Files
OwnCord/Server/api/upload_handler.go
T
J3vbandClaude Fable 5 f5faf82a60 infra: observability, backups, guardrails, and deployment hardening (#1376)
* docs: add infrastructure roadmap plan

Records the verified recommendations from an infrastructure review in three
tracks: raising the single-instance ceiling, cheap seams for a possible
multi-instance future, and ops hygiene. Includes explicit anti-recommendations
and sequencing. Security-sensitive detail is intentionally excluded per
docs/security.md.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* feat(server): real health checks and saturation metrics

/api/v1/metrics now exposes signals that were already computed in memory but
never surfaced: reconnect replay tier hits, event-persister counters, SQLite
writer-pool wait stats, aggregate per-client backpressure counters (including
previously invisible low-priority drops), and permission-cache hit/miss.

/health now returns a real verdict: hub dispatch-loop liveness, a bounded
database ping, and a free-disk check, returning 503 with a subsystem reason
when degraded. Checks are cached so the unauthenticated endpoint cannot
amplify load. The hub's panic breaker now exits the process so a supervisor
can restart it, instead of leaving broadcast delivery silently dead while
clients still appear online.

OTel instruments that were declared but never recorded are now wired
(ws_active_connections, ws_broadcast_latency_seconds, ws_messages_total,
ws_events_dropped_total, voice gauges) or removed (db_query_duration_seconds).
Also corrects the docs/api.md description of broadcast_drops, which counts
hub-queue overflow, not client send-queue overflow.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* feat(server): implement scheduled backups, retention, and backup verification

The backup_schedule and backup_retention settings have existed in the admin
panel and API since the initial schema but were never read by any code. The
15-minute maintenance loop now enforces them: a scheduled backup is taken
when the newest backup on disk is older than the schedule interval (manual
backups reset the clock), and retention prunes backups older than the
configured days while always keeping the newest one.

Backups are now verified with PRAGMA integrity_check immediately after
VACUUM INTO (a failed backup is removed rather than listed as restorable)
and again before a restore may overwrite the live database. A failed VACUUM
INTO also cleans up its partial output file — but never a pre-existing one.

The backup directory is configurable via a new backup.dir key (default
data/backups) so operators can point backups at another disk or an off-host
mount, mirroring the SetDatabasePath plumb.

Restore-handler tests now use real SQLite fixtures (the integrity gate
correctly refuses text files) with the mid-copy failure injected through a
test-only copy hook. Also adds audited gosec suppressions to the Windows
disk-free syscall added in the previous commit, which the Windows lint leg
flagged.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* feat(server): capacity and failure-mode guardrails

- server.max_ws_connections: optional cap on concurrent WebSocket clients,
  checked before the upgrade with a 503 + Retry-After; rejections are counted
  and exposed as ws_conn_rejects in /api/v1/metrics.
- Single-process database lock: an OS-level advisory lock (flock / exclusive
  handle) beside the SQLite file makes a second server process fail fast with
  a clear message instead of silently fighting the first over process-local
  state. A bounded retry covers the self-update/restore restart handoff, and
  the lock mechanism failing (e.g. network filesystems) only warns.
- Disk-space awareness: boot-time warnings for the data and backup volumes,
  plus a disk_free_mb metrics field, via a small cross-platform diskutil
  package (already used by /health).
- Upload storage failures: storage.Save now marks server-side filesystem
  failures with a sentinel (storage.ErrIO); handlers return 507 for those
  instead of blaming the client with a 400, and the emoji route stops echoing
  raw storage errors (which embed absolute paths) into responses.
- Unknown config keys now warn at startup — a typo like admin_alowed_cidrs
  previously kept the default silently while the operator believed the
  setting changed. Never fatal: newer servers tolerate older configs.
- Admin settings honesty: the three stored-but-inert settings (server_icon,
  max_upload_bytes, voice_quality) are shown read-only with a note pointing
  at the real config.yaml keys, instead of pretending to apply.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* perf(db): write-path efficiency and capacity knobs

