Refuse image requests whose Referer is on referer_blocklist (closes #90)

A new setting, referer_blocklist (PIXA_REFERER_BLOCKLIST), lists hosts
whose pages may not show pixa's images. Entries are written and matched
as allowlist_hosts are, with the same matcher. Both image routes check
the Referer before the signature, the cache and the upstream fetch, and
answer 403 with the JSON error, so a blocked request costs nothing and
is refused whether or not the image is cached. No Referer, or one that
does not parse, is served; README.md says this makes the list easy to
get around. An entry of either host list that is not a host name
(letters, digits, hyphens, underscores, dots, at most one leading dot)
or an IP address now aborts startup naming the setting and the entry.

Model: opus-5-5

.gitattributes

@@ -0,0 +1,4 @@
+# Every PR adds an entry at the top of TODO.md's Completed Steps; union keeps
+# both sides instead of conflicting. Git never reports a conflict here: read
+# the merged entries after every merge or rebase.
+TODO.md merge=union

.golangci.yml

@@ -10,14 +10,20 @@ run:
 
 linters:
   default: all
+  enable:
+    # Successor to the deprecated gomodguard. Named explicitly, rather than
+    # left to `default: all`, because it carries the module policy below.
+    - gomodguard_v2
   disable:
     # Genuinely incompatible with project patterns
     - exhaustruct      # Requires all struct fields
-    - depguard         # Dependency allow/block lists
     - godot            # Requires comments to end with periods
-    - wsl              # Deprecated, replaced by wsl_v5
     - wrapcheck        # Too verbose for internal packages
     - varnamelen       # Short names like db, id are idiomatic Go
+    # Deprecated: the warning is attached to the old name, so it is
+    # silenced by disabling that name, not by enabling the successor.
+    - wsl              # Deprecated, replaced by wsl_v5
+    - gomodguard       # Deprecated, replaced by gomodguard_v2
   settings:
     lll:
       line-length: 88
@@ -28,6 +34,64 @@ linters:
       max-complexity: 15
     dupl:
       threshold: 100
+    depguard:
+      # Test-support code must not be compiled into the shipped binary. A
+      # test-support package exists to hand a test privileges the program
+      # itself must never have, so a file that is not a test must not import
+      # one. Test files, and the files inside a package whose directory name
+      # ends in `test`, are where that code belongs, and are exempt.
+      #
+      # The deny list below is the one part of this file a repository is
+      # expected to extend, and the only part it may. depguard matches an
+      # import path against a list of prefixes, so it cannot be told "any path
+      # whose last segment ends in test"; a repository's own test-support
+      # packages have to be named here one at a time, by full import path,
+      # under a module path that differs from repository to repository. Add
+      # them; change nothing else.
+      rules:
+        test-support:
+          list-mode: lax
+          files:
+            - "$all"
+            - "!$test"
+            - "!**/*test/**"
+          deny:
+            - pkg: net/http/httptest
+              desc: >-
+                Test-support code belongs in test files and in packages whose
+                directory name ends in test, not in the shipped binary.
+    # Only decisions already recorded in the Go package defaults are
+    # listed here. Every entry matches the module path exactly.
+    gomodguard_v2:
+      blocked:
+        - module: github.com/rs/zerolog
+          recommendations:
+            - log/slog
+          reason: "Structured logging is stdlib log/slog."
+        # One entry per pre-fork module path, because the later releases
+        # are separate paths. A prefix match would be shorter but would
+        # also reach github.com/go-redis/redismock, the test double for
+        # the successor these entries recommend.
+        - module: github.com/go-redis/redis
+          recommendations:
+            - github.com/redis/go-redis/v9
+          reason: "Pre-fork module; use the maintained go-redis v9."
+        - module: github.com/go-redis/redis/v7
+          recommendations:
+            - github.com/redis/go-redis/v9
+          reason: "Pre-fork module; use the maintained go-redis v9."
+        - module: github.com/go-redis/redis/v8
+          recommendations:
+            - github.com/redis/go-redis/v9
+          reason: "Pre-fork module; use the maintained go-redis v9."
+        - module: github.com/sergi/go-diff
+          recommendations:
+            - github.com/aymanbagabas/go-udiff
+          reason: "No unified diff output; use go-udiff."
+        - module: github.com/hexops/gotextdiff
+          recommendations:
+            - github.com/aymanbagabas/go-udiff
+          reason: "Unmaintained fork; use go-udiff."
 
 issues:
   max-issues-per-linter: 0

README.md

@@ -18,7 +18,7 @@ make build
 # run with a config file: copy the example and set a real signing key
 # (the example placeholder is refused at startup), e.g. with
 # openssl rand -base64 32
-cp config.example.yml config.yml
+cp configs/config.example.yml config.yml
 $EDITOR config.yml   # replace the signing_key placeholder
 ./bin/pixad --config config.yml
 
@@ -34,6 +34,67 @@ else has a built-in default. A config file mounted at `/etc/pixa/config.yml`
 is optional: it is read when present, and an environment variable wins over
 the same setting in it.
 
+## Deployment
+
+pixa listens on plain HTTP and runs behind a reverse proxy that terminates TLS.
+[`configs/Caddyfile`](configs/Caddyfile) is an example for Caddy, chosen because
+it is the smallest correct one: Caddy gets the TLS certificate itself and does
+everything in this list without further settings. The reverse proxy must:
+
+- terminate TLS, as the login and generator pages work only over HTTPS (see
+  Routes);
+- pass the `Host`, `Origin` and `Referer` headers on unchanged, as pixa refuses
+  a form from those pages unless `Origin` or `Referer` names the host in `Host`,
+  builds encrypted URLs from `Host`, and checks `Referer` against
+  `referer_blocklist`;
+- set `X-Forwarded-For` to the client's address, with `trusted_proxies` set to
+  the address pixa sees the proxy's requests come from, so the login limit
+  counts each client by its own address (see `trusted_proxies` under
+  Configuration);
+- wait for pixa's answer for at least `downstream_timeout` (default `60s`), the
+  longest pixa takes to fetch, convert and send an image.
+
+It may also refuse `/metrics`, as the example does, so that only a scraper that
+reaches pixa directly can read it; pixa itself asks for the metrics username and
+password there.
+
+pixa does the rest itself: it checks signatures and encrypted URLs, applies the
+allowlist, refuses upstream hosts with private or local addresses, limits login
+attempts, upstream response size and image dimensions, and sends the security
+headers, `Strict-Transport-Security` included, with every response.
+
+The state directory (`state_dir`, `/var/lib/pixa` in the container) holds the
+database and the disk cache:
+
+- It needs a persistent volume: without one, every restart starts with an empty
+  cache. In the container, the startup script gives the directory to the user
+  pixa runs as (uid 65532) and sets its mode to `750`; outside it, that user
+  must be able to write the directory.
+- `cache_max_bytes` limits the source and transformed images together. The
+  database, the metadata files, the `.meta` file beside each transformed image
+  and files still being written come on top, and eviction runs in the
+  background, so the cache can pass the limit for a while: leave room on the
+  volume beyond it.
+- Set `cache_max_bytes` for a lasting deployment. Its default, worked out each
+  time pixa starts, is 75% of the sum of the space free on the volume and the
+  space the cached images already take, so a restart keeps the limit the cache
+  had, but anything else that fills or frees space on the volume moves it.
+
+A load balancer's health check can request `/.well-known/healthcheck.json`,
+which answers 200 whenever pixa is running, in maintenance mode too (see
+`maintenance_mode`).
+
+On SIGTERM or SIGINT pixa stops accepting connections, gives the requests in
+progress and the images being processed 5 seconds to finish, and exits: with 0,
+or with 1 when images were still being processed after those 5 seconds or
+another part of pixa failed to stop. A request not finished by then is cut off.
+`docker stop` waits 10 seconds before it kills the container.
+
+Outside Docker, pixa needs libvips (the image has 8.15) and libheif to run, as
+it uses libvips through CGO; building it also needs their development files,
+`pkg-config` and a C compiler. `script/bootstrap` installs all of these with
+nix, apt, brew or apk.
+
 ## Running under upaas
 
 What the [upaas](https://git.eeqj.de/sneak/upaas) app for pixa needs:
@@ -49,7 +110,7 @@ What the [upaas](https://git.eeqj.de/sneak/upaas) app for pixa needs:
     - `PIXA_ALLOWLIST_HOSTS`: upstream hosts served without a signature,
       comma-separated
     - `PIXA_CACHE_MAX_BYTES`: disk cache limit in bytes; `0` disables it;
-      default 75% of free space
+      default 75% of (free space + what the cache holds)
     - the rest are in the table under Configuration below
 - **Health check:** the image's `HEALTHCHECK` requests
   `/.well-known/healthcheck.json`. upaas reads the container's health 60
@@ -115,11 +176,13 @@ path under `/v1/` answers 200, in maintenance mode too.
   allowlisted (see Source Hosts). Answers: 200; 304 when `If-None-Match` matches
   the image's `ETag`; 400 for a URL or parameter that is not valid; 401 for a
   missing or wrong signature, a missing `exp` or an `exp` in the past; 403 when
-  the upstream host, or a host it redirects to, is `localhost`, ends in
-  `.localhost` or `.local`, or has an address in a blocked network (see
-  `blocked_networks`); 502 when the upstream answered with an error status, and
-  for 5 minutes after that for the same source URL; 503 when pixa is busy or in
-  maintenance mode; 500 for any other failure.
+  the request's `Referer` names a host in `referer_blocklist`, checked before
+  the signature, the cache and the upstream fetch; 403 when the upstream host,
+  or a host it redirects to, is `localhost`, ends in `.localhost` or `.local`,
+  or has an address in a blocked network (see `blocked_networks`); 502 when the
+  upstream answered with an error status, and for 5 minutes after that for the
+  same source URL; 503 when pixa is busy or in maintenance mode; 500 for any
+  other failure.
 - `GET` or `HEAD` `/v1/e/<token>/<name>` — an image through an encrypted URL
   (see Encrypted URLs). Needs: nothing but the URL. Answers: 200; 304 when
   `If-None-Match` matches the image's `ETag`; 400 for a token that does not
@@ -132,8 +195,9 @@ path under `/v1/` answers 200, in maintenance mode too.
 - `GET /.well-known/healthcheck.json` — JSON with `status` (`ok`), `now`,
   `uptime_seconds`, `uptime_human`, `version`, `appname` and
   `maintenance_mode`. Needs: nothing. Answers: 200, always.
-- `GET /static/<file>` — the script the login and generator pages load. Needs:
-  nothing. Answers: 200, or 404 for a file that does not exist.
+- `GET /static/<file>` — the stylesheet and script the login and generator
+  pages load. Needs: nothing. Answers: 200, or 404 for a file that does not
+  exist.
 - `GET /metrics` — Prometheus metrics (see Architecture). Needs: HTTP basic
   authentication with `metrics.username` and `metrics.password`. Answers: 200;
   401 without them; 404 when they are not set, as the route then does not exist.
@@ -331,6 +395,11 @@ and the URL is
 - **Suffix match**: `.example.com` — matches `cdn.example.com`,
   `images.example.com`, and `example.com`
 
+An IP address is matched exactly; write an IPv6 address without brackets. An
+entry that is neither a host name (letters, digits, hyphens, underscores and
+dots, with at most one leading dot) nor an IP address, such as one with a port
+or a `*.` wildcard, aborts startup.
+
 ### Configuration
 
 Every setting can be given as an environment variable, in a YAML config
@@ -365,8 +434,9 @@ tell. With no file, pixa uses the environment and the defaults.
 | `PORT`                               | `port`                          | Port to listen on; default `8080`                                            |
 | `PIXA_STATE_DIR`                     | `state_dir`                     | Directory for the database and the disk cache; default `/var/lib/pixa`       |
 | `PIXA_DB_URL`                        | `db_url`                        | SQLite database URL; default `state.sqlite3` in the state directory          |
-| `PIXA_CACHE_MAX_BYTES`               | `cache_max_bytes`               | Disk cache limit in bytes; `0` disables it; default 75% of free space        |
+| `PIXA_CACHE_MAX_BYTES`               | `cache_max_bytes`               | Disk cache limit in bytes; `0` disables it; default 75% of (free + cached)   |
 | `PIXA_ALLOWLIST_HOSTS`               | `allowlist_hosts`               | Upstream hosts served without a signature                                    |
+| `PIXA_REFERER_BLOCKLIST`             | `referer_blocklist`             | Hosts whose pages the image routes refuse with 403, by `Referer`             |
 | `PIXA_BLOCKED_NETWORKS`              | `blocked_networks`              | CIDR ranges never fetched from, on top of the built-in ones                  |
 | `PIXA_TRUSTED_PROXIES`               | `trusted_proxies`               | CIDR ranges of proxies whose `X-Forwarded-For` is believed; default RFC 1918 |
 | `PIXA_ALLOW_HTTP`                    | `allow_http`                    | Allow plain-HTTP upstreams, for testing only; default `false`                |
@@ -395,6 +465,18 @@ Key settings in more detail:
   that has no leading zero and is not the scheme's default. Any other value,
   including another scheme such as a browser extension's, aborts startup
 - `allowlist_hosts` — list of allowed upstream hosts
+- `referer_blocklist` — list of hosts whose pages may not show pixa's images, to
+  stop other sites hotlinking them. Entries are written and matched as for
+  `allowlist_hosts` (see Allowlist patterns), and an entry that is neither a
+  host name nor an IP address aborts startup. A request to `/v1/image/` or
+  `/v1/e/` whose `Referer` header names a listed host is refused with 403 before
+  its signature or token is checked and before the cache or the upstream host is
+  used, so it fetches nothing, and it is refused even when the image is cached.
+  A request with no `Referer`, or one that does not parse as a URL
+  with a host, is served, as many clients send none. So this is easily got
+  around: a site whose pages send no `Referer` (for example with
+  `Referrer-Policy: no-referrer`) is not stopped. It does not apply to the login
+  and generator pages. Default: empty
 - `blocked_networks` — list of CIDR ranges to refuse for SSRF protection,
   added to the always-enforced built-in ranges (loopback, private,
   link-local, CGNAT, benchmark, NAT64, and the like); an invalid CIDR
@@ -429,9 +511,17 @@ Key settings in more detail:
   waits for an upstream connection and for a processing slot (up to 10 seconds
   each), so keep it longer than `upstream_fetch_timeout` plus 20 seconds
 - `signing_key` — HMAC secret for URL signatures
+- `db_url` — the SQLite database to open; omitted, it is
+  `file:<state_dir>/state.sqlite3?_pragma=journal_mode(WAL)`, which keeps the
+  database in WAL mode. pixa adds `_pragma=busy_timeout(5000)` to any `db_url`,
+  so a write that finds another in progress waits up to five seconds for it
+  instead of failing. WAL mode comes only from the URL: keep
+  `_pragma=journal_mode(WAL)` in one you set
 - `cache_max_bytes` — disk cache size limit in bytes; `0` disables the
-  disk cache entirely; omitted defaults to 75% of the free space on
-  the filesystem containing `<state_dir>/cache/` (minimum 500 MiB)
+  disk cache entirely; omitted defaults to 75% of the sum of the free space on
+  the filesystem containing `<state_dir>/cache/` and the bytes of source and
+  transformed images the cache already holds, worked out at startup (minimum
+  500 MiB)
 - `upstream_connections` — the most connections to upstream hosts at once, all
   hosts together, on top of `upstream_connections_per_host`; default `64`. A
   fetch holds its connection until its image has been processed. A fetch that
@@ -451,7 +541,7 @@ Key settings in more detail:
   the container unhealthy, and upaas marks a deploy failed when its container
   is unhealthy. The login and URL generator pages and `/metrics` keep working
 
