Send only the Referer's scheme and host to Sentry (closes #548)

A Sentry event kept the Referer header verbatim, and a Referer can be
an entrypoint URL: a sender that follows the README's http:// to
https:// redirect sends the URL it was redirected from, and a browser
sends the 404 page's URL under /h/ when it fetches that page's
stylesheet. The Sentry hook now keeps only the Referer's scheme and
host. The access log's reduction moves, unchanged, into a new package,
internal/referer, as SchemeAndHost; the access log and the Sentry hook
both call it there. The comment above sentryKeepsHeader and the
README's Sentry passage say so, and the README's Package Layout lists
the new package.

Model: opus-5-5

Dockerfile

@@ -76,8 +76,6 @@ RUN --network=none node_modules/.bin/prettier --check '**/*.md'
 # depends on it.
 #
 # golangci/golangci-lint:v2.14.0 (Debian-based), 2026-09-24
-# Using Debian-based image because mattn/go-sqlite3 (CGO) does not
-# compile on Alpine musl (off64_t is a glibc type).
 FROM golangci/golangci-lint:v2.14.0@sha256:ad862ba6b3798cbe0fd9fd7408d498fd74fbd2623a92406b2fd3898faf0bf98f AS lint
 
 WORKDIR /src
@@ -167,9 +165,7 @@ RUN --mount=type=tmpfs,target=/tmp/tests,size=512m \
 
 # Build stage
 # golang:1.26.1-bookworm (Debian-based), 2026-03-17
-# Using Debian-based image because gorm.io/driver/sqlite pulls in
-# mattn/go-sqlite3 (CGO), which does not compile on Alpine musl. The image
-# ships git and make, which the version step below uses.
+# The image ships git and make, which the version step below uses.
 FROM golang:1.26.1-bookworm@sha256:4465644228bc2857a954b092167e12aa59c006a3492282a6c820bf4755fd64a4 AS builder
 
 # Nothing is wanted from the lint and test phases or from the stylesheet and
@@ -211,26 +207,11 @@ RUN version="$(make version VERSION="$VERSION")"; \
     fi
 
 # Builds through the Makefile's build target, which runs script/assets
-# (Alpine.js, extracted from its tarball in 3p/) first.
+# (Alpine.js, extracted from its tarball in 3p/) first. The target builds
+# with cgo off, so the binary is statically linked and runs on the Alpine
+# runtime image below as it is.
 RUN make build VERSION="$VERSION"
 
-# Rebuild with static linking for the Alpine runtime. The cgo binary the
-# step above builds is dynamically linked against glibc, which Alpine
-# (musl) does not have.
-#
-# A statically linked glibc still loads shared libraries at run time for
-# getaddrinfo and dlopen, and Alpine has none to load, so the build tags
-# keep both out of the binary: netgo and osusergo select Go's own DNS
-# resolver and user lookup instead of glibc's, and
-# sqlite_omit_load_extension leaves out SQLite extension loading, which
-# webhooker never uses.
-#
-# The static flags and tags go in through GO_LDFLAGS and GO_TAGS rather
-# than a -ldflags of their own: the build target composes them with the
-# -X that stamps the version, so this relink cannot silently drop the
-# stamp.
-RUN CGO_ENABLED=1 make build VERSION="$VERSION" GO_TAGS=netgo,osusergo,sqlite_omit_load_extension GO_LDFLAGS='-extldflags "-static"'
-
 # Runtime stage
 # alpine:3.21, 2026-03-17
 FROM alpine:3.21@sha256:c3f8e73fdb79deaebaa2037150150191b9dcbfba68b4a46d70103204c53f4709

Makefile

@@ -16,17 +16,6 @@ VERSION ?= $(shell script/version)
 # exists to correct.
 override VERSION := $(or $(strip $(VERSION)),$(shell script/version))
 
-# Extra linker flags and build tags for the build target. The static
-# relink in the Dockerfile passes -extldflags and its build tags through
-# these rather than a -ldflags of its own, so composing flags cannot drop
-# the version stamp.
-#
-# The build target itself always passes -trimpath and -s -w, as the Go
-# Dockerfile in REPO_POLICIES.md does: no build paths, symbol table or
-# debug information in the binary.
-GO_LDFLAGS ?=
-GO_TAGS ?=
-
 bootstrap:
 	@script/bootstrap
 