- channel_focus/mark_read now skip the read-state UPSERT when the stored row
  already matches (same last_message_id, no mentions) — refocus events fire
  at up to 10/s/user and every no-op write still occupied the single SQLite
  writer connection. The extra existence check runs on the reader pool, which
  doesn't serialize. Same shape as the session-touch throttle.
- DeleteExpiredSessions is now sargable: migration 031 normalizes legacy
  expiry formats to the RFC3339-Z layout the server writes and indexes
  expires_at, replacing the strftime full-table scan that ran on the writer
  every 15 minutes.
- Boot-time ANALYZE runs only when a migration actually applied; unchanged
  schemas get the cheap PRAGMA optimize instead (which also covers
  crash-restarts that never reached the shutdown optimize).
- The read/write SQL router gets a table-driven test with explicit expected
  values (INSERT ... RETURNING must hit the writer despite being :one).
- New knobs, all defaulting to current behavior: database.max_readers,
  security.auth_rate_limit_multiplier (for shared-NAT communities),
  event_persistence.replay_ring_size and replay_cold_limit.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* fix(server): shutdown lifecycle ordering

- The event pruner and maintenance loop are now joined (bounded) before the
  database closes: bgCtx cancellation used to run AFTER database.Close via
  LIFO defers, contradicting its own comment, and neither goroutine was ever
  waited on — a mid-tick scheduled backup or prune could still hold the
  writer while the pool tore down. StartEventPruner returns a done channel
  with the same join contract EventPersister.Stop already had.
- srv.Shutdown now runs before hub.GracefulStop, so in-flight HTTP handlers'
  broadcasts still reach a live hub and the event persister instead of
  vanishing from the replay/event store across a restart. Shutdown does not
  wait on hijacked WebSocket connections, so the swap adds no delay.
- GracefulStopContext threads the 30s shutdown budget into the hub: the 5s
  client-notice window (matching the countdown clients are shown) ends early
  when the budget expires, and is skipped entirely when nobody is connected —
  early-return startup paths and idle servers no longer sleep 5s for an
  audience of zero.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* build(deploy): systemd unit, compose hardening, boot-smoked releases, CI polish

- deploy/owncord.service: hardened systemd unit template with the two
  verified caveats encoded (install dir stays writable for self-update under
  ProtectSystem=strict; CAP_NET_BIND_SERVICE for ACME's :80), plus a
  'Linux (systemd)' deployment docs section — the Linux service story was
  previously 'Docker or nothing'.
- New 'Reverse Proxy Topology' docs section with a working nginx snippet and
  the correct signaling-vs-media distinction: /livekit/* is already proxied
  by the server, only WebRTC media ports must be directly reachable.
- docker-compose: log rotation, commented resource limits, and a healthcheck
  backed by a new 'chatserver healthcheck' subcommand (the distroless image
  has no shell) that probes /health without config side effects.
- release.yml: a concurrency group (queue, never cancel), and boot-smoke
  gates — the freshly built server binaries and the Docker image are cold
  booted and probed healthy BEFORE anything is signed or pushed. The release
  feed drives signed self-updates, so a binary that compiles but dies on
  boot previously would have shipped itself to every auto-updating instance.
- ci.yml: client-check/client-tests move to ubuntu with the reasoning
  recorded (no win32 code paths, LF enforced repo-wide); admin-e2e gets a
  written graduation criterion instead of an open-ended non-blocking status.
- docs: Tailscale guide notes the CGNAT range vs the default admin CIDRs;
  architecture overview records presence/voice state as the fifth
  single-instance blocker and the macOS client scope decision.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* perf(server): measured load tooling, narrowed invalidation, presence coalescing, storage and CIDR seams