-See `config.example.yml` for all options with defaults.
+See `configs/config.example.yml` for all options with defaults.
 
 ### Architecture
 

TODO.md

@@ -3,6 +3,8 @@
 * branch per issue from `next`
 * do the work in Next Step
 * move Next Step to the top of Completed Steps
+* `TODO.md` merges with git's union merge (`.gitattributes`), which never
+  reports a conflict: read the merged entries after every merge or rebase
 * move the top item of Future Steps into Next Step
 * commit (`TODO.md` changes in the same commit as the work)
 * open a PR based on `next`
@@ -25,15 +27,110 @@ The disk cache is now size-bounded with LRU eviction
 
 # Next Step
 
-P2: security: referer blacklist
+P2: security: per-IP rate limiting on the image routes
 
 # Completed Steps
 
+- 2026-10-04 referer blocklist (closes #90): `referer_blocklist`
+  (`PIXA_REFERER_BLOCKLIST`) lists hosts, written and matched as for
+  `allowlist_hosts` with the same matcher; an entry of either list that is
+  neither a host name (letters, digits, hyphens, underscores and dots, with at
+  most one leading dot) nor an IP address, such as one with a port or a `*.`
+  wildcard, aborts startup naming the setting and the entry. Both image routes
+  refuse a request whose `Referer` names a listed host with 403 and a JSON error
+  before the signature, the cache and the upstream fetch, so it fetches nothing
+  and is refused whether or not the image is cached. A request with no
+  `Referer`, or one that does not parse as a URL with a host, is served, so the
+  list is easily got around; `README.md` and `configs/config.example.yml` say
+  so. It does not apply to the login and generator pages.
+- 2026-10-04 fewer files in the repository root (closes #97):
+  `config.example.yml` moved unchanged to `configs/config.example.yml`, and
+  `README.md`, the comments in `internal/config/config.go` and the startup error
+  for the placeholder signing key name the new path; `scripts/manual-test.sh`
+  and its directory are deleted, as the handler tests in `internal/handlers`
+  cover every check it made except two: fetching a real image from the
+  internet, and a URL made on the generator page with a `ttl` answering 410 once
+  the `ttl` has passed (https://git.eeqj.de/sneak/pixa/issues/199);
+  `CONVENTIONS.md` is deleted, as `REPO_POLICIES.md` links the canonical Go HTTP
+  server conventions.
+- 2026-10-04 SQLite writes no longer fail with "database is locked" (closes
+  #198): pixa adds `_pragma=busy_timeout(5000)` to every `db_url`, so a write
+  that finds another in progress on another connection waits up to five seconds
+  for it, and the default `db_url` turns on WAL mode with
+  `_pragma=journal_mode(WAL)`. The old default's `_journal_mode=WAL` is not a
+  parameter the driver reads, so the database was never in WAL mode.
+- 2026-10-04 `TestPeriodicReconciliationAdoptsFileThatAppearsAfterStartup`
+  only passes through a periodic pass (closes #189): it slept for three
+  eviction intervals before writing its file, and a startup pass still running
+  then could adopt the file itself. It now holds the test database's only
+  connection until the startup pass waits for it after walking the empty
+  variant directory, writes the file and lets the connection go, as
+  `TestEvictionRunsOnPeriodicSchedule` does, so only a periodic reconciliation
+  pass can adopt the file. Test only.
+- 2026-10-04 logging in, logging out, the URL generator and `/v1/e/` have
+  handler tests (closes #77): new tests in `internal/handlers`, with no
+  network, check that `GET /` without a login session shows the login form; a
+  wrong key shows it again with an error and sets no session cookie; the right
+  key answers 303 to `/` with a session cookie marked `Secure`, `HttpOnly` and
+  `SameSite=Strict`, with which `GET /` shows the generator page; `GET /logout`
+  answers 303 to `/` with an empty session cookie sent with `Max-Age=0`;
+  `POST /generate` without a login session answers 303 to `/`; `/v1/e/` serves
+  the image for a valid token, answers 410 for an expired one and 400 for one
+  with a character changed, cut short or made with another signing key; and a
+  URL made on the generator page is served by `/v1/e/`. No code changes.
+- 2026-10-04 `TODO.md` merges with git's union merge (closes #190): a root
+  `.gitattributes`, copied from `sneak/prompts`, marks it `merge=union`, so two
+  branches that each add an entry at the top of Completed Steps merge without a
+  conflict and keep both entries. Git now never reports a conflict in
+  `TODO.md`: a real one keeps both versions of the lines, and two entries that
+  share an identical line can end up one inside the other, which a rebase can
+  do to an entry already on `next`. The Workflow above says to read the merged
+  entries after every merge or rebase.
+- 2026-10-04 the default `cache_max_bytes` no longer shrinks as the cache fills
+  (closes #184): for an omitted key, the cache works out the limit when it
+  opens, after the database is open, as 75% of the sum of the free space on the
+  filesystem containing `<state_dir>/cache/` and what the cache already holds by
+  its own size accounting, at least 500 MiB, so a cache filled to its limit
+  keeps that limit across a restart. The computation and its tests moved from
+  `internal/config` to `internal/imgcache`; the config only records whether the
+  key was set.
+- 2026-10-04 `TestEvictionRunsOnPeriodicSchedule` no longer races the evictor
+  (closes #183): it wrote each variant file and then inserted its accounting row
+  by hand, and a reconciliation pass between the two adopted the file first, so
+  the insert failed. It now writes the files only, while holding the test
+  database's only connection so the evictor's startup pass waits after walking
+  the empty variant directory; a periodic reconciliation pass then adopts the
+  files and the eviction pass after it evicts them. No other test in
+  `internal/imgcache` inserts a row by hand after starting the evictor. Test
+  only.
 - 2026-10-04 a config file pixa cannot read aborts startup (closes #176): of the
   places pixa looks for its config file on its own, only one where the file does
   not exist is passed over; any other error, such as a directory on the path
   that pixa may not enter, aborts startup naming the file, as a file that does
   not parse already did.
+- 2026-10-04 `.golangci.yml` re-vendored from the canonical copy (closes #57):
+  the deprecated `gomodguard` is switched off, so lint runs print no
+  deprecation warning; its successor `gomodguard_v2` runs with the shared
+  module block list, and `depguard` keeps `net/http/httptest` out of files that
+  are not tests. The tree needed no code changes.
+- 2026-10-04 the Content-Security-Policy allows no inline script or style
+  (closes #125): `script-src` and `style-src` are `'self'` only. The generator
+  page's two inline `onclick` handlers moved into
+  `internal/static/generator.js`, attached with `addEventListener`; the bundled
+  Tailwind script, which built styles in the browser, is replaced by a small
+  hand-written `internal/static/style.css` with only the rules the login and
+  generator pages use, the templates carrying a few plain class names in place
+  of Tailwind's. No build step. The pages keep their layout, not every pixel of
+  it.
+- 2026-10-04 deployment guide and example Caddy config (closes #89):
+  "Deployment" in `README.md` says what the reverse proxy in front of pixa must
+  do (terminate TLS; pass `Host`, `Origin` and `Referer` on unchanged; set
+  `X-Forwarded-For`, with `trusted_proxies` to match; wait at least
+  `downstream_timeout`; optionally refuse `/metrics`) and what pixa does itself,
+  that the state directory needs a persistent volume and what `cache_max_bytes`
+  counts, the health check for a load balancer, what a stop does and its exit
+  codes, and what running outside Docker needs; `configs/Caddyfile` is the
+  example, checked with `caddy validate`.
 - 2026-10-04 the metrics basic auth, CORS preflight, request logging and
   metrics recording have tests (closes #79): `MetricsAuth` on its own answers
   401 with a challenge without credentials or with a wrong username or password
@@ -476,7 +573,6 @@ P2: security: referer blacklist
 # Future Steps
 
 - P2: security
-    - per-IP rate limiting on the image routes
     - per-origin rate limiting
 - P2: HTTP response handling
     - Last-Modified headers
@@ -490,6 +586,3 @@ P2: security: referer blacklist
     - Prometheus performance metrics
     - integration tests for the image proxy flow
     - load tests to verify the 1k to 5k req/s target
-- P2: documentation
-    - deployment guide
-    - example nginx or caddy reverse proxy config

configs/Caddyfile

@@ -0,0 +1,17 @@
+# Example Caddy config for running pixa behind Caddy; see "Deployment" in
+# README.md. Replace images.example.com with pixa's public host name, and
+# 127.0.0.1:8080 with the address Caddy reaches pixa on.
+#
+# Caddy gets and renews the TLS certificate for the host name, passes the
+# Host, Origin and Referer headers on unchanged, sets X-Forwarded-For to the
+# client's address, and waits for pixa's answer with no time limit of its
+# own, so pixa's downstream_timeout is what ends a slow request.
+
+images.example.com
+
+# pixa asks for metrics.username and metrics.password on /metrics. This
+# line also keeps it off the public address, for a scraper that reaches
+# pixa directly; remove it to read /metrics through Caddy.
+respond /metrics 404
+
+reverse_proxy 127.0.0.1:8080

configs/config.example.yml

@@ -34,9 +34,10 @@ maintenance_mode: false
 state_dir: ./data
 
 # SQLite database URL (default:
-# file:<state_dir>/state.sqlite3?_journal_mode=WAL). An empty value aborts
-# startup; leave the key out to use the default.
-# db_url: "file:./data/state.sqlite3?_journal_mode=WAL"
+# file:<state_dir>/state.sqlite3?_pragma=journal_mode(WAL)). pixa adds
+# _pragma=busy_timeout(5000) to it. An empty value aborts startup; leave the
+# key out to use the default.
+# db_url: "file:./data/state.sqlite3?_pragma=journal_mode(WAL)"
 
 # Image proxy settings
 # HMAC signing key for URL signatures (required, at least 32 characters)
@@ -45,6 +46,8 @@ signing_key: "CHANGE_ME_generate_with_openssl_rand_base64_32"
 
 # Hosts that don't require signatures (default: none)
 # Use "." prefix for wildcard subdomain matching (e.g., ".example.com" matches "cdn.example.com")
+# An entry that is neither a host name nor an IP address (IPv6 without
+# brackets), such as one with a port or a "*." wildcard, aborts startup.
 allowlist_hosts:
   - s3.sneak.cloud
   - static.sneak.cloud
@@ -52,6 +55,16 @@ allowlist_hosts:
   - github.com
   - user-images.githubusercontent.com
 
+# Hosts whose pages may not show pixa's images, written as for
+# allowlist_hosts. A request to /v1/image/ or /v1/e/ whose Referer header
+# names one of them is answered 403 before anything is fetched, even when
+# the image is cached. A request with no Referer, or one that does not
+# parse, is served, so a site whose pages send no Referer is not stopped.
+# The login and generator pages are not covered. (default: none)
+# referer_blocklist:
+#   - leech.example
+#   - .hotlinker.example
+
 # Additional CIDR ranges to refuse when fetching upstream, extending the
 # SSRF protection. These are added to the always-enforced built-in ranges
 # (loopback, RFC 1918 private, link-local, CGNAT, benchmark, NAT64, and
@@ -128,8 +141,9 @@ access_control_allow_origin: "*"
 
 # Maximum disk cache size in bytes. Explicit values are used exactly as
 # given; 0 disables the disk cache entirely (every request fetches and
-# processes uncached). When omitted, the default is 75% of the free
-# space on the filesystem containing <state_dir>/cache/ at startup,
+# processes uncached). When omitted, the default is 75% of the sum of
+# the free space on the filesystem containing <state_dir>/cache/ and
+# the bytes of images the cache already holds, worked out at startup,
 # with a minimum of 500 MiB.
 # cache_max_bytes: 10737418240
 

internal/config/cache_max_bytes_internal_test.go

@@ -1,26 +1,10 @@
 package config
 
 import (
-	"errors"
-	"log/slog"
-	"os"
-	"path/filepath"
 	"strings"
 	"testing"
 )
 
