct upstream dials on a marked socket, as the other
protocols do; a proxied one dials through the xray dispatcher, so the
request leaves through the chosen outbound.

Two things the dispatcher path had wrong for tls:// upstreams: it sent
plain DNS over TCP whatever the protocol, and bootstrap replaced the
hostname by an IP without keeping it, so TLS had nothing to verify
against. The hostname is kept as ServerName on the upstream, the
dispatcher path wraps the connection in TLS for tcp-tls, and the direct
miekg client verifies the same name. The service adds https and tls
hostnames to the bootstrap list, which it skipped for every scheme
address before. CheckDnsUpstream accepts https:// URLs with a host and
still refuses quic://.

Verified on the VM with a raw UDP client against the module, names not
in its cache: https://dns.google/dns-query direct (bootstrap dns.google
→ 8.8.8.8, a v2raya_core connection to 8.8.8.8:443), https://1.1.1.1
/dns-query via proxy (the connection goes to the proxy server's :443),
tls://dns.google via proxy (rewritten to 8.8.8.8:853, TLS verified as
dns.google) and tls://1.1.1.1 direct all answer; the previous rules
restore unchanged. A unit test runs the DoH client against an httptest
TLS server.
2026-09-20 18:53:48 +10:00
Zakk c1e65ca44a fix(dns): refuse DoH and DoQ upstreams on save instead of failing every query
The DNS form's placeholder offered https://dns.google/dns-query, the
service turned https:// into protocol "https" and quic:// into "quic",
and the DNS module handed those to a miekg/dns client, which speaks udp,
tcp and tcp-tls only. Saving succeeded and every query on that rule then
failed, visible only in the log.

CheckDnsUpstream in kernel/v2ray accepts an address, tcp://, udp:// and
tls:// (DNS over TLS, which the module does support) and rejects https://
and quic:// with a message that names the alternatives; PutDnsRules
returns it as a 400 before anything is stored. The placeholder in the six
locales shows the supported forms and the docs say which schemes work.

Verified on the VM: https:// and quic:// rules come back BAD_REQUEST with
the message, tls://1.1.1.1 saves and the module logs
addr=1.1.1.1:853 protocol=tcp-tls; the previous rules restore unchanged.
2026-09-20 18:44:58 +10:00
Markson Hon 3f2d92d459 Merge pull request #2016 from Zakkaus/reinvent-1
refactor: use what the dependencies already provide
2026-09-20 16:43:02 +08:00
Zakk a7afb19ff6 fix(gui): one name and two places for the RoutingA rules; a clean lint
RoutingA could be opened from the dashboard's splitting tile (labelled
"RoutingA"), the proxies page's toolbar ("RoutingA rules"), the settings
row ("RoutingA rules" since the previous commit) and the tutorial. The
label is "RoutingA rules" everywhere it remains, and the proxies page
loses its button: rules are not a property of the node list, and the
dashboard tile that chooses the splitting mode and the settings row are
where a user looks for them.

The lint carried five warnings into every CI run: two unused imports in
AppShell.vue, one in a spec, a let that never changes, and v-html in the
log view, whose input highlight.js escapes.
2026-09-20 17:19:10 +10:00
Zakk e3a5f98b78 feat(gui): RoutingA templates in their own menu, seventeen of them; the row and dialog say what RoutingA is
The templates sat at the end of the syntax reference, behind five
collapsed sections, and the settings row and dialog were titled by the
bare name RoutingA. The dialog header now has a Templates menu next to
the reference button, in two groups: four full rule sets that replace the
rules after a confirmation (China direct; sites outside China through
the proxy; everything through the proxy; only common foreign services
through the proxy) and thirteen additions inserted at the cursor (ads,
streaming, social media, Telegram, AI, developer services, Chinese
endpoints of Apple/Google/Microsoft/Steam, Apple and Microsoft direct,
game platforms, speed tests, LAN, BitTorrent direct, block QUIC). The
reference is a syntax reference again. The settings row and the dialog
are titled "RoutingA rules" in every locale.

Every template uses only v2fly geosite/geoip categories and the three
built-in outbounds; a spec runs each through the editor's checker, and
on the VM each of the seventeen applied through the API while the core
was running.

The app bar's status button carries the spacing to the group button
itself: the class on the OutboundMenu component landed on the menu, not
its activator, so the two sat flush at every width below 1200.
2026-09-20 17:15:44 +10:00
Zakk 833af7dc74 refactor: use what the dependencies already provide
A read of the tree for home-grown code that a dependency or the platform
covers. Each item names what replaces it.