- Fix scripts/k6/ws-load.js against the real wire protocol: envelope-wrapped
  frames, correct message types (typing_start, presence_update), the correct
  /api/v1/ws path, and thresholds that fail a run where nobody authenticated
  or went ready — the script had drifted to pre-envelope framing and reported
  100% green while every auth failed on the first frame. A new
  workflow_dispatch-only load-baseline workflow boots a real server, seeds
  users through the setup/invite APIs, runs the script, and uploads the k6
  summary plus a metrics snapshot for before/after comparison.
- Role-scoped channel-override changes now evict only the affected role's
  members from the permission cache (fail-safe: unreadable member list still
  flushes everything). InvalidateAll here repopulated every connected user —
  two reads each — synchronously inside the admin request via
  RefreshChannelVisibility, a stampede that scaled with total population
  rather than the role's size. Same pattern the per-user override endpoints
  already used.
- Connect/disconnect presence broadcasts now pass through a 300ms latest-wins
  coalescer (QueuePresence): each un-coalesced presence change is a sequenced
  global broadcast (an O(clients) fan-out under seqMu), so a reconnect storm
  fired O(users) of them from the connect critical path. A flap inside the
  window collapses to its final state; the wire format, seq ordering, and
  replay behaviour are unchanged, and the delivery path (BroadcastPresence)
  is untouched.
- Storage seam: api handlers now consume a FileStore interface (consumer-side,
  same pattern as service.Store) with Open returning a seekable storage.File —
  writing down the contract (range-request seeks included) an alternative
  backend would have to meet, without building one.
- The metrics surfaces and the LiveKit webhook/health endpoints get their own
  allowlist keys (metrics_allowed_cidrs, livekit_webhook_allowed_cidrs, both
  defaulting to admin_allowed_cidrs), so a central Prometheus scraper or an
  externally-hosted LiveKit no longer requires widening the admin panel's
  perimeter. Startup now also warns when admin_allowed_cidrs is customized
  while trusted_proxies is empty — behind a proxy or container network the
  check would otherwise compare the proxy's private address, not the client's.
- The container healthcheck probe now PINS the server's own certificate from
  disk (VerifyConnection, exact-match) instead of skipping TLS verification,
  addressing the CodeQL finding on the previous commit; WebPKI verification
  is used when no local cert exists (ACME).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* fix(server): address self-review findings on the hardening branch

Seven fixes from a high-effort review of the full branch diff:

- healthcheck CLI now works under tls.mode acme: it overrides ServerName
  with the configured domain for WebPKI verification instead of pinning a
  cert that doesn't exist (or is stale) in that mode. Previously an ACME
  deployment's container healthcheck failed forever.
- /health pings the READER pool (new db.PingRead): the writer ping queued
  behind a scheduled backup's VACUUM INTO and reported the server degraded
  for the whole backup — which an autoheal watchdog would turn into a
  nightly mid-backup restart.
- /health runs its cached checks under context.WithoutCancel so a probe
  that disconnects mid-request cannot poison the shared cache with a false
  degraded verdict for the next 5 seconds.
- The token CLI uses a new db.OpenShared that skips the single-process
  lock: minting a token against a running server is safe under WAL and was
  a documented workflow the lock had broken.
- The per-user TOTP failure cap is no longer scaled by
  security.auth_rate_limit_multiplier — that knob exists for per-IP limits;
  scaling the only cross-IP brute-force defence multiplied an attacker's
  distributed guess budget. Mirrors the unscaled per-user login threshold.
- A direct presence_update now drops the user's queued entry in the
  connect/disconnect coalescer, so a stale connect-time presence can no
  longer flush 300ms later over the user's fresher chosen status.
- The scheduled-backup filename collision loop breaks on any stat error
  and bounds its suffix probing, instead of spinning the maintenance
  goroutine forever on a persistent EACCES.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

* test(admin): real SQLite fixture for the merged Close-failure restore test

TestHandleRestoreBackup_RestartsWhenCloseFails arrived from main (#1375)
with a plain-text backup fixture; this branch's restore handler verifies
backups with integrity_check before touching the live database, so the text
fixture was (correctly) refused with 400 before the Close-failure branch
under test was reached. Use a real backup via BackupToSafe, matching the
other restore tests.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017RtDNHSYWwPKArL8MsRdbj