-// Static errors returned by the stub free-space probes below.
-var (
-	errTestStatfsFailed     = errors.New("statfs failed")
-	errTestProbeNotExpected = errors.New("probe must not be called")
-)
-
-// discardLogger returns a logger that swallows all output, for tests
-// that exercise code paths which log.
-func discardLogger() *slog.Logger {
-	return slog.New(slog.DiscardHandler)
-}
-
 // TestCacheMaxBytesExplicitValueUsedWithoutFloor verifies that an
 // explicitly configured cache_max_bytes value is used exactly as
 // given: the 500 MiB floor applies only to the computed default, never
@@ -151,155 +135,30 @@ func TestCacheMaxBytesInvalidValuesAbortStartup(t *testing.T) {
 	}
 }
 
-// TestComputeDefaultCacheMaxBytesUses75PercentOfFreeSpace verifies the
-// computed default is 75% of the probed free space when that exceeds
-// the floor.
-func TestComputeDefaultCacheMaxBytesUses75PercentOfFreeSpace(t *testing.T) {
+// TestCacheMaxBytesExplicitIsRecorded verifies that an omitted
+// cache_max_bytes is recorded as not explicit, so the cache works out
+// the default when it opens, and that an explicit zero is recorded as
+// explicit, so it disables the disk cache instead.
+func TestCacheMaxBytesExplicitIsRecorded(t *testing.T) {
 	t.Parallel()
 
-	// 4 GiB free -> 3 GiB default.
-	probe := func(string) (uint64, error) { return 4294967296, nil }
+	signingKeyLine := "signing_key: " + validTestSigningKey + "\n"
 
-	got, err := ComputeDefaultCacheMaxBytes(t.TempDir(), probe)
-	if err != nil {
-		t.Fatalf("ComputeDefaultCacheMaxBytes returned error: %v", err)
-	}
-
-	if got != 3221225472 {
-		t.Errorf("ComputeDefaultCacheMaxBytes = %d, want 3221225472 (75%% of 4 GiB)",
-			got)
-	}
-}
-
-// TestComputeDefaultCacheMaxBytesAppliesFloorToComputedDefault
-// verifies that when 75% of free space is below 500 MiB, the computed
-// default is floored at DefaultCacheMaxBytesFloor.
-func TestComputeDefaultCacheMaxBytesAppliesFloorToComputedDefault(t *testing.T) {
-	t.Parallel()
-
-	cases := []struct {
-		name      string
-		freeBytes uint64
-	}{
-		{name: "100 MiB free", freeBytes: 104857600},
-		{name: "zero free", freeBytes: 0},
-		{name: "just below floor threshold", freeBytes: 699050665},
-	}
-
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			t.Parallel()
-
-			probe := func(string) (uint64, error) { return tc.freeBytes, nil }
-
-			got, err := ComputeDefaultCacheMaxBytes(t.TempDir(), probe)
-			if err != nil {
-				t.Fatalf("ComputeDefaultCacheMaxBytes returned error: %v", err)
-			}
-
-			if got != DefaultCacheMaxBytesFloor {
-				t.Errorf("ComputeDefaultCacheMaxBytes = %d, want floor %d",
-					got, DefaultCacheMaxBytesFloor)
-			}
-		})
-	}
-}
-
-// TestComputeDefaultCacheMaxBytesPropagatesProbeError verifies that a
-// failing free-space probe produces an error naming the config key,
-// instead of a silently wrong default.
-func TestComputeDefaultCacheMaxBytesPropagatesProbeError(t *testing.T) {
-	t.Parallel()
-
-	probe := func(string) (uint64, error) { return 0, errTestStatfsFailed }
-
-	_, err := ComputeDefaultCacheMaxBytes(t.TempDir(), probe)
-	if err == nil {
-		t.Fatal("probe failure must produce an error, got nil")
-	}
-
-	t.Logf("got expected error: %v", err)
-
-	if !strings.Contains(err.Error(), keyCacheMaxBytes) {
-		t.Errorf("error %q does not name the config key cache_max_bytes", err.Error())
-	}
-}
-
-// TestResolveCacheMaxBytesComputesDefaultWhenOmitted verifies that an
-// omitted cache_max_bytes key resolves to the computed default, that
-// the probe is pointed at <state_dir>/cache/ (which must be created
-// first so statfs measures the right filesystem), and that the result
-// lands on the Config.
-func TestResolveCacheMaxBytesComputesDefaultWhenOmitted(t *testing.T) {
-	t.Parallel()
-
-	c, err := configFromYAML(t, "signing_key: "+validTestSigningKey+"\n")
+	omitted, err := configFromYAML(t, signingKeyLine)
 	if err != nil {
 		t.Fatalf("minimal config should be valid, got error: %v", err)
 	}
 
-	c.StateDir = t.TempDir()
-	wantCacheDir := filepath.Join(c.StateDir, "cache")
-
-	var probedPath string
-
-	// 4 GiB free -> 3 GiB default.
-	probe := func(path string) (uint64, error) {
-		probedPath = path
-
-		return 4294967296, nil
+	if omitted.CacheMaxBytesExplicit {
+		t.Error("omitted cache_max_bytes recorded as explicit")
 	}
 
-	err = c.resolveCacheMaxBytes(discardLogger(), probe)
+	zero, err := configFromYAML(t, signingKeyLine+"cache_max_bytes: 0\n")
 	if err != nil {
-		t.Fatalf("resolveCacheMaxBytes returned error: %v", err)
+		t.Fatalf("cache_max_bytes: 0 must be accepted, got error: %v", err)
 	}
 
-	if c.CacheMaxBytes != 3221225472 {
-		t.Errorf("CacheMaxBytes = %d, want computed default 3221225472",
-			c.CacheMaxBytes)
-	}
-
-	if probedPath != wantCacheDir {
-		t.Errorf("free space probed at %q, want cache directory %q",
-			probedPath, wantCacheDir)
-	}
-
-	info, err := os.Stat(wantCacheDir)
-	if err != nil || !info.IsDir() {
-		t.Errorf("cache directory %q was not created before probing: info=%v err=%v",
-			wantCacheDir, info, err)
-	}
-}
-
-// TestResolveCacheMaxBytesDoesNotOverrideExplicitValue verifies that
-// an explicitly configured value survives resolution untouched and
-// that the free-space probe is never consulted for it.
-func TestResolveCacheMaxBytesDoesNotOverrideExplicitValue(t *testing.T) {
-	t.Parallel()
-
-	yamlContent := "signing_key: " + validTestSigningKey + "\ncache_max_bytes: 1024\n"
-
-	c, err := configFromYAML(t, yamlContent)
-	if err != nil {
-		t.Fatalf("explicit cache_max_bytes must be accepted, got error: %v", err)
-	}
-
-	c.StateDir = t.TempDir()
-
-	probe := func(string) (uint64, error) {
-		t.Error("free-space probe must not be consulted for explicit values")
-
-		return 0, errTestProbeNotExpected
-	}
-
-	err = c.resolveCacheMaxBytes(discardLogger(), probe)
-	if err != nil {
-		t.Fatalf("resolveCacheMaxBytes returned error: %v", err)
-	}
-
-	if c.CacheMaxBytes != 1024 {
-		t.Errorf("CacheMaxBytes = %d, want explicit 1024 (no floor, no recompute)",
-			c.CacheMaxBytes)
+	if !zero.CacheMaxBytesExplicit {
+		t.Error("cache_max_bytes: 0 not recorded as explicit")
 	}
 }

internal/config/cachesize.go

@@ -1,116 +0,0 @@
-package config
-
-import (
-	"fmt"
-	"log/slog"
-	"math"
-	"os"
-	"path/filepath"
-	"syscall"
-)
-
-// DefaultCacheMaxBytesFloor is the minimum computed default for the
-// cache_max_bytes setting: 500 MiB. The floor applies only to the
-// computed default (when the key is omitted from the configuration),
-// never to explicitly configured values.
-const DefaultCacheMaxBytesFloor int64 = 524288000
-
-// cacheDirPerms is the permission mode for the cache directory created
-// before probing free space, matching the state directory permissions.
-const cacheDirPerms = 0o750
-
-// freeSpaceFractionNumerator and freeSpaceFractionDenominator express
-// the 75% share of free space used for the computed default limit as
-// integer arithmetic (dividing before multiplying avoids overflow).
-const (
-	freeSpaceFractionNumerator   uint64 = 3
-	freeSpaceFractionDenominator uint64 = 4
-)
-
-// FreeSpaceProbeFunc reports the number of free bytes available on the
-// filesystem containing path. It is a function type so tests can
-// inject a fake probe instead of depending on the host disk.
-type FreeSpaceProbeFunc func(path string) (uint64, error)
-
-// defaultFreeSpaceProbe reports free filesystem bytes via statfs on
-// the given path, as available to unprivileged processes.
-func defaultFreeSpaceProbe(path string) (uint64, error) {
-	var stat syscall.Statfs_t
-
-	err := syscall.Statfs(path, &stat)
-	if err != nil {
-		return 0, err
-	}
-
-	if stat.Bsize < 0 {
-		return 0, fmt.Errorf("%w %d for %q", errNegativeBlockSize, stat.Bsize, path)
-	}
-
-	blockSize := uint64(stat.Bsize)
-
-	return stat.Bavail * blockSize, nil
-}
-
-// ComputeDefaultCacheMaxBytes returns the default cache size limit for
-// the filesystem containing cacheDir: 75% of the free bytes reported
-// by probe, with a floor of DefaultCacheMaxBytesFloor.
-func ComputeDefaultCacheMaxBytes(
-	cacheDir string, probe FreeSpaceProbeFunc,
-) (int64, error) {
-	freeBytes, err := probe(cacheDir)
-	if err != nil {
-		return 0, fmt.Errorf("config key %q: cannot determine free space for %q: %w",
-			"cache_max_bytes", cacheDir, err)
-	}
-
-	computed := freeBytes / freeSpaceFractionDenominator * freeSpaceFractionNumerator
-	computed = min(computed, math.MaxInt64)
-
-	// gosec cannot see that min() above bounds computed, so it reads
-	// this conversion as potentially overflowing. It cannot: computed is
-	// at most math.MaxInt64 on every path here.
-	//nolint:gosec // G115: clamped to MaxInt64 by min above
-	limit := int64(computed)
-	limit = max(limit, DefaultCacheMaxBytesFloor)
-
-	return limit, nil
-}
-
-// resolveCacheMaxBytes finalizes CacheMaxBytes after state_dir
-// validation: an explicitly configured value is kept as-is (no floor
-// applies), while an omitted key receives the computed default based
-// on free space in <state_dir>/cache/. The cache directory is created
-// first so statfs measures the filesystem that will actually hold the
-// cache. The effective limit is logged either way.
-func (c *Config) resolveCacheMaxBytes(
-	log *slog.Logger, probe FreeSpaceProbeFunc,
-) error {
-	if !c.cacheMaxBytesExplicit {
-		cacheDir := filepath.Join(c.StateDir, "cache")
-
-		err := os.MkdirAll(cacheDir, cacheDirPerms)
-		if err != nil {
-			return fmt.Errorf("config key %q: cannot create cache directory %q: %w",
-				keyCacheMaxBytes, cacheDir, err)
-		}
-
-		limit, err := ComputeDefaultCacheMaxBytes(cacheDir, probe)
-		if err != nil {
-			return err
-		}
-
-		c.CacheMaxBytes = limit
-
-		log.Info("computed default cache size limit from free space",
-			"cache_max_bytes", limit,
-			"cache_dir", cacheDir,
-		)
-	}
-
-	log.Info("effective cache size limit",
-		"cache_max_bytes", c.CacheMaxBytes,
-		"cache_disabled", c.CacheMaxBytes == 0,
-	)
-
-	return nil
-}

internal/config/config.go

@@ -11,10 +11,12 @@ import (
 	"net/url"
 	"os"
 	"path/filepath"
+	"regexp"
 	"runtime"
 	"sort"
 	"strconv"
 	"strings"
+	"syscall"
 	"time"
 
 	"git.eeqj.de/sneak/smartconfig"
@@ -47,6 +49,7 @@ const (
 	keyMetricsPassword            = "metrics.password"
 	keySigningKey                 = "signing_key"
 	keyAllowlistHosts             = "allowlist_hosts"
+	keyRefererBlocklist           = "referer_blocklist"
 	keyAllowHTTP                  = "allow_http"
 	keyUpstreamConnectionsPerHost = "upstream_connections_per_host"
 	keyUpstreamConnections        = "upstream_connections"
@@ -61,7 +64,7 @@ const (
 )
 