- IPv4 CIDR membership was a binary-string trie (infra/dataStructure/trie)
  fed with CIDRs rendered as bit strings; net.IPNet.Contains, which the
  same file already used for IPv6, does it. The trie package is gone.
- The response replay cache was an LRU with its own linked list, indexes
  and timeout strategy (infra/dataStructure/lru), of which one fixed-size
  instance existed; hashicorp/golang-lru/v2, already in go.mod, is that
  cache, typed. go mod tidy also dropped go-nanoid, which nothing imported.
- The HTTP client for subscription updates through the local proxy carried
  a CONNECT dialer of its own (pkg/util/proxyWithHttp); http.Transport's
  Proxy handles socks5:// and http:// URLs, resolves hostnames on the
  SOCKS proxy and uses CONNECT for https. Verified on the VM: a
  subscription update succeeds in both "pac" and "proxy" modes.
- core/dns/qtype.go listed nine type names in two switches before falling
  back to miekg/dns tables that contain the same names; dns.Type.String
  and dns.StringToType are the tables.
- The GUI tracked prefers-color-scheme with its own media query and
  toggled a theme-dark class nothing reads any more; Vuetify's "system"
  theme follows the OS itself. Verified in headless Chrome: auto follows
  the OS in both directions, an explicit choice overrides it.
- The AUR packages carried .INSTALL hooks whose only content was a
  coloured "Completed!" banner, duplicated in the deb/rpm after-install
  scripts; pacman and apt report that themselves. The hooks are gone, the
  release workflow stops copying them, and after_install/after_upgrade
  keep only daemon-reload and the restart.
- The tap release job downloaded eight artifacts in eight steps;
  actions/download-artifact takes a pattern with merge-multiple.
2026-09-20 16:40:34 +10:00
Markson Hon f2be654fca Merge pull request #2014 from Zakkaus/asset-runtime-dir 2026-09-20 14:36:37 +08:00
Zakk 70d244622e fix(install): OpenRC respawn stays at its defaults, as the systemd unit's start limit does 2026-09-20 16:18:54 +10:00
Zakk 0506a1f448 fix(install): OpenRC runs v2rayA under supervise-daemon; a new user's ~/.config is created; the ignored flag goes
The scripts backgrounded v2raya with start-stop-daemon and a pidfile,
and stop_post ran pkill -x v2raya_core to catch a core that outlived a
killed service, which also took the cores of v2raya-user.<user>
instances. Under supervise-daemon the service sits in its own process
group and cgroup: a stop takes the core with it, and a crash respawns
the service, whose start kills an orphaned core holding the ports
itself. stop_post is gone.

v2raya-user.initd ran checkpath on ~/.config/v2raya, which fails on an
account that has no ~/.config yet ("could not open .config"), so a fresh
user could not start the service; checkpath makes one level at a time,
and ~/.config is created first.

The conf.d defaults passed --log-disable-timestamp together with
--log-file, and the flag is ignored whenever a log file is set; the
default now names the log file alone and says why. The descriptions say
what v2rayA is today, and the --lite help line matches the binary.

Verified on the Gentoo VM (OpenRC 0.63): stop clears the 51 tproxy
rules, the fwmark rule and the core; start restores them; kill -9 of
v2raya is followed by a respawn within seconds, the orphaned core is
killed by the new instance and the dashboard reports it running; a
v2raya-user.tuser instance for a new account starts as that user,
survives a system service stop, and stops cleanly.
2026-09-20 16:15:05 +10:00
Zakk 17e6c48dff fix(service): a user without a session keeps its assets in the configuration directory
The core's asset directory defaulted to v2raya/ under the XDG runtime
directory, and the lookup created the links there with xdg.RuntimeFile.
A service user has no session, so /run/user/<uid> does not exist and the
user cannot create it; the first lookup failed with "could not create any
of the following paths: /run/user/1001/v2raya" and the start died, while
geoip.dat and geosite.dat sat in /usr/share/v2raya. That is every
OpenRC v2raya-user.<user> instance and any lite start outside a desktop
session. When the runtime directory cannot be made, the configuration
directory, which the process owns, holds the links instead.

Verified on the Gentoo VM: /etc/init.d/v2raya-user.tuser starts, reports
lite and links /usr/share/v2raya/*.dat into ~/.config/v2raya.
2026-09-20 16:08:25 +10:00
Markson Hon 49c8a2e168 Merge pull request #2013 from Zakkaus/readme-macos
docs: the Homebrew service runs as root for tun or as the user for the system proxy
2026-09-20 14:04:22 +08:00