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-08-15 20:50:47 +02:00

453 lines
16 KiB
Go
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
package api
import (
"database/sql"
"errors"
"image"
_ "image/gif"
_ "image/jpeg"
_ "image/png"
"io"
"log/slog"
"mime"
"net/http"
"path/filepath"
"strings"
"time"
"unicode"
"unicode/utf8"
"github.com/go-chi/chi/v5"
"github.com/google/uuid"
"github.com/owncord/server/auth"
"github.com/owncord/server/db"
"github.com/owncord/server/permissions"
"github.com/owncord/server/service"
"github.com/owncord/server/storage"
)
// uploadResponse is the JSON shape returned by POST /api/v1/uploads.
type uploadResponse struct {
ID string `json:"id"`
Filename string `json:"filename"`
Size int64 `json:"size"`
Mime string `json:"mime"`
URL string `json:"url"`
Width *int `json:"width,omitempty"`
Height *int `json:"height,omitempty"`
}
// sanitizeUploadFilename cleans an upload filename: strips control and
// invisible formatting characters, removes path separators, and truncates to a
// safe length.
func sanitizeUploadFilename(name string) string {
// Strip path components — use only the base name.
name = filepath.Base(name)
// filepath.Base only understands the *server* OS's separator, so a
// backslash survives on a Linux server and is then a path separator on the
// victim's Windows client, where the name is pre-filled into a save dialog.
if i := strings.LastIndexByte(name, '\\'); i >= 0 {
name = name[i+1:]
}
// Remove control characters, invisible formatting characters, and any
// residual forward slash.
var sb strings.Builder
for _, r := range name {
// unicode.Cf covers the bidi overrides (U+202AU+202E, U+2066U+2069):
// invisible characters that reorder how the name renders, so an
// attachment can display a harmless-looking extension to every other
// member of the channel while really being an executable script — and
// the same string is what the native save dialog pre-fills. This is the
// rule auth.ValidateUsername already applies to usernames.
//
// A forward slash is dropped too: filepath.Base("/") returns "/" (root
// is its own basename), so an upload literally named "/" would otherwise
// slip through the reserved-name check below with a path separator
// intact. Any residual '/' is unsafe as a basename, so strip it here.
if unicode.IsControl(r) || unicode.In(r, unicode.Cf) || r == '/' {
continue
}
sb.WriteRune(r)
}
name = strings.TrimSpace(sb.String())
// Truncate to the filesystem limit. Slicing by byte offset can land in the
// middle of a multibyte rune, so trim back to the last full rune to keep the
// result valid UTF-8 (an invalid name misbehaves in JSON encoding, on disk,
// and in the client's download-name handling).
if len(name) > maxUploadFilenameLength {
name = name[:maxUploadFilenameLength]
for len(name) > 0 && !utf8.ValidString(name) {
name = name[:len(name)-1]
}
}
if name == "" || name == "." || name == ".." {
name = "unnamed"
}
return name
}
// isUnsafeInlineMIME returns true for MIME types that could execute active
// content (scripts, markup) if served inline under the OwnCord origin.
func isUnsafeInlineMIME(mimeType string) bool {
// Normalize: take the base type before any parameters (e.g. "text/html; charset=utf-8").
base := strings.SplitN(mimeType, ";", 2)[0]
base = strings.TrimSpace(strings.ToLower(base))
switch base {
case "text/html", "application/xhtml+xml",
"image/svg+xml", "text/xml", "application/xml",
"application/pdf",
"text/xsl", "text/xslt":
return true
}
return false
}
// safeStorageErrorMessage maps a storage.Save error to a client-safe
// "upload rejected" body. Full detail always goes to slog.Warn at the call
// site — this only decides what crosses the HTTP boundary. storage.Save's
// failure messages are built with fmt.Errorf("... %s", dst) / %w around
// path-bearing OS errors (creating the file, syncing it, or the destination
// resolving outside the storage dir), so echoing them verbatim hands any