 // placeholderSigningKey is the dummy signing_key shipped in
-// config.example.yml. It is 45 characters, so it passes the length
+// configs/config.example.yml. It is 45 characters, so it passes the length
 // check, but it is public in this repository and must be rejected at
 // startup so no deployment ever signs URLs with it.
 const placeholderSigningKey = "CHANGE_ME_generate_with_openssl_rand_base64_32"
@@ -87,23 +90,20 @@ var (
 	errMustBeAtLeastOne  = errors.New("must be at least 1")
 	errValueTooShort     = errors.New("value too short")
 	errPlaceholderKey    = errors.New(
-		"is the placeholder from config.example.yml; " +
+		"is the placeholder from configs/config.example.yml; " +
 			"generate a real key with: openssl rand -base64 32")
 	errMustBeSetTogether = errors.New("must be set together")
 	errMustNotBeNegative = errors.New("must not be negative")
 	errOverflowsInt64    = errors.New("overflows a 64-bit integer")
-	errNegativeBlockSize = errors.New(
-		"statfs reported negative block size")
-	errValueNull = errors.New(
+	errValueNull         = errors.New(
 		"value is null; omit the key entirely to use the default")
 	errValuesNull = errors.New(
 		"value is null; omit a key entirely to use its default")
-	errNotBareHostname = errors.New(
-		"must be a bare hostname without scheme, path, or whitespace")
-	errNoHostnameLabels = errors.New("contains no hostname labels")
-	errNotADuration     = errors.New("not a duration such as 30s or 2m")
-	errMustBePositive   = errors.New("must be positive")
-	errNotAnOrigin      = errors.New(
+	errNotAHost = errors.New("must be a host name such as " +
+		"cdn.example.com or .example.com, or an IP address")
+	errNotADuration   = errors.New("not a duration such as 30s or 2m")
+	errMustBePositive = errors.New("must be positive")
+	errNotAnOrigin    = errors.New(
 		`not "*" or an origin such as https://example.com`)
 )
 
@@ -131,6 +131,10 @@ type Config struct {
 	AllowHTTP                  bool     // Allow non-TLS upstream (testing only)
 	UpstreamConnectionsPerHost int      // Max concurrent connections per upstream host
 
+	// RefererBlocklist holds host patterns, matched as AllowlistHosts is: the
+	// image routes refuse a request whose Referer names a matching host.
+	RefererBlocklist []string
+
 	// UpstreamConnections is the most concurrent connections to all
 	// upstream hosts together, on top of the per-host limit.
 	// MaxConcurrentProcessing is the most images processed at once.
@@ -170,18 +174,19 @@ type Config struct {
 	// address, and an explicit list replaces the default.
 	TrustedProxies []netip.Prefix
 
-	// CacheMaxBytes is the disk cache size limit in bytes. Zero
-	// disables the disk cache entirely. When cache_max_bytes is
-	// omitted from the configuration, this holds the computed default
-	// (75% of free space on the filesystem containing
-	// <state_dir>/cache/, floored at DefaultCacheMaxBytesFloor).
+	// CacheMaxBytes is the disk cache size limit in bytes. Only an
+	// explicit zero (CacheMaxBytesExplicit true) disables the disk
+	// cache. Zero with CacheMaxBytesExplicit false means
+	// cache_max_bytes was omitted, and the cache works out the default
+	// limit when it opens.
 	CacheMaxBytes int64
 
-	// cacheMaxBytesExplicit records whether cache_max_bytes was
+	// CacheMaxBytesExplicit records whether cache_max_bytes was
 	// explicitly set, in the environment or the configuration file.
-	// Explicit values are used exactly as given; only an omitted key
-	// gets the computed default (and its floor) in resolveCacheMaxBytes.
-	cacheMaxBytesExplicit bool
+	// Explicit values are used exactly as given; for an omitted key the
+	// cache works out the default limit when it opens (see
+	// imgcache.CacheConfig.UseDefaultMaxBytes).
+	CacheMaxBytesExplicit bool
 }
 
 // New creates a new Config instance from the environment and the
@@ -218,9 +223,13 @@ func New(_ fx.Lifecycle, params Params) (*Config, error) {
 		return nil, err
 	}
 
-	err = c.resolveCacheMaxBytes(log, defaultFreeSpaceProbe)
-	if err != nil {
-		return nil, err
+	// An omitted cache_max_bytes is worked out and logged when the
+	// cache opens.
+	if c.CacheMaxBytesExplicit {
+		log.Info("effective cache size limit",
+			"cache_max_bytes", c.CacheMaxBytes,
+			"cache_disabled", c.CacheMaxBytes == 0,
+		)
 	}
 
 	if c.Debug {
@@ -266,7 +275,6 @@ func newFromSmartConfig(sc *smartconfig.Config) (*Config, error) {
 	}
 
 	loader := &strictLoader{sc: sc}
-
 	c := &Config{
 		Debug:           loader.boolVal(keyDebug, false),
 		MaintenanceMode: loader.boolVal(keyMaintenanceMode, false),
@@ -294,16 +302,17 @@ func newFromSmartConfig(sc *smartconfig.Config) (*Config, error) {
 			keyAccessControlAllowOrigin, DefaultAccessControlAllowOrigin),
 		DownstreamTimeout: loader.durationVal(
 			keyDownstreamTimeout, DefaultDownstreamTimeout),
-		CacheMaxBytes:   loader.int64Val(keyCacheMaxBytes, 0),
-		BlockedNetworks: blockedNetworks,
-		TrustedProxies:  trustedProxies,
+		CacheMaxBytes:    loader.int64Val(keyCacheMaxBytes, 0),
+		BlockedNetworks:  blockedNetworks,
+		TrustedProxies:   trustedProxies,
+		RefererBlocklist: loader.hostListVal(keyRefererBlocklist),
 	}
 
-	// The computed default for cache_max_bytes needs a validated
-	// state_dir, so it is resolved later (resolveCacheMaxBytes); here
-	// we only record whether the operator set the key explicitly.
+	// The default for an omitted cache_max_bytes is worked out when
+	// the cache opens; here we only record whether the operator set
+	// the key explicitly.
 	if _, present := lookupValue(sc, keyCacheMaxBytes); present {
-		c.cacheMaxBytesExplicit = true
+		c.CacheMaxBytesExplicit = true
 	}
 
 	// Build DBURL from StateDir if not explicitly set. The derived URL
@@ -316,7 +325,8 @@ func newFromSmartConfig(sc *smartconfig.Config) (*Config, error) {
 				settingName(keyDBURL), errValueEmpty)
 		}
 
-		c.DBURL = fmt.Sprintf("file:%s/state.sqlite3?_journal_mode=WAL", c.StateDir)
+		// The driver sets the journal mode only through a _pragma parameter.
+		c.DBURL = fmt.Sprintf("file:%s/state.sqlite3?_pragma=journal_mode(WAL)", c.StateDir)
 	}
 
 	if loader.err != nil {
@@ -415,7 +425,8 @@ func isKnownConfigKey(key string) bool {
 		keyUpstreamConnectionsPerHost, keyUpstreamConnections,
 		keyMaxConcurrentProcessing, keyCacheMaxBytes, keyBlockedNetworks,
 		keyTrustedProxies, keyAccessControlAllowOrigin, keyUpstreamFetchTimeout,
-		keyUpstreamMaxResponseSize, keyDownstreamTimeout, "env":
+		keyUpstreamMaxResponseSize, keyDownstreamTimeout, keyRefererBlocklist,
+		"env":
 		return true
 	}
 
@@ -438,6 +449,7 @@ func envVarNames() map[string]string {
 		keyMetricsPassword:            "PIXA_METRICS_PASSWORD",
 		keySigningKey:                 "PIXA_SIGNING_KEY",
 		keyAllowlistHosts:             "PIXA_ALLOWLIST_HOSTS",
+		keyRefererBlocklist:           "PIXA_REFERER_BLOCKLIST",
 		keyAllowHTTP:                  "PIXA_ALLOW_HTTP",
 		keyUpstreamConnectionsPerHost: "PIXA_UPSTREAM_CONNECTIONS_PER_HOST",
 		keyUpstreamConnections:        "PIXA_UPSTREAM_CONNECTIONS",
@@ -549,8 +561,8 @@ func (c *Config) ensureStateDirWritable() error {
 }
 
 // validateSigningKey checks that the signing key is present, long
-// enough, and not the public placeholder from config.example.yml. The
-// key value itself is never echoed in error messages.
+// enough, and not the public placeholder from configs/config.example.yml.
+// The key value itself is never echoed in error messages.
 func (c *Config) validateSigningKey() error {
 	if c.SigningKey == "" {
 		return fmt.Errorf("%s: %w", settingName(keySigningKey), errValueRequired)
@@ -607,7 +619,7 @@ func (c *Config) validate() error {
 	}
 
 	for _, host := range c.AllowlistHosts {
-		err := validateAllowlistHost(host)
+		err := validateHostPattern(keyAllowlistHosts, host)
 		if err != nil {
 			return err
 		}
@@ -730,25 +742,24 @@ func (c *Config) validateConcurrencyLimits() error {
 	return nil
 }
 
-// validateAllowlistHost checks that an allowlist_hosts entry is a bare
-// hostname, optionally with a leading dot for suffix matching. URLs,
-// paths, and whitespace indicate a misconfigured entry. An entry with
-// no hostname labels (such as ".") is rejected: the allowlist matcher
-// treats a leading dot as a suffix pattern, so a bare "." would match
-// any upstream host written in FQDN trailing-dot form and effectively
-// disable URL signing.
-func validateAllowlistHost(host string) error {
-	if strings.Contains(host, "://") || strings.ContainsAny(host, "/ \t") {
-		return fmt.Errorf("%s: entry %q %w",
-			settingName(keyAllowlistHosts), host, errNotBareHostname)
+// hostNamePattern matches a host name: letters, digits, hyphens, underscores
+// and dots, optionally after one leading dot.
+var hostNamePattern = regexp.MustCompile(`^\.?[A-Za-z0-9_-][A-Za-z0-9_.-]*$`)
+
+// validateHostPattern checks that an entry of the named key, allowlist_hosts
+// or referer_blocklist, is an IP address or a host name, the host name
+// optionally with one leading dot for suffix matching. Anything else, such as
+// a URL, a port or a "*." wildcard, can never match a host name that resolves,
+// so it is refused.
+// So is "." alone: the allowlist matcher would match it against any host
+// written with a trailing dot, which in allowlist_hosts disables URL signing.
+func validateHostPattern(key, host string) error {
+	_, err := netip.ParseAddr(host)
+	if err == nil || hostNamePattern.MatchString(host) {
+		return nil
 	}
 
-	if strings.Trim(host, ".") == "" {
-		return fmt.Errorf("%s: entry %q %w",
-			settingName(keyAllowlistHosts), host, errNoHostnameLabels)
-	}
-
-	return nil
+	return fmt.Errorf("%s: entry %q %w", settingName(key), host, errNotAHost)
 }
 
 // loadConfigFile loads configuration from the PIXA_CONFIG_PATH env var
@@ -779,10 +790,12 @@ func loadConfigFile(log *slog.Logger, appName string) (*smartconfig.Config, erro
 	for _, path := range configPaths {
 		cleanPath := filepath.Clean(path)
 
-		// Only a config file that does not exist is skipped. One that
-		// cannot be read or does not parse is a fatal startup error.
+		// Only a config file that does not exist is skipped, including
+		// one whose path runs through a file, such as under a HOME of
+		// /dev/null. One that cannot be read or does not parse is a
+		// fatal startup error.
 		_, statErr := os.Stat(cleanPath)
-		if errors.Is(statErr, fs.ErrNotExist) {
+		if errors.Is(statErr, fs.ErrNotExist) || errors.Is(statErr, syscall.ENOTDIR) {
 			continue
 		}
 
@@ -876,6 +889,19 @@ func (l *strictLoader) boolVal(key string, defaultVal bool) bool {
 	return val
 }
 
+func (l *strictLoader) hostListVal(key string) []string {
+	if l.err != nil {
+		return nil
+	}
+
+	val, err := parseHostList(l.sc, key)
+	if err != nil {
+		l.err = err
+	}
+
+	return val
+}
+
 // getString returns the string value for key, or defaultVal if the key
 // is omitted. A present value that is not a string, or is explicitly
 // null, is an error.
@@ -1173,7 +1199,7 @@ func parseCIDRList(sc *smartconfig.Config, key string) ([]netip.Prefix, error) {
 		return nil, errNullConfigValue(key)
 	}
 
-	entries, err := cidrListEntries(raw, key)
+	entries, err := listEntries(raw, key)
 	if err != nil {
 		return nil, err
 	}
@@ -1193,11 +1219,41 @@ func parseCIDRList(sc *smartconfig.Config, key string) ([]netip.Prefix, error) {
 	return prefixes, nil
 }
 
-// cidrListEntries extracts the raw entries of the named CIDR-list key as
-// trimmed, non-empty strings, from either a YAML list of strings or a
-// comma-separated string; an empty string is an empty list, as for
-// allowlist_hosts. Any other shape is a configuration error.
-func cidrListEntries(raw any, key string) ([]string, error) {
+// parseHostList parses the value of the named config key into host patterns,
+// or returns nil if the key is omitted. It accepts a YAML list of strings or a
+// comma-separated string. An explicitly null value, a wrong type, an empty
+// entry, a non-string entry, or an entry validateHostPattern rejects aborts
+// startup naming the key and the offending value.
+func parseHostList(sc *smartconfig.Config, key string) ([]string, error) {
+	raw, ok := lookupValue(sc, key)
+	if !ok {
+		return nil, nil
+	}
+
+	if raw == nil {
+		return nil, errNullConfigValue(key)
+	}
+
+	entries, err := listEntries(raw, key)
+	if err != nil {
+		return nil, err
+	}
+
+	for _, entry := range entries {
+		err := validateHostPattern(key, entry)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	return entries, nil
+}
+
+// listEntries extracts the raw entries of the named list key as trimmed,
+// non-empty strings, from either a YAML list of strings or a comma-separated
+// string; an empty string is an empty list, as for allowlist_hosts. Any other
+// shape is a configuration error.
+func listEntries(raw any, key string) ([]string, error) {
 	switch val := raw.(type) {
 	case []any:
 		entries := make([]string, 0, len(val))

internal/config/config_validation_internal_test.go

@@ -1,14 +1,18 @@
 package config
 
 import (
+	"database/sql"
 	"log/slog"
 	"os"
 	"path/filepath"
+	"slices"
 	"strings"
 	"testing"
 	"time"
 