@@ -54,8 +43,12 @@ fmt-check:
 check:
 	@script/check
 
+# Built with cgo off, -trimpath and -s -w, as the Go Dockerfile in
+# REPO_POLICIES.md does: the binary carries no build paths, symbol table or
+# debug information, and on Linux, the image's included, it is statically
+# linked.
 build: assets
-	go build -trimpath -tags '$(GO_TAGS)' -ldflags '$(strip -s -w -X main.version=$(VERSION) $(GO_LDFLAGS))' -o bin/webhooker ./cmd/webhooker
+	CGO_ENABLED=0 go build -trimpath -ldflags '-s -w -X main.version=$(VERSION)' -o bin/webhooker ./cmd/webhooker
 
 run: build
 	./bin/webhooker

README.md

@@ -1393,10 +1393,10 @@ It uses:
   management
 - **[go-chi](https://github.com/go-chi/chi)** for HTTP routing
 - **[GORM](https://gorm.io)** for database access with
-  **[modernc.org/sqlite](https://pkg.go.dev/modernc.org/sqlite)** as the runtime
-  SQLite driver. Note: `gorm.io/driver/sqlite` transitively depends on
-  `mattn/go-sqlite3`, which requires CGO at build time (see [Docker](#docker)
-  section)
+  **[modernc.org/sqlite](https://pkg.go.dev/modernc.org/sqlite)**, which is pure
+  Go, as the SQLite driver. `gorm.io/driver/sqlite` also imports
+  `mattn/go-sqlite3`, whose stub for builds without cgo is linked into the
+  binary but never used
 - **[slog](https://pkg.go.dev/log/slog)** (stdlib) for structured logging with
   TTY detection (text for dev, JSON for prod)
 - **[gorilla/sessions](https://github.com/gorilla/sessions)** for encrypted
@@ -2135,9 +2135,7 @@ endpoint. That is the JSON key; the error text for a missing one reads
 `webhook_url is required`, which is the message, not the key.
 
 The database uses the
-[modernc.org/sqlite](https://pkg.go.dev/modernc.org/sqlite) driver at runtime,
-though CGO is required at build time due to the transitive `mattn/go-sqlite3`
-dependency from `gorm.io/driver/sqlite`.
+[modernc.org/sqlite](https://pkg.go.dev/modernc.org/sqlite) driver.
 
 #### Schema
 
@@ -2496,6 +2494,9 @@ are both taken unconditionally, the body precisely because these handlers call
 body with `(redacted)`, drops cookies and the remote-address environment, and
 reduces the headers to a fixed allowlist — `Accept`, `Content-Length`,
 `Content-Type`, `Host`, `Origin`, `Referer`, `User-Agent` and `X-Request-Id`.
+`Referer` is sent as its scheme and host only, because its path can be an
+entrypoint URL; `internal/referer.SchemeAndHost` does that reduction for both
+the access log's `referer` field and this hook.
 
 The same hook rewrites the request URL. The SDK builds it as
 `scheme://host/path` from the concrete path, which on the receiver route is
@@ -2566,8 +2567,9 @@ receiver route. What survives still names the failing route — scheme, host,
 route pattern, method — and `X-Request-Id` ties the event to the local access
 log line that holds the rest. Nothing dropped is needed for the likeliest use,
 debugging a CSRF rejection. Its three inputs are the TLS decision, `Origin` and
-`Referer`; the latter two are kept, and the first is the scheme of the retained
-URL, because the SDK derives that scheme from
+`Referer`; the latter two are kept (`Referer` as its scheme and host, which is
+all the check compares), and the first is the scheme of the retained URL,
+because the SDK derives that scheme from
 `r.TLS != nil || r.Header.Get("X-Forwarded-Proto") == "https"`. That predicate
 is the SDK's own and is stricter than `internal/reqtls.IsTLS`, which this
 service now uses everywhere it decides transport: the SDK reports `http` for the
@@ -3099,6 +3101,8 @@ webhooker/
 │   │   ├── ratelimit.go       # Per-IP rate limiting middleware (go-chi/httprate)
 │   │   ├── loginguard.go      # Login failure counters and the Argon2id verification semaphore
 │   │   └── middlewaretest/    # Test support: a Middleware for tests in other packages
+│   ├── referer/
+│   │   └── referer.go         # SchemeAndHost: cuts a Referer down to its scheme and host, for the access log and the Sentry hook
 │   ├── reqtls/
 │   │   └── reqtls.go          # IsTLS: the one TLS predicate, r.TLS or X-Forwarded-Proto
 │   ├── server/
@@ -3468,16 +3472,12 @@ fixed independently of the compiler's:
    a step's log, and then fails. `make test` builds this stage alone.
 6. **Builder stage** (`golang:1.26.1-bookworm`) — depends on the lint and test
    phases and the `css-check` and `markdown-check` stages passing (it copies a
-   file from each), runs `make build` (which extracts Alpine.js from `3p/`
-   first), and then rebuilds the binary with `CGO_ENABLED=1` and static linking
-   so it runs on musl, with the build tags `netgo`, `osusergo` and
-   `sqlite_omit_load_extension` so it needs none of glibc's shared libraries at
-   run time. Both builds go through `make build`, the relink adding its
-   `-extldflags` via `GO_LDFLAGS` and its tags via `GO_TAGS`, so neither can
-   drop the `-X` that stamps the version. The version is the `VERSION` build arg
-   if one is given, otherwise derived from the `.git` in the context, and the
-   stage fails if a context with `.git` would stamp an empty version, `dev` or
-   `unknown` (see [Version stamping](#version-stamping)).
+   file from each), and runs `make build` (which extracts Alpine.js from `3p/`
+   first). `make build` compiles with cgo off, so the binary is statically
+   linked and runs on the Alpine runtime image. The version is the `VERSION`
+   build arg if one is given, otherwise derived from the `.git` in the context,
+   and the stage fails if a context with `.git` would stamp an empty version,
+   `dev` or `unknown` (see [Version stamping](#version-stamping)).
 7. **Runtime stage** (`alpine:3.21`) — copies the static binary and
    `deploy/docker-entrypoint.sh`, creates the `/var/lib/webhooker` directory for
    all SQLite databases, exposes port 8080, and includes a health check against
@@ -3489,11 +3489,6 @@ The lint and test phases invoke `gofmt`, `golangci-lint` and `go test` directly
 rather than `make fmt-check`, `make lint` and `make test`: those targets build
 docker stages, which would need a docker daemon inside this build.
 
