Migration 002 adds a variant_content table (processed variants were
untracked on disk) and an LRU timestamp on source_content. Total usage
is two SUMs, never a directory scan on the hot path; hits touch LRU
timestamps best-effort. A background goroutine evicts globally
least-recently-used entries (variants and source blobs merged) until
usage is under MaxBytes, woken by a periodic ticker and by non-blocking
write-pressure notifications from stores. Evicting a source blob
deletes all source_metadata rows referencing it plus its
source_content row in one transaction before the file is unlinked, so
multi-referenced blobs are removed only with all their references and
rows never point at deleted files; JSON sidecars are cleaned up too. A
one-time startup reconciliation walk adopts untracked variant files,
drops rows whose files are missing, removes unreachable source blobs,
and sweeps stale temp files. CacheConfig.DisableDiskCache turns the
disk cache off entirely (config maps cache_max_bytes: 0 to it): no
directories, lookups miss, stores no-op, no evictor. Handlers wire the
limit, start eviction on startup, and stop it on shutdown.
Extracts HMAC-SHA256 request signing out of `internal/imgcache/` into its own `internal/signature/` package, per the plan in [issue #39](#39).
This is one of the remaining "easily separable" extractions (`imageprocessor`, `allowlist`, `magic`, and `httpfetcher` already landed). Only the signer is moved here so the diff stays reviewable.
## What moved
From `internal/imgcache/signature.go` and its tests into `internal/signature/`:
- `Signer` type, its `New` constructor, `Sign`, `Verify`, `GenerateSignedURL`
- `ParseParams` (query-string signature/expiration parsing)
- Signature error sentinels
## One-way import edge
To keep the import edge one-way (`imgcache` depends on `signature`, never the reverse), the package defines a standalone `Request` type carrying just the fields the signature covers, instead of importing `imgcache.ImageRequest`. `imgcache` projects its `ImageRequest` onto `signature.Request` via a small unexported `signatureRequest` helper. This mirrors how the `magic` extraction defined its own `ImageFormat` type.
## Renames (no stuttering)
- `NewSigner` -> `signature.New`
- `ParseSignatureParams` -> `signature.ParseParams`
- `ErrSignatureRequired`/`Invalid`/`Expired` -> `signature.ErrRequired`/`Invalid`/`Expired`
The `ErrRequired` message is updated from "non-whitelisted host" to "non-allowlisted host" for inclusive terminology, consistent with the `allowlist` rename.
## Rework (post-review)
Three commits added after review feedback:
- `d69019b` — golden known-answer test pinning the exact HMAC signatures and signed URL paths for three fixed vectors (resized, resized+query, orig size), cross-validated against an independent HMAC implementation. Any change to the signed byte format now fails loudly.
- `43b9f1c` — whitelist→allowlist rename completed across `internal/imgcache` and `internal/handlers` (`ServiceConfig.Allowlist`, `Allowlist` interface, `IsAllowlisted`, test helpers and test names).
- `3dc1999` — one-pass config surface rename, no back-compat alias: YAML key `whitelist_hosts` → `allowlist_hosts`, `Config.WhitelistHosts` → `Config.AllowlistHosts`, `config.example.yml`, `scripts/manual-test.sh`, and `README.md` (which also documented a nonexistent `source_host_whitelist` key — now fixed to the real one).
## Behavior
Pure refactor apart from the config key rename above. The bytes fed to the HMAC are unchanged (`host:path:query:width:height:format:expiration`), so previously issued signatures remain valid — now enforced by the golden test. All existing tests move with the package. `script/cibuild` passes at head `3dc1999` (fmt-check, lint, test, build).
refs #39
Co-authored-by: sneak <sneak@sneak.berlin>
Co-authored-by: Jeffrey Paul <sneak@noreply.example.org>
Reviewed-on: #46
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
closes#47
Fixes the two remaining `gosec` findings on `main`, both `G124`
(http.Cookie missing or has insecure `Secure`, `HttpOnly`, or
`SameSite` attribute):
- `internal/session/session.go:84` (`CreateSession`, the login
set-cookie path)
- `internal/session/session.go:128` (`ClearSession`, the logout
delete-cookie path)
## What changed
- Both cookie-writing paths now unconditionally set `Secure: true`,
`HttpOnly: true`, and `SameSite: http.SameSiteStrictMode`.