 	"git.eeqj.de/sneak/smartconfig"
+
+	_ "modernc.org/sqlite" // SQLite driver registration
 )
 
 // validTestSigningKey is a 32-character signing key that satisfies the
@@ -94,12 +98,43 @@ func TestOmittedValuesUseDefaults(t *testing.T) {
 		t.Errorf("AllowlistHosts = %v, want empty", c.AllowlistHosts)
 	}
 
-	wantDBURL := "file:" + DefaultStateDir + "/state.sqlite3?_journal_mode=WAL"
+	wantDBURL := "file:" + DefaultStateDir +
+		"/state.sqlite3?_pragma=journal_mode(WAL)"
 	if c.DBURL != wantDBURL {
 		t.Errorf("DBURL = %q, want derived default %q", c.DBURL, wantDBURL)
 	}
 }
 
+// TestDefaultDBURLOpensTheDatabaseInWALMode opens the db_url derived from
+// state_dir with the SQLite driver pixad uses and checks that the database
+// is in WAL mode: the driver ignores any parameter it does not know.
+func TestDefaultDBURLOpensTheDatabaseInWALMode(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine+"state_dir: "+t.TempDir()+"\n")
+	if err != nil {
+		t.Fatalf("config with only state_dir set should be valid, got: %v", err)
+	}
+
+	db, err := sql.Open("sqlite", c.DBURL)
+	if err != nil {
+		t.Fatalf("failed to open %q: %v", c.DBURL, err)
+	}
+
+	t.Cleanup(func() { _ = db.Close() })
+
+	var journalMode string
+
+	err = db.QueryRowContext(t.Context(), "PRAGMA journal_mode").Scan(&journalMode)
+	if err != nil {
+		t.Fatalf("failed to read the journal mode of %q: %v", c.DBURL, err)
+	}
+
+	if journalMode != "wal" {
+		t.Errorf("journal mode of %q = %q, want wal", c.DBURL, journalMode)
+	}
+}
+
 func TestExplicitValidValuesAreUsed(t *testing.T) {
 	t.Parallel()
 
@@ -182,6 +217,25 @@ func TestCommaSeparatedAllowlistStillSupported(t *testing.T) {
 	}
 }
 
+// TestAllowlistHostsAcceptsUnderscore checks that an upstream host name with
+// an underscore, which pixa can fetch from, is accepted as an entry.
+func TestAllowlistHostsAcceptsUnderscore(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine+`allowlist_hosts:
+  - my_bucket.example.com
+  - .my_bucket.example.org
+`)
+	if err != nil {
+		t.Fatalf("host names with an underscore should load, got error: %v", err)
+	}
+
+	want := []string{"my_bucket.example.com", ".my_bucket.example.org"}
+	if !slices.Equal(c.AllowlistHosts, want) {
+		t.Errorf("AllowlistHosts = %v, want %v", c.AllowlistHosts, want)
+	}
+}
+
 // runAbortCases asserts that each case's config aborts startup with an
 // error message mentioning every expected substring.
 func runAbortCases(t *testing.T, cases []abortCase) {
@@ -284,6 +338,20 @@ func invalidHostAndCredentialCases() []abortCase {
 				keyAllowlistHosts, "example.com/images",
 			},
 		},
+		{
+			name: "allowlist host with wildcard",
+			yaml: signingKeyLine + "allowlist_hosts:\n  - \"*.example.com\"\n",
+			wantErrSubstrings: []string{
+				keyAllowlistHosts, "*.example.com",
+			},
+		},
+		{
+			name: "allowlist host with port",
+			yaml: signingKeyLine + "allowlist_hosts:\n  - example.com:8443\n",
+			wantErrSubstrings: []string{
+				keyAllowlistHosts, "example.com:8443",
+			},
+		},
 		{
 			name:              "allowlist host with whitespace",
 			yaml:              signingKeyLine + "allowlist_hosts:\n  - \"exa mple.com\"\n",

internal/config/env_internal_test.go

@@ -65,6 +65,7 @@ func TestEnvironmentSetsEveryKey(t *testing.T) {
 	t.Setenv("PIXA_METRICS_PASSWORD", "metricspass")
 	t.Setenv("PIXA_SIGNING_KEY", validTestSigningKey)
 	t.Setenv("PIXA_ALLOWLIST_HOSTS", "s3.sneak.cloud,.example.com")
+	t.Setenv("PIXA_REFERER_BLOCKLIST", "hotlinker.example,.leech.example")
 	t.Setenv("PIXA_ALLOW_HTTP", "true")
 	t.Setenv("PIXA_UPSTREAM_CONNECTIONS_PER_HOST", "5")
 	t.Setenv("PIXA_UPSTREAM_CONNECTIONS", "10")
@@ -93,12 +94,13 @@ func TestEnvironmentSetsEveryKey(t *testing.T) {
 		MetricsPassword:            "metricspass",
 		SigningKey:                 validTestSigningKey,
 		AllowlistHosts:             []string{testHostS3, ".example.com"},
+		RefererBlocklist:           []string{"hotlinker.example", ".leech.example"},
 		AllowHTTP:                  true,
 		UpstreamConnectionsPerHost: 5,
 		UpstreamConnections:        10,
 		MaxConcurrentProcessing:    3,
 		CacheMaxBytes:              1024,
-		cacheMaxBytesExplicit:      true,
+		CacheMaxBytesExplicit:      true,
 		BlockedNetworks:            []netip.Prefix{netip.MustParsePrefix("203.0.113.0/24")},
 		TrustedProxies:             []netip.Prefix{netip.MustParsePrefix("192.0.2.0/24")},
 		AccessControlAllowOrigin:   "https://app.example.com",

internal/config/referer_blocklist_internal_test.go

@@ -0,0 +1,166 @@
+package config
+
+import (
+	"slices"
+	"testing"
+)
+
+// TestRefererBlocklistParsed loads a referer_blocklist with a host and a
+// pattern starting with "." and checks both are kept in order.
+func TestRefererBlocklistParsed(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine+`referer_blocklist:
+  - leech.example
+  - .hotlinker.example
+`)
+	if err != nil {
+		t.Fatalf("valid referer_blocklist should load, got error: %v", err)
+	}
+
+	want := []string{"leech.example", ".hotlinker.example"}
+	if !slices.Equal(c.RefererBlocklist, want) {
+		t.Errorf("RefererBlocklist = %v, want %v", c.RefererBlocklist, want)
+	}
+}
+
+// TestRefererBlocklistAcceptsIPAddresses checks that IPv4 and IPv6 addresses,
+// the IPv6 one written without brackets, are accepted as entries.
+func TestRefererBlocklistAcceptsIPAddresses(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine+`referer_blocklist:
+  - 192.0.2.7
+  - "2001:db8::7"
+`)
+	if err != nil {
+		t.Fatalf("IP address entries should load, got error: %v", err)
+	}
+
+	want := []string{"192.0.2.7", "2001:db8::7"}
+	if !slices.Equal(c.RefererBlocklist, want) {
+		t.Errorf("RefererBlocklist = %v, want %v", c.RefererBlocklist, want)
+	}
+}
+
+// TestRefererBlocklistAcceptsUnderscore checks that a host name with an
+// underscore, which a page can be served from, is accepted as an entry.
+func TestRefererBlocklistAcceptsUnderscore(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine+`referer_blocklist:
+  - my_site.leech.example
+  - .my_site.hotlinker.example
+`)
+	if err != nil {
+		t.Fatalf("host names with an underscore should load, got error: %v", err)
+	}
+
+	want := []string{"my_site.leech.example", ".my_site.hotlinker.example"}
+	if !slices.Equal(c.RefererBlocklist, want) {
+		t.Errorf("RefererBlocklist = %v, want %v", c.RefererBlocklist, want)
+	}
+}
+
+// TestRefererBlocklistOmittedIsEmpty checks that an omitted key blocks no
+// referer.
+func TestRefererBlocklistOmittedIsEmpty(t *testing.T) {
+	t.Parallel()
+
+	c, err := configFromYAML(t, signingKeyLine)
+	if err != nil {
+		t.Fatalf("minimal config should be valid, got error: %v", err)
+	}
+
+	if len(c.RefererBlocklist) != 0 {
+		t.Errorf("RefererBlocklist = %v, want empty", c.RefererBlocklist)
+	}
+}
+
+// TestRefererBlocklistInvalidAbortsStartup checks that an entry that is not a
+// host, or a value that is not a list of them, aborts startup with an error
+// naming the key and the entry.
+func TestRefererBlocklistInvalidAbortsStartup(t *testing.T) {
+	t.Parallel()
+
+	runAbortCases(t, []abortCase{
+		{
+			name: "entry with a scheme",
+			yaml: signingKeyLine + "referer_blocklist:\n  - https://leech.example\n",
+			wantErrSubstrings: []string{
+				keyRefererBlocklist, "https://leech.example",
+			},
+		},
+		{
+			name: "entry with a path",
+			yaml: signingKeyLine + "referer_blocklist:\n  - leech.example/page\n",
+			wantErrSubstrings: []string{
+				keyRefererBlocklist, "leech.example/page",
+			},
+		},
+		{
+			name: "wildcard entry",
+			yaml: signingKeyLine + "referer_blocklist:\n  - \"*.leech.example\"\n",
+			wantErrSubstrings: []string{
+				keyRefererBlocklist, "*.leech.example",
+			},
+		},
+		{
+			name: "entry with a port",
+			yaml: signingKeyLine + "referer_blocklist:\n  - leech.example:8080\n",
+			wantErrSubstrings: []string{
+				keyRefererBlocklist, "leech.example:8080",
+			},
+		},
+		{
+			name: "two leading dots",
+			yaml: signingKeyLine + "referer_blocklist:\n  - ..leech.example\n",
+			wantErrSubstrings: []string{
+				keyRefererBlocklist, "..leech.example",
+			},
+		},
+		{
+			name:              "dot only",
+			yaml:              signingKeyLine + "referer_blocklist:\n  - \".\"\n",
+			wantErrSubstrings: []string{keyRefererBlocklist, `"."`},
+		},
+		{
+			name:              "empty entry",
+			yaml:              signingKeyLine + "referer_blocklist:\n  - \"\"\n",
+			wantErrSubstrings: []string{keyRefererBlocklist},
+		},
+		{
+			name:              "entry not a string",
+			yaml:              signingKeyLine + "referer_blocklist:\n  - 42\n",
+			wantErrSubstrings: []string{keyRefererBlocklist, "42"},
+		},
+		{
+			name:              "null value",
+			yaml:              signingKeyLine + "referer_blocklist:\n",
+			wantErrSubstrings: []string{keyRefererBlocklist, nullValueText},
+		},
+	})
+}
+
+// TestRefererBlocklistFromEnvironment checks that PIXA_REFERER_BLOCKLIST
+// takes comma-separated entries, and that an entry in it that is not a host
+// aborts startup naming the variable and the entry.
+func TestRefererBlocklistFromEnvironment(t *testing.T) {
+	t.Setenv("PIXA_SIGNING_KEY", validTestSigningKey)
+	t.Setenv("PIXA_REFERER_BLOCKLIST", " leech.example , .hotlinker.example ")
+
+	c, err := newFromSmartConfig(nil)
+	if err != nil {
+		t.Fatalf("valid PIXA_REFERER_BLOCKLIST should load, got error: %v", err)
+	}
+
+	want := []string{"leech.example", ".hotlinker.example"}
+	if !slices.Equal(c.RefererBlocklist, want) {
+		t.Errorf("RefererBlocklist = %v, want %v", c.RefererBlocklist, want)
+	}
+
+	t.Setenv("PIXA_REFERER_BLOCKLIST", "leech.example,https://hotlinker.example")
+
+	_, err = newFromSmartConfig(nil)
+	wantStartupError(t, err, "PIXA_REFERER_BLOCKLIST", "https://hotlinker.example")
+}

