serve exits 78 (EX_CONFIG) when the configuration file it loaded is
wrong: format, syntax, keys or values. No file, an unreadable one, the
database and ports still exit 1, so systemd keeps retrying what can
heal. The deb unit and the NixOS module set RestartPreventExitStatus=78.
A parse error gives a position and fixed words, never the file's text.
Updates #1926
settings mirrors headscale's keys one to one, headscale's checks judge the
rest, and the local CLI gets the socket from a variable instead of a file.
headscale always runs as headscale:headscale, which the module creates;
user and group are read-only, so setting either fails evaluation, and stay
only because nixpkgs' headplane module reads them. A settings.cli key
fails evaluation instead of being dropped with a warning, and the unit no
longer gets CAP_CHOWN.
nix/example-configuration.nix goes: the options carry its examples and
defaults.
Updates #1926
The environment alone is a complete configuration and client commands
need a file only for an API key, so the loader stops warning; serve logs
at info level that it runs on defaults and the environment.
Updates #1926
tls.cert_path, tls.key_path and tls.acme.* replace the flat keys, which
still work. Warn on them only when they turn TLS on: NixOS always sets them.
Updates #2527
resolveSingleUser reported every non-single result as "multiple users
match query", including zero matches. An explicit --identifier 0 was
sent to the API, which treats id=0 as no filter, so it listed every
user and failed as ambiguous. Return a not-found error for zero matches,
list the matching users when several match, and reject a non-positive
identifier before calling the API.
A *ListenerBindError that wraps syscall.EADDRINUSE now ends with a
"sudo ss -tlnp 'sport = :PORT'" pointer, and one wrapping
syscall.EACCES with a CAP_NET_BIND_SERVICE / setcap pointer. The
underlying chain is preserved via fmt.Errorf("%w"), so errors.Is /
errors.As continue to walk to the typed bind error and the syscall
errno.
Drop the "headscale ran into an error and had to shut down" wrap,
which only restated the symptom.
Export types.PortFromAddr so the classifier can render the port
number in the hint.
Updates #3227
Address golangci-lint findings in users_test.go: replace inline
"if err := ...; err != nil" statements with plain assignments
(noinlineerr) and add blank lines between statements in the fake
server's filter loop and before an early return (wsl_v5).
resolveSingleUser returned the raw --identifier flag value instead of
the identifier of the matched user. Renaming with --name therefore sent
OldId=0 to the API and failed with "user not found". Return the matched
user's identifier and pass it straight to the rename endpoint.
The new full-tree golangci-lint check reports issues the --new-from-rev
diff lint hid: nine wsl_v5 whitespace gaps, a prealloc, and an unparam
(setCSRFCookie never errored, so drop the return and update callers).
gocyclo on the central UpdateNodeFromMapRequest and an SA1019 NetMap
deprecation in an integration helper are suppressed with reasons.
newRemoteClient prepended "https://" unconditionally, so
HEADSCALE_CLI_ADDRESS=https://host became https://https://host and dialed
the host "https". Default a bare host to https, keep an explicit scheme.
With the v1 API served by Huma and the CLI on the HTTP client, nothing
uses the gRPC service or its generated code. Delete proto/, the generated
gen/go and gen/openapiv2, grpcv1.go, the buf config, the proto .Proto()
helpers and gRPC config, and the proto build tooling from the flake and
CI.
Replace the gRPC client with the generated HTTP client across every
command: locally over the unix socket without auth (matching the previous
local gRPC socket), remotely over TLS with a Bearer API key. Output
rendering and integration tests move to the HTTP client types; the
transport changes, the assertions do not.