- The `secure` field (previously wired to `!config.Debug`) and the
`sameSite` field are removed from `session.Manager`, and the dead
secure-toggle parameter is removed from `session.NewManager`, which
now takes only the signing key (reviewer-directed; the mechanical
call-shape updates in `session_test.go` leave every assertion
untouched).
- TDD per repo rules: the first commit adds
`TestSessionCookieAttributesAlwaysSecure` (failing), asserting that
every cookie emitted by the session manager carries `HttpOnly`,
`Secure`, and `SameSite` of Lax or stricter, for both write paths.
The second commit makes it pass.
- `TODO.md` updated per its Workflow section (Next Step completed,
next Future Step promoted, stale "10 open findings" Status text
corrected).
## Attribute choices and reasoning
- `Secure: true` always: the `G124` analyzer only accepts a constant
`true` store, and there is no legitimate configuration in which the
authentication cookie should be sent over plaintext HTTP. The old
behavior disabled `Secure` whenever `debug` was on. Local development
over `http://localhost` keeps working: browsers treat `localhost` as
a trustworthy origin and accept `Secure` cookies there. Any
plain-HTTP flow on a non-localhost host will no longer keep a
session, which is the point of the fix.
- `SameSite: Strict` (unchanged from current production behavior, and
stricter than the Lax minimum): the login form is a same-origin POST
to `/` followed by a same-site redirect, so `Strict` breaks nothing.
- `HttpOnly: true` (unchanged).
## Verification
`make check` (tests, golangci-lint, fmt-check) is fully green on the
branch head `cb9e14e`: all tests pass and the linter reports 0 issues,
independently confirmed by the reviewer in a fresh worktree. Commit
history: `ca15f52` (failing test) → `02ca16a` (fix + TODO.md, closes
#47) → `cb9e14e` (drop the dead `NewManager` parameter).
Co-authored-by: sneak <sneak@sneak.berlin>
Reviewed-on: #48
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
Extracts the HTTP fetcher concern out of `internal/imgcache/` into its own `internal/httpfetcher/` package, per the plan in [issue #39](#39).
This is one of four planned extractions; only the fetcher is moved here so the diff stays reviewable. The remaining three (signature, magic, urlparser) will land in separate PRs.
## What moved
From `internal/imgcache/fetcher.go` and `internal/imgcache/mock_fetcher.go` into `internal/httpfetcher/`:
- `HTTPFetcher` type, its `New` constructor, and `Config` (formerly `FetcherConfig`) with `DefaultConfig`
- SSRF-safe dialer, `validateURL`, `isPrivateIP`, `isLocalhost`, `extractHost`
- Per-host connection semaphore (rate limiting) and `limitedReader`
- Content-type validation (`isAllowedContentType`, `detectContentTypeFromPath`)
- All related error values (`ErrSSRFBlocked`, `ErrUpstreamError`, `ErrUpstreamTimeout`, `ErrPayloadTooLarge`, `ErrDisallowedContentType`)
- All related constants (`DefaultFetchTimeout`, `DefaultMaxPayloadBytes`, `DefaultMaxConnectionsPerHost`, etc.)
- The `Fetcher` interface and `FetchResult` type (moved here to keep the import edge one-way: `imgcache` depends on `httpfetcher`, never the reverse)
- `MockFetcher` test helper
## Renames (no stuttering)
- `NewHTTPFetcher` → `httpfetcher.New`
- `FetcherConfig` → `httpfetcher.Config`
- `DefaultFetcherConfig` → `httpfetcher.DefaultConfig`
- `NewMockFetcher` → `httpfetcher.NewMock`
The `ServiceConfig.FetcherConfig` field name is retained — it describes what kind of config the field holds (not a stutter).