internal/database/concurrent_writes_internal_test.go

@@ -0,0 +1,153 @@
+package database
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"log/slog"
+	"path/filepath"
+	"sync"
+	"testing"
+
+	"sneak.berlin/go/pixa/internal/config"
+)
+
+// TestConcurrentWritesAllSucceed opens a database the way pixad does and
+// writes to it from several goroutines at once, so the writes run on
+// separate connections, as one request's writes and the background eviction
+// pass do. Every write must succeed, none failing with "database is locked",
+// whether or not db_url already has parameters, and the parameters it has
+// must still apply.
+func TestConcurrentWritesAllSucceed(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name            string
+		query           string
+		wantJournalMode string
+	}{
+		{
+			name:            "db_url without parameters",
+			query:           "",
+			wantJournalMode: "delete",
+		},
+		{
+			name:            "db_url with the WAL parameter",
+			query:           "?_pragma=journal_mode(WAL)",
+			wantJournalMode: "wal",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			dbURL := "file:" + filepath.Join(t.TempDir(), "state.sqlite3") + tt.query
+
+			d := &Database{
+				log:    slog.New(slog.DiscardHandler),
+				config: &config.Config{DBURL: dbURL},
+			}
+
+			err := d.connect(t.Context())
+			if err != nil {
+				t.Fatalf("failed to connect to %q: %v", dbURL, err)
+			}
+
+			t.Cleanup(func() { _ = d.db.Close() })
+
+			writeConcurrently(t, d.db)
+
+			var journalMode string
+
+			err = d.db.QueryRowContext(t.Context(), "PRAGMA journal_mode").
+				Scan(&journalMode)
+			if err != nil {
+				t.Fatalf("failed to read the journal mode: %v", err)
+			}
+
+			if journalMode != tt.wantJournalMode {
+				t.Errorf("journal mode = %q, want %q", journalMode, tt.wantJournalMode)
+			}
+		})
+	}
+}
+
+// writeConcurrently runs writeLikeOneRequest from several goroutines at once
+// and checks that every write was made.
+func writeConcurrently(t *testing.T, db *sql.DB) {
+	t.Helper()
+
+	const (
+		writers       = 4
+		requestsEach  = 20
+		totalRequests = writers * requestsEach
+	)
+
+	ctx := t.Context()
+
+	var wg sync.WaitGroup
+
+	for writer := range writers {
+		wg.Go(func() {
+			for request := range requestsEach {
+				key := fmt.Sprintf("%d-%d", writer, request)
+
+				err := writeLikeOneRequest(ctx, db, key)
+				if err != nil {
+					t.Errorf("writer %d: %v", writer, err)
+
+					return
+				}
+			}
+		})
+	}
+
+	wg.Wait()
+
+	var hits, sources int
+
+	err := db.QueryRowContext(ctx, `
+		SELECT hit_count, (SELECT COUNT(*) FROM source_content)
+		FROM cache_stats WHERE id = 1
+	`).Scan(&hits, &sources)
+	if err != nil {
+		t.Fatalf("failed to count the writes: %v", err)
+	}
+
+	if hits != totalRequests || sources != totalRequests {
+		t.Errorf("hit_count = %d and %d source_content rows, want %d of each",
+			hits, sources, totalRequests)
+	}
+}
+
+// writeLikeOneRequest makes the writes one request and the eviction pass
+// make: it counts a cache hit, stores a source, records a transformed image
+// and deletes that record again.
+func writeLikeOneRequest(ctx context.Context, db *sql.DB, key string) error {
+	_, err := db.ExecContext(ctx,
+		`UPDATE cache_stats SET hit_count = hit_count + 1 WHERE id = 1`)
+	if err != nil {
+		return fmt.Errorf("counting a cache hit: %w", err)
+	}
+
+	_, err = db.ExecContext(ctx, `INSERT INTO source_content
+		(content_hash, content_type, size_bytes) VALUES (?, 'image/png', 1)`, key)
+	if err != nil {
+		return fmt.Errorf("storing a source: %w", err)
+	}
+
+	_, err = db.ExecContext(ctx, `INSERT INTO variant_content
+		(cache_key, size_bytes, content_type) VALUES (?, 1, 'image/png')`, key)
+	if err != nil {
+		return fmt.Errorf("recording a transformed image: %w", err)
+	}
+
+	_, err = db.ExecContext(ctx,
+		`DELETE FROM variant_content WHERE cache_key = ?`, key)
+	if err != nil {
+		return fmt.Errorf("evicting a transformed image: %w", err)
+	}
+
+	return nil
+}

internal/database/database.go

@@ -243,7 +243,17 @@ func (s *Database) DB() *sql.DB {
 }
 
 func (s *Database) connect(ctx context.Context) error {
-	dbURL := s.config.DBURL
+	// Requests and the eviction pass write on separate connections. With
+	// a busy timeout, a write that finds another one in progress waits up
+	// to five seconds for it instead of failing at once with "database is
+	// locked". The driver runs each _pragma parameter on every connection
+	// it opens.
+	separator := "?"
+	if strings.Contains(s.config.DBURL, "?") {
+		separator = "&"
+	}
+
+	dbURL := s.config.DBURL + separator + "_pragma=busy_timeout(5000)"
 
 	s.log.Info("connecting to database", "url", dbURL)
 

internal/handlers/auth_session_internal_test.go

@@ -0,0 +1,243 @@
+package handlers
+
+import (
+	"log/slog"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"regexp"
+	"strings"
+	"testing"
+
+	"sneak.berlin/go/pixa/internal/imgcache"
+	"sneak.berlin/go/pixa/internal/session"
+)
+
+// formatField is the generator form's format field name.
+const formatField = "format"
+
+// Markers telling the login page from the generator page.
+const (
+	loginForm     = `action="/"`
+	loginKeyInput = `name="key"`
+	generatorForm = `action="/generate"`
+)
+
+// generatedURLPattern extracts the path of the URL the generator page shows.
+// The test router runs with debug on, so the URL starts with http, and its
+// host is httptest's default request host.
+var generatedURLPattern = regexp.MustCompile(
+	`value="http://example\.com(/v1/e/[^"]+)"`)
+
+// findSessionCookie returns the session cookie rec sets, or nil if it sets
+// none.
+func findSessionCookie(rec *httptest.ResponseRecorder) *http.Cookie {
+	for _, c := range rec.Result().Cookies() {
+		if c.Name == session.CookieName {
+			return c
+		}
+	}
+
+	return nil
+}
+
+// TestHandleRoot_NoSession_ShowsLoginForm verifies that GET / without a
+// login session shows the login form.
+func TestHandleRoot_NoSession_ShowsLoginForm(t *testing.T) {
+	t.Parallel()
+
+	_, srv := newCSRFTestRouter(t)
+
+	rec := httptest.NewRecorder()
+	srv.ServeHTTP(rec, httptest.NewRequestWithContext(
+		t.Context(), http.MethodGet, "/", nil))
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d", rec.Code, http.StatusOK)
+	}
+
+	body := rec.Body.String()
+	if !strings.Contains(body, loginForm) || !strings.Contains(body, loginKeyInput) {
+		t.Errorf("page is not the login form: %s", body)
+	}
+}
+
+// TestLoginPost_WrongKey_ShowsErrorWithoutSession verifies that a wrong key
+// shows the login form again with an error, and sets no session cookie.
+func TestLoginPost_WrongKey_ShowsErrorWithoutSession(t *testing.T) {
+	t.Parallel()
+
+	_, srv := newCSRFTestRouter(t)
+
+	cookies, token := csrfCredentials(t, srv, nil)
+
+	rec := postForm(srv, "/", cookies, url.Values{
+		loginKeyField:  {"wrong-signing-key-fedcba9876543210"},
+		csrfTokenField: {token},
+	})
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d", rec.Code, http.StatusOK)
+	}
+
+	body := rec.Body.String()
+	if !strings.Contains(body, loginForm) || !strings.Contains(body, loginKeyInput) {
+		t.Errorf("page is not the login form: %s", body)
+	}
+
+	if !strings.Contains(body, "Invalid signing key") {
+		t.Error("login form does not show the error")
+	}
+
+	if c := findSessionCookie(rec); c != nil {
+		t.Errorf("wrong key set a session cookie: %s", c)
+	}
+}
+
+// TestLoginPost_RightKey_SetsSessionCookie verifies that the right key answers
+// 303 to / with a session cookie marked Secure, HttpOnly and SameSite=Strict,
+// and that GET / with that cookie shows the generator page.
+func TestLoginPost_RightKey_SetsSessionCookie(t *testing.T) {
+	t.Parallel()
+
+	_, srv := newCSRFTestRouter(t)
+
+	cookies, token := csrfCredentials(t, srv, nil)
+
+	rec := postForm(srv, "/", cookies, url.Values{
+		loginKeyField:  {testSigningKey},
+		csrfTokenField: {token},
+	})
+
+	if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/" {
+		t.Fatalf("status = %d, Location = %q, want %d to /",
+			rec.Code, rec.Header().Get("Location"), http.StatusSeeOther)
+	}
+
+	sessionCookie := findSessionCookie(rec)
+	if sessionCookie == nil {
+		t.Fatal("right key set no session cookie")
+	}
+
+	t.Logf("Set-Cookie: %s", sessionCookie)
+
+	if !sessionCookie.Secure {
+		t.Error("session cookie is not Secure")
+	}
+
+	if !sessionCookie.HttpOnly {
+		t.Error("session cookie is not HttpOnly")
+	}
+
+	if sessionCookie.SameSite != http.SameSiteStrictMode {
+		t.Errorf("session cookie SameSite = %v, want Strict", sessionCookie.SameSite)
+	}
+
+	req := httptest.NewRequestWithContext(t.Context(), http.MethodGet, "/", nil)
+	req.AddCookie(sessionCookie)
+
+	rec = httptest.NewRecorder()
+	srv.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK ||
+		!strings.Contains(rec.Body.String(), generatorForm) {
+		t.Errorf("GET / with the session cookie: status = %d, "+
+			"want %d and the generator page", rec.Code, http.StatusOK)
+	}
+}
+
+// TestHandleLogout_ClearsSessionCookie verifies that GET /logout answers 303
+// to / and replaces the session cookie with an empty one sent with
+// Max-Age=0, which makes the browser delete it.
+func TestHandleLogout_ClearsSessionCookie(t *testing.T) {
+	t.Parallel()
+
+	h, _ := newCSRFTestRouter(t)
+
+	req := httptest.NewRequestWithContext(
+		t.Context(), http.MethodGet, "/logout", nil)
+	req.AddCookie(newSessionCookie(t, h))
+
+	rec := httptest.NewRecorder()
+	h.HandleLogout().ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/" {
+		t.Fatalf("status = %d, Location = %q, want %d to /",
+			rec.Code, rec.Header().Get("Location"), http.StatusSeeOther)
+	}
+
+	t.Logf("Set-Cookie: %s", rec.Header().Get("Set-Cookie"))
+
+	sessionCookie := findSessionCookie(rec)
+	if sessionCookie == nil {
+		t.Fatal("logout did not set the session cookie")
+	}
+
+	if sessionCookie.Value != "" {
+		t.Errorf("session cookie value = %q, want empty", sessionCookie.Value)
+	}
+
+	// net/http reads a Max-Age=0 attribute back as MaxAge -1.
+	if sessionCookie.MaxAge != -1 {
+		t.Errorf("session cookie MaxAge = %d, want -1 (Max-Age=0)",
+			sessionCookie.MaxAge)
+	}
+}
+
+// TestGeneratePost_NoSession_RedirectsToLogin verifies that POST /generate
+// with a valid CSRF token but no login session answers 303 to / and makes no
+// URL.
+func TestGeneratePost_NoSession_RedirectsToLogin(t *testing.T) {
+	t.Parallel()
+
+	_, srv := newCSRFTestRouter(t)
+
+	cookies, token := csrfCredentials(t, srv, nil)
+
+	rec := postForm(srv, "/generate", cookies, url.Values{
+		sourceURLField: {testSourceURL},
+		csrfTokenField: {token},
+	})
+
+	if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/" {
+		t.Fatalf("status = %d, Location = %q, want %d to /",
+			rec.Code, rec.Header().Get("Location"), http.StatusSeeOther)
+	}
+
+	if strings.Contains(rec.Body.String(), "/v1/e/") {
+		t.Error("a URL was made without a login session")
+	}
+}
+
+// TestGeneratePost_URLServesImage verifies that the URL the generator page
+// makes is served by /v1/e/. The image route runs on handlers of its own,
+// made with the same signing key.
+func TestGeneratePost_URLServesImage(t *testing.T) {
+	t.Parallel()
+
+	_, imageSrv := newSignedHostServer(t, slog.New(slog.DiscardHandler))
+
+	rec := generatePost(t, url.Values{
+		sourceURLField: {"https://" + signedHost + photoPath},
+		widthField:     {"50"},
+		heightField:    {"50"},
+		formatField:    {string(imgcache.FormatJPEG)},
+	})
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("POST /generate status = %d, want %d", rec.Code, http.StatusOK)
+	}
+
+	match := generatedURLPattern.FindStringSubmatch(rec.Body.String())
+	if match == nil {
+		t.Fatalf("generator page shows no URL: %s", rec.Body.String())
+	}
+
+	t.Logf("generated URL path: %s", match[1])
+
+	imageRec := httptest.NewRecorder()
+	imageSrv.ServeHTTP(imageRec, httptest.NewRequestWithContext(
+		t.Context(), http.MethodGet, match[1], nil))
+
+	requireServedPhoto(t, imageRec)
+}