// authenticated user the server's absolute storage layout the moment a save
// fails (disk full, permission change, read-only mount). The two validation
// failures below are the only ones that never embed a path, so they're the
// only ones whose detail is forwarded.
func safeStorageErrorMessage(err error) string {
msg := err.Error()
switch {
case strings.HasPrefix(msg, "blocked file type:"),
strings.HasPrefix(msg, "file exceeds maximum size"):
return "upload rejected: " + msg
default:
return "upload rejected"
}
}
// writeStorageSaveError maps a storage.Save failure onto the right HTTP
// class: server-side filesystem failures (storage.ErrIO — disk full,
// permissions, read-only mount) become 507 so they are distinguishable from
// bad uploads in any status dashboard; everything else stays the client's
// 400. Detail never crosses the HTTP boundary either way (path leakage —
// see safeStorageErrorMessage).
func writeStorageSaveError(w http.ResponseWriter, saveErr error, what string) {
if errors.Is(saveErr, storage.ErrIO) {
slog.Error(what+" failed: server storage error", "error", saveErr)
writeJSON(w, http.StatusInsufficientStorage, errorResponse{
Error: "STORAGE_ERROR",
Message: "upload failed: server storage error",
})
return
}
slog.Warn(what+" rejected", "error", saveErr)
writeJSON(w, http.StatusBadRequest, errorResponse{
Error: "BAD_REQUEST",
Message: safeStorageErrorMessage(saveErr),
})
}
// MountUploadRoutes registers upload and file-serving endpoints.
// allowedOrigins controls the Access-Control-Allow-Origin header on served files.
//
// permSvc MUST be non-nil — handleServeFile dereferences it to enforce
// per-channel ACLs on every file download. A nil permSvc would panic for
// any authenticated file request, so we fail fast at mount time rather
// than let the first user hit a 500.
func MountUploadRoutes(r chi.Router, database *db.DB, store FileStore, limiter *auth.RateLimiter, allowedOrigins []string, permSvc *service.PermissionService) {
if permSvc == nil {
panic("api: MountUploadRoutes requires a non-nil PermissionService")
}
// Upload requires authentication and a higher body size limit (100 MB).
r.With(
AuthMiddleware(database),
MaxBodySize(uploadMaxBodySize),
).Post("/api/v1/uploads", handleUpload(database, store, limiter))
// File serving requires authentication for channel-level access control.
r.With(AuthMiddleware(database)).Get("/api/v1/files/{id}", handleServeFile(database, store, allowedOrigins, permSvc))
}
func handleUpload(database *db.DB, store FileStore, limiter *auth.RateLimiter) http.HandlerFunc {
return func(w http.ResponseWriter, r *http.Request) {
// BUG-131: Per-user upload rate limit to prevent disk exhaustion.
user, ok := r.Context().Value(UserKey).(*db.User)
if ok && user != nil {
uploadKey := auth.Key("upload", user.ID)
if !limiter.Allow(uploadKey, uploadRateLimitPerMinute, time.Minute) {
writeJSON(w, http.StatusTooManyRequests, errorResponse{
Error: "RATE_LIMITED",
Message: "upload rate limit exceeded, try again later",
})
return
}
}
// Limit request body size to prevent abuse.
r.Body = http.MaxBytesReader(w, r.Body, uploadMaxBodySize)
// Parse multipart form — 10 MB in memory, rest on disk.
if err := r.ParseMultipartForm(multipartMemoryLimit); err != nil {
writeJSON(w, http.StatusBadRequest, errorResponse{
Error: "BAD_REQUEST",
Message: "invalid multipart form",
})
return
}
file, header, err := r.FormFile("file")
if err != nil {
writeJSON(w, http.StatusBadRequest, errorResponse{
Error: "BAD_REQUEST",
Message: "missing file field",
})
return
}
defer file.Close() //nolint:errcheck
// Generate UUID for storage.
fileID := uuid.New().String()
// Detect MIME type from actual file bytes (never trust client header).
var sniffBuf [512]byte
n, readErr := file.Read(sniffBuf[:])
if readErr != nil && !errors.Is(readErr, io.EOF) && !errors.Is(readErr, io.ErrUnexpectedEOF) {
writeJSON(w, http.StatusBadRequest, errorResponse{
Error: "BAD_REQUEST",
Message: "failed to read uploaded file",