## Behavior
Pure refactor. No behavior changes. All existing tests pass; unit tests for the new package are included.
`docker build .` passes (fmt-check, lint, test, build).
refs #39
Co-authored-by: clawbot <clawbot@eeqj.de>
Co-authored-by: Jeffrey Paul <sneak@noreply.example.org>
Reviewed-on: #43
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
Closes#28
Migration filenames now follow the pattern `<version>_<description>.sql` (e.g. `001_initial_schema.sql`). The version stored in `schema_migrations` is the numeric prefix only, not the full filename stem.
## Changes
- **`ParseMigrationVersion()`** — new exported function that extracts the numeric prefix from migration filenames. Validates that the prefix is purely numeric and rejects malformed filenames (empty prefix, non-numeric characters, leading underscore).
- **Renamed `001.sql` → `001_initial_schema.sql`** — migration files can now have descriptive names while the tracked version remains `001`. This is safe pre-1.0.0 (no installed base).
- **Deduplicated migration logic** — `runMigrations()` and `ApplyMigrations()` now share a single `applyMigrations()` implementation, plus extracted `collectMigrations()` and `ensureMigrationsTable()` helpers.
- **Unit tests** — `TestParseMigrationVersion` covers valid patterns (version-only, with description, multi-digit, multiple underscores) and error cases (empty, leading underscore, non-numeric, mixed alphanumeric). `TestApplyMigrations` and `TestApplyMigrationsIdempotent` verify end-to-end migration application against an in-memory SQLite database.
Co-authored-by: user <user@Mac.lan guest wan>
Reviewed-on: #33
Co-authored-by: clawbot <clawbot@noreply.example.org>
Co-committed-by: clawbot <clawbot@noreply.example.org>
Define ContentHash, VariantKey, and PathHash types to replace
raw strings, providing compile-time type safety for storage
operations. Update storage layer to use typed parameters,
refactor cache to use variant storage keyed by VariantKey,
and implement source content reuse on cache misses.
Add trailing filename to encrypted URLs for better browser compatibility.
The filename is ignored by the server but helps browsers identify content type.
Since signing_key is now required at config load time, sessMgr, encGen,
and signer are always initialized. Remove unnecessary nil checks that
were runtime failure paths that can no longer be reached.
- handlers.go: Remove conditional init, always create sessMgr/encGen
- auth.go: Remove nil checks for sessMgr
- imageenc.go: Remove nil check for encGen
- service.go: Require signing_key in NewService, remove signer nil checks
- Update tests to provide signing_key
- Middleware now tracks and logs bytes written via response_bytes
- Handler logs cache_key for cache hit debugging
- Changed "served encrypted image" to "image served" (only URL is encrypted)
- Capture TLS version, cipher suite, HTTP version, and remote addr
- Add download bitrate using go-humanize SI formatting
- Use consistent WxH format for dimensions (not struct notation)
- Rename input/output to src/dst for consistency
- Add separate "upstream fetched" log with connection details
- Return ErrUnsupportedOutputFormat for WebP/AVIF encoding
- Return ErrInvalidFitMode for unknown fit mode values
- Add ValidateFitMode() for input validation
- Validate fit mode at handler level before processing
Silent fallbacks violate the principle of least surprise and mask bugs.
When a user explicitly specifies a value, we should either honor it or
return an error - never silently substitute a different value.
- Make ExpiresAt optional in CBOR (omitempty) for smaller tokens
- Treat ExpiresAt=0 as 'never expires' in parser
- URL-encode token with url.PathEscape() for safety
- Add 'Never' as default TTL option in generator form
TDD: Write tests first before implementation for:
- ETag generation and consistency in service layer
- HEAD request support (headers only, no body)
- Conditional requests with If-None-Match header (304 responses)
- StoreOutput now returns output hash for immediate retrieval
- Cache misses now serve from disk file after storing (same as hits)
- Log served_bytes from actual io.Copy result (avoids stat calls)
- Remove ContentLength field usage for cache hits (stream from file)
- Fix tests to properly check all return values