internal/handlers/cache_max_bytes_internal_test.go

@@ -0,0 +1,147 @@
+package handlers
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"go.uber.org/fx/fxtest"
+
+	"sneak.berlin/go/pixa/internal/config"
+	"sneak.berlin/go/pixa/internal/database"
+	"sneak.berlin/go/pixa/internal/globals"
+	"sneak.berlin/go/pixa/internal/logger"
+)
+
+// TestNewCacheConfigFromCacheMaxBytes checks the cache configuration
+// built from cache_max_bytes: omitted, the cache works out the default
+// limit; 0 turns the disk cache off; a positive value is the limit,
+// unchanged.
+func TestNewCacheConfigFromCacheMaxBytes(t *testing.T) {
+	t.Parallel()
+
+	const oneGiB = 1 << 30
+
+	cases := []struct {
+		name                   string
+		cacheMaxBytes          int64
+		cacheMaxBytesExplicit  bool
+		wantMaxBytes           int64
+		wantUseDefaultMaxBytes bool
+		wantDisableDiskCache   bool
+	}{
+		{
+			name:                   "cache_max_bytes omitted",
+			cacheMaxBytes:          0,
+			cacheMaxBytesExplicit:  false,
+			wantMaxBytes:           0,
+			wantUseDefaultMaxBytes: true,
+			wantDisableDiskCache:   false,
+		},
+		{
+			name:                   "cache_max_bytes: 0",
+			cacheMaxBytes:          0,
+			cacheMaxBytesExplicit:  true,
+			wantMaxBytes:           0,
+			wantUseDefaultMaxBytes: false,
+			wantDisableDiskCache:   true,
+		},
+		{
+			name:                   "cache_max_bytes: 1 GiB",
+			cacheMaxBytes:          oneGiB,
+			cacheMaxBytesExplicit:  true,
+			wantMaxBytes:           oneGiB,
+			wantUseDefaultMaxBytes: false,
+			wantDisableDiskCache:   false,
+		},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+
+			cfg := &config.Config{
+				CacheMaxBytes:         tc.cacheMaxBytes,
+				CacheMaxBytesExplicit: tc.cacheMaxBytesExplicit,
+			}
+
+			got := newCacheConfig(cfg, nil)
+			t.Logf("MaxBytes = %d, UseDefaultMaxBytes = %v, DisableDiskCache = %v",
+				got.MaxBytes, got.UseDefaultMaxBytes, got.DisableDiskCache)
+
+			if got.MaxBytes != tc.wantMaxBytes {
+				t.Errorf("MaxBytes = %d, want %d", got.MaxBytes, tc.wantMaxBytes)
+			}
+
+			if got.UseDefaultMaxBytes != tc.wantUseDefaultMaxBytes {
+				t.Errorf("UseDefaultMaxBytes = %v, want %v",
+					got.UseDefaultMaxBytes, tc.wantUseDefaultMaxBytes)
+			}
+
+			if got.DisableDiskCache != tc.wantDisableDiskCache {
+				t.Errorf("DisableDiskCache = %v, want %v",
+					got.DisableDiskCache, tc.wantDisableDiskCache)
+			}
+		})
+	}
+}
+
+// TestDiskCacheOffOnlyForExplicitZeroCacheMaxBytes starts the handlers
+// once with cache_max_bytes omitted and once with cache_max_bytes: 0,
+// and checks by whether the cache directories were created that the
+// disk cache is on in the first case and off in the second.
+func TestDiskCacheOffOnlyForExplicitZeroCacheMaxBytes(t *testing.T) {
+	t.Parallel()
+
+	cases := []struct {
+		name                  string
+		cacheMaxBytesExplicit bool
+		wantDiskCache         bool
+	}{
+		{name: "cache_max_bytes omitted", cacheMaxBytesExplicit: false, wantDiskCache: true},
+		{name: "cache_max_bytes: 0", cacheMaxBytesExplicit: true, wantDiskCache: false},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+
+			stateDir := t.TempDir()
+			cfg := &config.Config{
+				SigningKey:            testSigningKey,
+				StateDir:              stateDir,
+				DBURL:                 "file:" + filepath.Join(stateDir, "state.sqlite3"),
+				CacheMaxBytes:         0,
+				CacheMaxBytesExplicit: tc.cacheMaxBytesExplicit,
+			}
+
+			lc := fxtest.NewLifecycle(t)
+
+			log, err := logger.New(lc, logger.Params{Globals: &globals.Globals{}})
+			if err != nil {
+				t.Fatalf("logger.New() error = %v", err)
+			}
+
+			db, err := database.New(lc, database.Params{Logger: log, Config: cfg})
+			if err != nil {
+				t.Fatalf("database.New() error = %v", err)
+			}
+
+			_, err = New(lc, Params{Logger: log, Database: db, Config: cfg})
+			if err != nil {
+				t.Fatalf("New() error = %v", err)
+			}
+
+			lc.RequireStart()
+			t.Cleanup(lc.RequireStop)
+
+			_, err = os.Stat(filepath.Join(stateDir, "cache", "variants"))
+			gotDiskCache := err == nil
+
+			if gotDiskCache != tc.wantDiskCache {
+				t.Errorf("cache directories created = %v, want %v",
+					gotDiskCache, tc.wantDiskCache)
+			}
+		})
+	}
+}

internal/handlers/handlers.go

@@ -9,6 +9,7 @@ import (
 	"time"
 
 	"go.uber.org/fx"
+	"sneak.berlin/go/pixa/internal/allowlist"
 	"sneak.berlin/go/pixa/internal/config"
 	"sneak.berlin/go/pixa/internal/database"
 	"sneak.berlin/go/pixa/internal/encurl"
@@ -40,6 +41,10 @@ type Handlers struct {
 	sessMgr     *session.Manager
 	encGen      *encurl.Generator
 	csrfProtect func(http.Handler) http.Handler
+
+	// refererBlocklist matches the hosts of referer_blocklist; its IsAllowed
+	// reports whether a URL's host is on that list.
+	refererBlocklist *allowlist.HostAllowList
 }
 
 // New creates a new Handlers instance.
@@ -50,11 +55,12 @@ func New(lc fx.Lifecycle, params Params) (*Handlers, error) {
 	}
 
 	s := &Handlers{
-		log:         params.Logger.Get(),
-		hc:          params.Healthcheck,
-		db:          params.Database,
-		config:      params.Config,
-		csrfProtect: csrfProtect,
+		log:              params.Logger.Get(),
+		hc:               params.Healthcheck,
+		db:               params.Database,
+		config:           params.Config,
+		csrfProtect:      csrfProtect,
+		refererBlocklist: allowlist.New(params.Config.RefererBlocklist),
 	}
 
 	lc.Append(fx.Hook{
@@ -80,18 +86,25 @@ func (s *Handlers) WaitForProcessing(ctx context.Context) int {
 	return s.imgSvc.WaitForProcessing(ctx)
 }
 
+// newCacheConfig builds the image cache's configuration from cfg.
+// cache_max_bytes: 0 disables the disk cache entirely; any other value
+// is the eviction limit in bytes; when it is omitted, the cache works
+// out the default limit itself.
+func newCacheConfig(cfg *config.Config, log *slog.Logger) imgcache.CacheConfig {
+	return imgcache.CacheConfig{
+		StateDir:           cfg.StateDir,
+		CacheTTL:           imgcache.DefaultCacheTTL,
+		NegativeTTL:        imgcache.DefaultNegativeTTL,
+		MaxBytes:           cfg.CacheMaxBytes,
+		UseDefaultMaxBytes: !cfg.CacheMaxBytesExplicit,
+		DisableDiskCache:   cfg.CacheMaxBytesExplicit && cfg.CacheMaxBytes == 0,
+		Logger:             log,
+	}
+}
+
 // initImageService initializes the image cache and service.
 func (s *Handlers) initImageService() error {
-	// Create the cache. cache_max_bytes: 0 disables the disk cache
-	// entirely; any other value is the eviction limit in bytes.
-	cache, err := imgcache.NewCache(s.db.DB(), imgcache.CacheConfig{
-		StateDir:         s.config.StateDir,
-		CacheTTL:         imgcache.DefaultCacheTTL,
-		NegativeTTL:      imgcache.DefaultNegativeTTL,
-		MaxBytes:         s.config.CacheMaxBytes,
-		DisableDiskCache: s.config.CacheMaxBytes == 0,
-		Logger:           s.log,
-	})
+	cache, err := imgcache.NewCache(s.db.DB(), newCacheConfig(s.config, s.log))
 	if err != nil {
 		return err
 	}

internal/handlers/image.go

@@ -21,6 +21,10 @@ import (
 // /v1/image/<host>/<path>/<width>x<height>.<format>
 func (s *Handlers) HandleImage() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
+		if s.refuseBlockedReferer(w, r) {
+			return
+		}
+
 		req, ok := s.parseImageRequest(w, r)
 		if !ok {
 			return
@@ -248,6 +252,23 @@ func cacheControl(expires time.Time) string {
 	return fmt.Sprintf("public, max-age=%d, immutable", int64(maxAge/time.Second))
 }
 
+// refuseBlockedReferer answers 403 with a JSON error when the request's Referer
+// names a host on referer_blocklist, and reports whether it answered. A request
+// with no Referer, or one that does not parse as a URL with a host, is not
+// refused.
+func (s *Handlers) refuseBlockedReferer(
+	w http.ResponseWriter, r *http.Request,
+) bool {
+	referer, err := url.Parse(r.Referer())
+	if err != nil || !s.refererBlocklist.IsAllowed(referer) {
+		return false
+	}
+
+	s.respondError(w, "referer blocked", http.StatusForbidden)
+
+	return true
+}
+
 // notModified sets the ETag header to etag and, when the request's
 // If-None-Match is that ETag, answers 304 Not Modified. It reports whether it
 // answered. An empty etag sets no header and never answers.

internal/handlers/imageenc.go

@@ -22,6 +22,10 @@ import (
 // browsers identify the content type.
 func (s *Handlers) HandleImageEnc() http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
+		if s.refuseBlockedReferer(w, r) {
+			return
+		}
+
 		ctx := r.Context()
 		start := time.Now()
 

internal/handlers/imageenc_token_internal_test.go

@@ -0,0 +1,126 @@
+package handlers
+
+import (
+	"image/jpeg"
+	"log/slog"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+	"time"
+
+	"sneak.berlin/go/pixa/internal/encurl"
+)
+
+// requireServedPhoto requires that rec answers 200 with the JPEG at photoPath
+// on signedHost at the 50x50 that encPhotoURL and the generator tests ask for.
+func requireServedPhoto(t *testing.T, rec *httptest.ResponseRecorder) {
+	t.Helper()
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want %d; body %q",
+			rec.Code, http.StatusOK, rec.Body.String())
+	}
+
+	contentType := rec.Header().Get("Content-Type")
+	if contentType != "image/jpeg" {
+		t.Errorf("Content-Type = %q, want image/jpeg", contentType)
+	}
+
+	img, err := jpeg.DecodeConfig(rec.Body)
+	if err != nil {
+		t.Fatalf("body is not a JPEG: %v", err)
+	}
+
+	if img.Width != 50 || img.Height != 50 {
+		t.Errorf("image is %dx%d, want 50x50", img.Width, img.Height)
+	}
+}
+
+// TestHandleImageEnc_ValidToken_ServesImage verifies that a token made with
+// the signing key serves the image it asks for.
+func TestHandleImageEnc_ValidToken_ServesImage(t *testing.T) {
+	t.Parallel()
+
+	h, srv := newSignedHostServer(t, slog.New(slog.DiscardHandler))
+
+	rec := httptest.NewRecorder()
+	srv.ServeHTTP(rec, httptest.NewRequestWithContext(
+		t.Context(), http.MethodGet, encPhotoURL(t, h), nil))
+
+	requireServedPhoto(t, rec)
+}
+
+// TestHandleImageEnc_RejectedToken verifies that a token that has expired
+// answers 410, and that a token with one character changed, a token cut
+// short, and a token made with another signing key answer 400. The server
+// would serve the photo for a token it accepted.
+func TestHandleImageEnc_RejectedToken(t *testing.T) {
+	t.Parallel()
+
+	h, srv := newSignedHostServer(t, slog.New(slog.DiscardHandler))
+
+	photo := encurl.Payload{
+		SourceHost: signedHost,
+		SourcePath: photoPath,
+		Width:      50,
+		Height:     50,
+	}
+
+	valid, err := h.encGen.Generate(&photo)
+	if err != nil {
+		t.Fatalf("Generate() error = %v", err)
+	}
+
+	expiredPhoto := photo
+	expiredPhoto.ExpiresAt = time.Now().Add(-time.Minute).Unix()
+
+	expired, err := h.encGen.Generate(&expiredPhoto)
+	if err != nil {
+		t.Fatalf("Generate() error = %v", err)
+	}
+
+	otherGen, err := encurl.NewGenerator("another-signing-key-fedcba9876543210")
+	if err != nil {
+		t.Fatalf("encurl.NewGenerator() error = %v", err)
+	}
+
+	otherKey, err := otherGen.Generate(&photo)
+	if err != nil {
+		t.Fatalf("Generate() error = %v", err)
+	}
+
+	// Changing a character in the middle always changes the decoded bytes;
+	// the last character of unpadded base64 can carry unused bits.
+	middle := len(valid) / 2
+
+	replacement := "A"
+	if valid[middle] == 'A' {
+		replacement = "B"
+	}
+
+	changed := valid[:middle] + replacement + valid[middle+1:]
+
+	tests := []struct {
+		name       string
+		token      string
+		wantStatus int
+	}{
+		{"expired", expired, http.StatusGone},
+		{"one character changed", changed, http.StatusBadRequest},
+		{"cut short", valid[:middle], http.StatusBadRequest},
+		{"another signing key", otherKey, http.StatusBadRequest},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			rec := getEncToken(srv, tt.token)
+			t.Logf("GET /v1/e/%s/img.jpg: %d %s", tt.token, rec.Code, rec.Body)
+
+			if rec.Code != tt.wantStatus {
+				t.Errorf("status = %d, want %d", rec.Code, tt.wantStatus)
+			}
+		})
+	}
+}