-The lint phase and the Go stages use Debian rather than Alpine because
-`gorm.io/driver/sqlite` pulls in `mattn/go-sqlite3`, which needs CGO and does
-not compile against musl. Only the final binary is statically linked, which is
-what lets it run on the Alpine runtime image.
-
 `script/cibuild` is the CI gate: it runs `script/bootstrap`, then
 `script/check`, then builds the image, whose build runs the lint and test phases
 and the stylesheet and Markdown checks again. A build that succeeds is a repo

internal/middleware/middleware.go

@@ -22,6 +22,7 @@ import (
 	"sneak.berlin/go/webhooker/internal/config"
 	"sneak.berlin/go/webhooker/internal/globals"
 	"sneak.berlin/go/webhooker/internal/logfield"
+	"sneak.berlin/go/webhooker/internal/referer"
 	"sneak.berlin/go/webhooker/internal/session"
 )
 
@@ -48,10 +49,6 @@ const (
 	// pick the size of the line it writes.
 	redactedQuery = "?(redacted)"
 
-	// redactedReferer stands in for a Referer the access log cannot
-	// reduce to its scheme and host; see accessLogReferer.
-	redactedReferer = "(redacted)"
-
 	// maxLogRequestIDBytes bounds the request id, which is also
 	// client-supplied: chi's RequestID middleware passes an inbound
 	// X-Request-Id header through verbatim. Its generated form is an
@@ -342,27 +339,6 @@ func accessLogURL(r *http.Request, status int) string {
 	return pattern
 }
 
-// accessLogReferer returns the value for the access log's referer
-// field: the Referer's scheme and host, never its path or query. The
-// path can be an entrypoint URL, which carries the UUID: a sender that
-// follows a redirect from http:// to https:// sends the URL it was
-// redirected from, and a browser sends the 404 page's URL under /h/
-// when it fetches that page's stylesheet. A Referer without both a
-// scheme and a host is logged as redactedReferer.
-func accessLogReferer(r *http.Request) string {
-	referer := r.Referer()
-	if referer == "" {
-		return ""
-	}
-
-	parsed, err := url.Parse(referer)
-	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
-		return redactedReferer
-	}
-
-	return parsed.Scheme + "://" + parsed.Host
-}
-
 // Logging returns middleware that logs each HTTP request with
 // timing and metadata.
 func (s *Middleware) Logging() func(http.Handler) http.Handler {
@@ -413,7 +389,8 @@ func (s *Middleware) Logging() func(http.Handler) http.Handler {
 						requestID, maxLogRequestIDBytes,
 					),
 					"referer", logfield.Truncate(
-						accessLogReferer(r), logfield.MaxBytes,
+						referer.SchemeAndHost(r.Referer()),
+						logfield.MaxBytes,
 					),
 					"proto", r.Proto,
 					"remoteIP", RemoteIP(r),