})
return
}
detectedMime := http.DetectContentType(sniffBuf[:n])
// Seek back so the full content is available for storage.
if _, seekErr := file.Seek(0, 0); seekErr != nil {
writeJSON(w, http.StatusInternalServerError, errorResponse{
Error: "INTERNAL_ERROR",
Message: "failed to process uploaded file",
})
return
}
mime := detectedMime
// Store file on disk (validates file type via magic bytes).
writtenBytes, saveErr := store.Save(fileID, file)
if saveErr != nil {
writeStorageSaveError(w, saveErr, "file upload")
return
}
// Extract image dimensions if the file is an image.
var width, height *int
if strings.HasPrefix(mime, "image/") {
f, openErr := store.Open(fileID)
if openErr == nil {
cfg, _, decErr := image.DecodeConfig(f)
f.Close() //nolint:errcheck
if decErr == nil {
w2, h2 := cfg.Width, cfg.Height
width = &w2
height = &h2
} else {
slog.Warn("failed to decode image dimensions", "id", fileID, "error", decErr)
}
}
}
// Insert attachment record in DB (unlinked — message_id is NULL).
user, _ = r.Context().Value(UserKey).(*db.User)
safeFilename := sanitizeUploadFilename(header.Filename)
if err := database.CreateAttachment(r.Context(), fileID, user.ID, safeFilename, fileID, mime, writtenBytes, width, height); err != nil {
// Clean up stored file on DB failure.
_ = store.Delete(fileID)
slog.Error("failed to create attachment record", "error", err)
writeJSON(w, http.StatusInternalServerError, errorResponse{
Error: "INTERNAL_ERROR",
Message: "failed to save attachment",
})
return
}
slog.Info("file uploaded", "id", fileID, "filename", safeFilename, "size", writtenBytes, "mime", mime)
writeJSON(w, http.StatusCreated, uploadResponse{
ID: fileID,
Filename: safeFilename,
Size: writtenBytes,
Mime: mime,
URL: "/api/v1/files/" + fileID,
Width: width,
Height: height,
})
}
}
func handleServeFile(database *db.DB, store FileStore, allowedOrigins []string, permSvc *service.PermissionService) http.HandlerFunc {
return func(w http.ResponseWriter, r *http.Request) {
fileID := chi.URLParam(r, "id")
if fileID == "" {
http.NotFound(w, r)
return
}
user, _ := r.Context().Value(UserKey).(*db.User)
role, _ := r.Context().Value(RoleKey).(*db.Role)
// Look up attachment metadata with channel context.
aa, err := database.GetAttachmentWithChannel(r.Context(), fileID)
if err != nil {
slog.Error("failed to look up attachment", "id", fileID, "error", err)
writeJSON(w, http.StatusInternalServerError, errorResponse{
Error: "INTERNAL_ERROR",
Message: "internal server error",
})
return
}
if aa == nil {
http.NotFound(w, r)
return
}
// A soft-deleted message's attachments must stop being servable the
// moment the message is deleted — the client shows a tombstone, but
// without this check the file stays reachable by URL forever (no
// sweep can ever reclaim a linked row either, since the only reaper
// requires message_id IS NULL). Checked before the ACL branch so it
// also covers admins, matching the tombstone applying to everyone.
//
// Queried directly rather than through database.GetMessage: that
// wrapper's SELECT list carries every message column, and the
// `deleted` flag is the only one this check needs.
if aa.MessageID != nil {
var deleted bool
deletedErr := database.QueryRowContext(r.Context(),
`SELECT deleted FROM messages WHERE id = ?`, *aa.MessageID).Scan(&deleted)
switch {
case errors.Is(deletedErr, sql.ErrNoRows):
// No message row — leave ACL to decide (unlinked-shaped by now).
case deletedErr != nil:
slog.Error("failed to look up message for attachment", "id", fileID, "error", deletedErr)
writeJSON(w, http.StatusInternalServerError, errorResponse{
Error: "INTERNAL_ERROR",
Message: "internal server error",
})
return
case deleted:
http.NotFound(w, r)
return
}
}
// ── Access control ──────────────────────────────────────────────
isAdmin := role != nil && permissions.HasAdmin(role.Permissions)
// DM participation is required of everyone, including admins — this
// matches every other DM read gate in the codebase (requireChannelRead,