internal/referer/referer.go

@@ -0,0 +1,32 @@
+// Package referer reduces a Referer header to its scheme and host, so
+// that a URL path, which on the receiver route carries the entrypoint
+// UUID, never reaches a log line or an error report.
+package referer
+
+import "net/url"
+
+// redacted stands in for a Referer that cannot be reduced to its
+// scheme and host.
+const redacted = "(redacted)"
+
+// SchemeAndHost returns a Referer's scheme and host, never its path or
+// query. The path can be an entrypoint URL, which carries the UUID: a
+// sender that follows a redirect from http:// to https:// sends the
+// URL it was redirected from, and a browser sends the 404 page's URL
+// under /h/ when it fetches that page's stylesheet. A Referer without
+// both a scheme and a host comes back as "(redacted)".
+//
+// It is the access log's referer field, and the Referer header the
+// Sentry hook sends.
+func SchemeAndHost(referer string) string {
+	if referer == "" {
+		return ""
+	}
+
+	parsed, err := url.Parse(referer)
+	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
+		return redacted
+	}
+
+	return parsed.Scheme + "://" + parsed.Host
+}

internal/referer/referer_test.go

@@ -0,0 +1,67 @@
+package referer_test
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"sneak.berlin/go/webhooker/internal/referer"
+)
+
+// entrypointUUID is the secret part of an entrypoint URL; no case may
+// return it.
+const entrypointUUID = "0b6f7d3e-6a39-4e3c-9f0a-2c4d8e1b5a77"
+
+// redacted is what SchemeAndHost returns for a Referer it cannot
+// reduce to a scheme and host.
+const redacted = "(redacted)"
+
+func TestSchemeAndHost(t *testing.T) {
+	t.Parallel()
+
+	for _, tc := range []struct {
+		name    string
+		referer string
+		want    string
+	}{
+		{
+			name:    "no Referer",
+			referer: "",
+			want:    "",
+		},
+		{
+			name: "entrypoint URL with a query",
+			referer: "http://webhooker.example/h/" + entrypointUUID +
+				"?src=ci",
+			want: "http://webhooker.example",
+		},
+		{
+			name:    "host with a port",
+			referer: "https://webhooker.example:8443/h/" + entrypointUUID,
+			want:    "https://webhooker.example:8443",
+		},
+		{
+			name:    "path without a scheme or host",
+			referer: "/h/" + entrypointUUID,
+			want:    redacted,
+		},
+		{
+			name:    "host without a scheme",
+			referer: "webhooker.example/h/" + entrypointUUID,
+			want:    redacted,
+		},
+		{
+			name:    "not a URL",
+			referer: "http://[::1/h/" + entrypointUUID,
+			want:    redacted,
+		},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+
+			got := referer.SchemeAndHost(tc.referer)
+
+			assert.Equal(t, tc.want, got)
+			assert.NotContains(t, got, entrypointUUID)
+		})
+	}
+}

internal/server/sentry.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/getsentry/sentry-go"
 	"github.com/go-chi/chi"
+	"sneak.berlin/go/webhooker/internal/referer"
 )
 
 // sentryRedacted stands in for a withheld field on every event shipped
@@ -253,6 +254,7 @@ func sentryTransactionName(name, pattern string) string {
 // gorilla/csrf accepts in place of the form field — and the shared
 // secrets senders put on the receiver route (X-Gitlab-Token and the
 // per-provider signature headers) would otherwise ship verbatim.
+// The Referer it keeps is cut down to its scheme and host.
 func keptSentryHeaders(headers map[string]string) map[string]string {
 	if len(headers) == 0 {
 		return headers
@@ -261,9 +263,15 @@ func keptSentryHeaders(headers map[string]string) map[string]string {
 	kept := make(map[string]string, len(headers))
 
 	for name, value := range headers {
-		if sentryKeepsHeader(name) {
-			kept[name] = value
+		if !sentryKeepsHeader(name) {
+			continue
 		}
+
+		if http.CanonicalHeaderKey(name) == "Referer" {
+			value = referer.SchemeAndHost(value)
+		}
+
+		kept[name] = value
 	}
 
 	return kept
@@ -271,11 +279,14 @@ func keptSentryHeaders(headers map[string]string) map[string]string {
 
 // sentryKeepsHeader reports whether a request header is routing or
 // content metadata rather than client-chosen payload. Referer is kept
-// on the reasoning that it is browser-set, that the only query
-// parameters in this service's own URLs are the login page's `next`,
-// `notice` and the event log's `show`, and that Referrer-Policy is set to
-// strict-origin-when-cross-origin. X-Request-Id ties the event to the
-// local access log line, which holds the rest of the detail.
+// only as its scheme and host, through referer.SchemeAndHost, which
+// the access log uses too: its path can be an entrypoint URL, which
+// carries the UUID, as when a sender follows a redirect from http://
+// to https:// or a browser fetches the 404 page's stylesheet from
+// under /h/. The scheme and host are all gorilla/csrf compares a
+// Referer by, so a CSRF rejection is still readable from the event.
+// X-Request-Id ties the event to the local access log line, which
+// holds the rest of the detail.
 func sentryKeepsHeader(name string) bool {
 	switch http.CanonicalHeaderKey(name) {
 	case "Accept",

internal/server/sentry_test.go

@@ -325,6 +325,37 @@ func TestSentryScrub_ReplacesTheCapabilityPathWithTheRoutePattern(
 	)
 }
 
+// TestSentryScrub_KeepsOnlyTheRefererSchemeAndHost sends a receiver
+// request whose Referer is the entrypoint URL, as a sender's is after
+// it follows the README's redirect from http:// to https://. The event
+// keeps the Referer's scheme and host, and carries the UUID nowhere.
+func TestSentryScrub_KeepsOnlyTheRefererSchemeAndHost(t *testing.T) {
+	t.Parallel()
+
+	event := onlyEvent(t, sentryCase{
+		scrub:  true,
+		panics: true,
+		request: func(client *sentry.Client) *http.Request {
+			req := sentryReceiverRequest(client)
+			req.Header.Set(
+				"Referer",
+				"http://webhooker.example/h/"+sentryReceiverUUID,
+			)
+
+			return req
+		},
+	}.capture(t))
+
+	assert.NotContains(
+		t, marshalEvent(t, event), sentryReceiverUUID,
+	)
+	assert.Equal(
+		t,
+		"http://webhooker.example",
+		event.Request.Headers["Referer"],
+	)
+}
+
 // TestSentryScrub_KeepsTheRoutingContext checks the hook does not cost
 // the debugging signal: the route, its scheme and host, the method and
 // the metadata headers still identify what failed. On a static route

internal/versionscript/version_script_test.go

@@ -153,19 +153,15 @@ func TestVersion_IsStableAcrossInvocations(t *testing.T) {
 	require.Equal(t, first, second)
 }
 
-// The build target is the only place composing linker flags. If a
-// future edit drops either half, the binary silently reports "dev"
-// again (the defect this package exists for) or fails to link on
-// Alpine. If it stops passing GO_TAGS, the image's static binary links
-// glibc's getaddrinfo and dlopen again.
-func TestMakefile_BuildComposesVersionAndExtraFlags(t *testing.T) {
+// The build target is the only place that stamps the version. If a
+// future edit drops the -X flag, the binary silently reports "dev"
+// again, the defect this package exists for.
+func TestMakefile_BuildStampsTheVersion(t *testing.T) {
 	t.Parallel()
 
 	makefile := read(t, makefilePath)
 
 	require.Contains(t, makefile, "-X main.version=$(VERSION)")
-	require.Contains(t, makefile, "$(GO_LDFLAGS)")
-	require.Contains(t, makefile, "-tags '$(GO_TAGS)'")
 	require.Contains(t, makefile, "VERSION ?= $(shell script/version)")
 }
 
@@ -224,8 +220,8 @@ func requireMake(t *testing.T) {
 	}
 }
 
-// Every compile in the image goes through the build target, so the
-// static relink cannot replace the flags that carry the stamp.
+// The image compiles through the build target, so it carries the
+// target's version stamp.
 func TestDockerfile_BuildsThroughTheMakeTarget(t *testing.T) {
 	t.Parallel()
 
@@ -234,10 +230,7 @@ func TestDockerfile_BuildsThroughTheMakeTarget(t *testing.T) {
 	require.NotContains(t, dockerfile, "go build",
 		"a raw go build bypasses the Makefile's -X flag")
 	require.Contains(t, dockerfile, "ARG VERSION")
-	require.Contains(t, dockerfile,
-		`make build VERSION="$VERSION"`+
-			` GO_TAGS=netgo,osusergo,sqlite_omit_load_extension`+
-			` GO_LDFLAGS='-extldflags "-static"'`)
+	require.Contains(t, dockerfile, `RUN make build VERSION="$VERSION"`)
 }
 
 func read(t *testing.T, path string) string {