// PermissionService.RequireChannelAccess, checkSendPermission), none of
// which have an admin bypass. Checked ahead of the `!isAdmin` block so
// the admin bypass below cannot skip it.
if aa.ChannelID != nil && aa.ChannelType == "dm" {
if user == nil {
writeJSON(w, http.StatusForbidden, errorResponse{
Error: "FORBIDDEN",
Message: "you do not have access to this file",
})
return
}
ok, dmErr := database.IsDMParticipant(r.Context(), user.ID, *aa.ChannelID)
if dmErr != nil || !ok {
writeJSON(w, http.StatusForbidden, errorResponse{
Error: "FORBIDDEN",
Message: "you do not have access to this file",
})
return
}
}
if !isAdmin {
if aa.ChannelID == nil {
// An unlinked attachment that some user's avatar points at is
// readable by every authenticated user: an avatar has to be
// visible to the people who see the messages it sits next to.
// The check is by the exact URL the column stores, so the file
// stops being public the instant the avatar is replaced.
isAvatar, avatarErr := database.IsAvatarFileURL(r.Context(), service.AvatarFileURL(fileID))
if avatarErr != nil {
slog.Error("failed to check avatar file", "id", fileID, "error", avatarErr)
}
switch {
case isAvatar:
// Public while in use — fall through to serving.
// Unlinked attachment — only the uploader may access.
// M-2: Legacy rows (NULL uploader_id) are now denied rather than
// served to any authenticated user.
case aa.UploaderID == nil:
slog.Warn("legacy attachment access denied (NULL uploader_id)", "id", fileID)
writeJSON(w, http.StatusForbidden, errorResponse{
Error: "FORBIDDEN",
Message: "you do not have access to this file",
})
return
case user == nil || *aa.UploaderID != user.ID:
writeJSON(w, http.StatusForbidden, errorResponse{
Error: "FORBIDDEN",
Message: "you do not have access to this file",
})
return
}
} else if aa.ChannelType != "dm" {
// Linked attachment in a guild channel — check channel
// permissions. The DM case is handled unconditionally above.
if user == nil || !permSvc.HasChannelPerm(r.Context(), user.ID, *aa.ChannelID, permissions.ReadMessages) {
writeJSON(w, http.StatusForbidden, errorResponse{
Error: "FORBIDDEN",
Message: "you do not have access to this file",
})
return
}
}
}
// Open file from storage.
f, err := store.Open(aa.StoredAs)
if err != nil {
http.NotFound(w, r)
return
}
defer f.Close() //nolint:errcheck
// Set headers before ServeContent to ensure correct MIME type.
w.Header().Set("Content-Type", aa.MimeType)
// BUG-118: Force download for MIME types that could execute content
// under the OwnCord origin (HTML, SVG, XML, PDF).
disposition := "inline"
if isUnsafeInlineMIME(aa.MimeType) {
disposition = "attachment"
}
w.Header().Set("Content-Disposition", mime.FormatMediaType(disposition, map[string]string{"filename": aa.Filename}))
// These downloads are access-controlled, so they must never be stored by
// shared/proxy caches (info-leak). Mark private and force revalidation.
// W3-4: no-cache forces revalidation on every use, so a max-age is dead
// weight alongside it — private + no-cache expresses the intent exactly.
w.Header().Set("Cache-Control", "private, no-cache")
// The Access-Control-Allow-Origin header below reflects the request
// Origin, so responses vary by Origin and must not be cross-served.
w.Header().Set("Vary", "Origin")
// CORS: allow webview to read the response body using configured origins.
if origin := r.Header.Get("Origin"); origin != "" {
for _, allowed := range allowedOrigins {
if allowed == "*" || strings.EqualFold(allowed, origin) {
w.Header().Set("Access-Control-Allow-Origin", origin)
w.Header().Set("Access-Control-Expose-Headers", "Content-Type, Content-Length")
break
}
}
}
w.Header().Set("X-Content-Type-Options", "nosniff")
// Use the actual file modification time so If-Modified-Since works correctly.
var modTime time.Time
if info, statErr := f.Stat(); statErr == nil {
modTime = info.ModTime()
}
http.ServeContent(w, r, aa.Filename, modTime, f)
}
}