internal/handlers/referer_blocklist_internal_test.go

@@ -0,0 +1,185 @@
+package handlers
+
+import (
+	"context"
+	"log/slog"
+	"net/http"
+	"net/http/httptest"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/go-chi/chi/v5"
+	"sneak.berlin/go/pixa/internal/allowlist"
+	"sneak.berlin/go/pixa/internal/encurl"
+	"sneak.berlin/go/pixa/internal/httpfetcher"
+	"sneak.berlin/go/pixa/internal/imgcache"
+)
+
+// blockedReferer is a page on leech.example, which newRefererRoutes puts on
+// referer_blocklist.
+const blockedReferer = "https://leech.example/page.html"
+
+// countingFetcher passes each fetch on to the fetcher it holds and counts it.
+type countingFetcher struct {
+	httpfetcher.Fetcher
+
+	fetches atomic.Int32
+}
+
+// Fetch counts the fetch and passes it on.
+func (f *countingFetcher) Fetch(
+	ctx context.Context, url string,
+) (*httpfetcher.FetchResult, error) {
+	f.fetches.Add(1)
+
+	return f.Fetcher.Fetch(ctx, url)
+}
+
+// newRefererRoutes returns both image routes of a Handlers whose
+// referer_blocklist is "leech.example" and ".hotlinker.example", the
+// Handlers, and the fetcher the routes fetch through. The JPEG at photoPath
+// exists on allowlistedHost and on signedHost.
+func newRefererRoutes(t *testing.T) (http.Handler, *Handlers, *countingFetcher) {
+	t.Helper()
+
+	fetcher := &countingFetcher{
+		Fetcher: newPhotoFetcher(t, allowlistedHost, signedHost),
+	}
+
+	cache, err := imgcache.NewCache(setupTestDB(t), imgcache.CacheConfig{
+		StateDir:    t.TempDir(),
+		CacheTTL:    time.Hour,
+		NegativeTTL: 5 * time.Minute,
+	})
+	if err != nil {
+		t.Fatalf("imgcache.NewCache() error = %v", err)
+	}
+
+	svc, err := imgcache.NewService(&imgcache.ServiceConfig{
+		Cache:      cache,
+		Fetcher:    fetcher,
+		SigningKey: testSigningKey,
+		Allowlist:  []string{allowlistedHost},
+	})
+	if err != nil {
+		t.Fatalf("imgcache.NewService() error = %v", err)
+	}
+
+	encGen, err := encurl.NewGenerator(testSigningKey)
+	if err != nil {
+		t.Fatalf("encurl.NewGenerator() error = %v", err)
+	}
+
+	h := &Handlers{
+		log:    slog.New(slog.DiscardHandler),
+		imgSvc: svc,
+		encGen: encGen,
+		refererBlocklist: allowlist.New(
+			[]string{"leech.example", ".hotlinker.example"}),
+	}
+
+	r := chi.NewRouter()
+	r.Get("/v1/image/*", h.HandleImage())
+	r.Get("/v1/e/{token}/*", h.HandleImageEnc())
+
+	return r, h, fetcher
+}
+
+// getWithReferer sends a GET for target to routes with referer as its
+// Referer header, or with none when referer is empty, and returns the
+// response.
+func getWithReferer(
+	t *testing.T, routes http.Handler, target, referer string,
+) *httptest.ResponseRecorder {
+	t.Helper()
+
+	req := httptest.NewRequestWithContext(t.Context(), http.MethodGet, target, nil)
+	if referer != "" {
+		req.Header.Set("Referer", referer)
+	}
+
+	rec := httptest.NewRecorder()
+
+	routes.ServeHTTP(rec, req)
+	t.Logf("GET %s with Referer %q: %d", target, referer, rec.Code)
+
+	return rec
+}
+
+// TestRefererBlocklist verifies that both image routes refuse a request whose
+// Referer names a host on referer_blocklist with 403 and the JSON error,
+// without fetching from the upstream host, and serve a request with no
+// Referer, one that does not parse, or one naming any other host. Hosts are
+// matched as allowlist_hosts matches them.
+func TestRefererBlocklist(t *testing.T) {
+	t.Parallel()
+
+	cases := []struct {
+		name    string
+		referer string
+		want    int
+	}{
+		{"no referer", "", http.StatusOK},
+		{"unlisted host", "https://unlisted.example/page.html", http.StatusOK},
+		{"unparseable", "%zz", http.StatusOK},
+		{"listed host", blockedReferer, http.StatusForbidden},
+		{"subdomain of listed host", "https://www.leech.example/", http.StatusOK},
+		{"subdomain of dot pattern", "https://www.hotlinker.example/a.html",
+			http.StatusForbidden},
+		{"dot pattern without its dot", "https://hotlinker.example/",
+			http.StatusForbidden},
+		{"host continuing past dot pattern",
+			"https://hotlinker.example.evil.example/", http.StatusOK},
+	}
+
+	// The photo's URL on each image route.
+	photoURLs := map[string]func(t *testing.T, h *Handlers) string{
+		"plain URL": func(t *testing.T, _ *Handlers) string {
+			t.Helper()
+
+			return photoURL(allowlistedHost)
+		},
+		"encrypted URL": encPhotoURL,
+	}
+
+	for urlName, photoURLFor := range photoURLs {
+		for _, tc := range cases {
+			t.Run(urlName+", "+tc.name, func(t *testing.T) {
+				t.Parallel()
+
+				routes, h, fetcher := newRefererRoutes(t)
+
+				rec := getWithReferer(t, routes, photoURLFor(t, h), tc.referer)
+
+				if tc.want == http.StatusOK {
+					requireServedPhoto(t, rec)
+
+					return
+				}
+
+				checkErrorBody(t, rec, http.StatusForbidden, "referer blocked")
+
+				if n := fetcher.fetches.Load(); n != 0 {
+					t.Errorf("upstream fetched %d times, want 0", n)
+				}
+			})
+		}
+	}
+}
+
+// TestBlockedRefererRefusedWhenImageIsCached verifies that a request whose
+// Referer is on referer_blocklist is refused even when the image it asks for
+// is already cached, so the answer does not depend on the cache.
+func TestBlockedRefererRefusedWhenImageIsCached(t *testing.T) {
+	t.Parallel()
+
+	routes, h, _ := newRefererRoutes(t)
+
+	for _, target := range []string{photoURL(allowlistedHost), encPhotoURL(t, h)} {
+		requireServedPhoto(t, getWithReferer(t, routes, target, ""))
+
+		rec := getWithReferer(t, routes, target, blockedReferer)
+		checkErrorBody(t, rec, http.StatusForbidden, "referer blocked")
+	}
+}

internal/imgcache/cache.go

@@ -37,11 +37,16 @@ type CacheConfig struct {
 	NegativeTTL time.Duration
 
 	// MaxBytes is the disk cache size limit in bytes that eviction
-	// enforces. Zero means no limit is enforced (no eviction). The
-	// config layer supplies the computed default when the operator
-	// omits cache_max_bytes.
+	// enforces. Zero means no limit is enforced (no eviction).
 	MaxBytes int64
 
+	// UseDefaultMaxBytes makes NewCache replace MaxBytes with the
+	// default limit: 75% of the sum of the space free on the filesystem
+	// holding the cache and the bytes the cache already holds, at least
+	// DefaultCacheMaxBytesFloor. The config layer sets this when the
+	// operator omits cache_max_bytes.
+	UseDefaultMaxBytes bool
+
 	// DisableDiskCache turns the disk cache off entirely: no cache
 	// directories are created, lookups always miss, stores are
 	// no-ops, and no eviction machinery runs. The config layer sets
@@ -95,6 +100,14 @@ type Cache struct {
 
 // NewCache creates a new cache instance.
 func NewCache(db *sql.DB, config CacheConfig) (*Cache, error) {
+	return newCache(db, config, defaultFreeSpaceProbe)
+}
+
+// newCache is NewCache with the free-space probe passed in, so tests
+// can fake the free space the default limit is worked out from.
+func newCache(
+	db *sql.DB, config CacheConfig, probe FreeSpaceProbeFunc,
+) (*Cache, error) {
 	log := config.Logger
 	if log == nil {
 		log = slog.Default()
@@ -145,6 +158,20 @@ func NewCache(db *sql.DB, config CacheConfig) (*Cache, error) {
 	c.variants = variants
 	c.srcMetadata = srcMetadata
 
+	if config.UseDefaultMaxBytes {
+		limit, err := c.computeDefaultMaxBytes(context.Background(), probe)
+		if err != nil {
+			return nil, err
+		}
+
+		c.config.MaxBytes = limit
+
+		log.Info("computed default cache size limit from free space and cache contents",
+			"cache_max_bytes", limit,
+			"cache_dir", filepath.Join(config.StateDir, "cache"),
+		)
+	}
+
 	return c, nil
 }
 

internal/imgcache/cachesize.go

@@ -0,0 +1,89 @@
+package imgcache
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"math"
+	"path/filepath"
+	"syscall"
+)
+
+// DefaultCacheMaxBytesFloor is the minimum computed default for the
+// cache_max_bytes setting: 500 MiB. The floor applies only to the
+// computed default (when the key is omitted from the configuration),
+// never to explicitly configured values.
+const DefaultCacheMaxBytesFloor int64 = 524288000
+
+// freeSpaceFractionNumerator and freeSpaceFractionDenominator express
+// the 75% share used for the computed default limit as integer
+// arithmetic (dividing before multiplying avoids overflow).
+const (
+	freeSpaceFractionNumerator   uint64 = 3
+	freeSpaceFractionDenominator uint64 = 4
+)
+
+var errNegativeBlockSize = errors.New("statfs reported negative block size")
+
+// FreeSpaceProbeFunc reports the number of free bytes available on the
+// filesystem containing path. It is a function type so tests can
+// inject a fake probe instead of depending on the host disk.
+type FreeSpaceProbeFunc func(path string) (uint64, error)
+
+// defaultFreeSpaceProbe reports free filesystem bytes via statfs on
+// the given path, as available to unprivileged processes.
+func defaultFreeSpaceProbe(path string) (uint64, error) {
+	var stat syscall.Statfs_t
+
+	err := syscall.Statfs(path, &stat)
+	if err != nil {
+		return 0, err
+	}
+
+	if stat.Bsize < 0 {
+		return 0, fmt.Errorf("%w %d for %q", errNegativeBlockSize, stat.Bsize, path)
+	}
+
+	blockSize := uint64(stat.Bsize)
+
+	return stat.Bavail * blockSize, nil
+}
+
+// computeDefaultMaxBytes returns the default cache size limit: 75% of
+// the sum of the free bytes probe reports for <state_dir>/cache/ and
+// the bytes the cache already holds, with a floor of
+// DefaultCacheMaxBytesFloor. Counting what the cache holds keeps the
+// limit from shrinking as the cache fills.
+func (c *Cache) computeDefaultMaxBytes(
+	ctx context.Context, probe FreeSpaceProbeFunc,
+) (int64, error) {
+	cacheDir := filepath.Join(c.config.StateDir, "cache")
+
+	freeBytes, err := probe(cacheDir)
+	if err != nil {
+		return 0, fmt.Errorf(
+			"default cache_max_bytes: cannot determine free space for %q: %w",
+			cacheDir, err)
+	}
+
+	usedBytes, err := c.UsageBytes(ctx)
+	if err != nil {
+		return 0, err
+	}
+
+	// Both terms are at most math.MaxInt64, so the sum cannot overflow.
+	//nolint:gosec // G115: UsageBytes sums file sizes, never negative
+	spaceBytes := min(freeBytes, math.MaxInt64) + uint64(usedBytes)
+
+	computed := spaceBytes / freeSpaceFractionDenominator * freeSpaceFractionNumerator
+	computed = min(computed, math.MaxInt64)
+
+	// gosec cannot see that min() above bounds computed, so it reads
+	// this conversion as potentially overflowing. It cannot: computed is
+	// at most math.MaxInt64 on every path here.
+	//nolint:gosec // G115: clamped to MaxInt64 by min above
+	limit := int64(computed)
+	limit = max(limit, DefaultCacheMaxBytesFloor)
+
+	return limit, nil
+}

internal/imgcache/cachesize_internal_test.go

@@ -0,0 +1,188 @@
+package imgcache
+
+import (
+	"errors"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// Static errors returned by the stub free-space probes below.
+var (
+	errTestStatfsFailed     = errors.New("statfs failed")
+	errTestProbeNotExpected = errors.New("probe must not be called")
+)
+
+// TestComputeDefaultMaxBytesCountsWhatTheCacheHolds verifies that the
+// default limit is 75% of the free space plus what the cache already
+// holds, so a cache filled to its limit keeps that limit across a
+// restart instead of shrinking to 75% of the space left free.
+func TestComputeDefaultMaxBytesCountsWhatTheCacheHolds(t *testing.T) {
+	t.Parallel()
+
+	cache, _ := newEvictionTestCache(t, 1<<30)
+
+	// Empty cache, 4 GiB free -> 3 GiB default.
+	got, err := cache.computeDefaultMaxBytes(t.Context(),
+		func(string) (uint64, error) { return 4294967296, nil })
+	if err != nil {
+		t.Fatalf("computeDefaultMaxBytes returned error: %v", err)
+	}
+
+	t.Logf("default for an empty cache with 4 GiB free: %d", got)
+
+	if got != 3221225472 {
+		t.Errorf("default for an empty cache = %d, want 3221225472 (75%% of 4 GiB)",
+			got)
+	}
+
+	// The cache now holds those 3 GiB, which leaves 1 GiB free.
+	_, err = cache.db.ExecContext(t.Context(),
+		`INSERT INTO variant_content (cache_key, size_bytes, content_type)
+		 VALUES (?, ?, ?)`,
+		string(testVariantKeyOne), 3221225472, testContentTypeWebP,
+	)
+	if err != nil {
+		t.Fatalf("failed to insert variant accounting row: %v", err)
+	}
+
+	got, err = cache.computeDefaultMaxBytes(t.Context(),
+		func(string) (uint64, error) { return 1073741824, nil })
+	if err != nil {
+		t.Fatalf("computeDefaultMaxBytes returned error: %v", err)
+	}
+
+	t.Logf("default for a cache holding 3 GiB with 1 GiB free: %d", got)
+
+	if got != 3221225472 {
+		t.Errorf("default for a cache holding 3 GiB with 1 GiB free = %d, "+
+			"want 3221225472 (75%% of 1 GiB + 3 GiB)", got)
+	}
+}
+
+// TestComputeDefaultMaxBytesAppliesFloor verifies that when 75% of the
+// free space plus what the cache holds is below 500 MiB, the default
+// is floored at DefaultCacheMaxBytesFloor.
+func TestComputeDefaultMaxBytesAppliesFloor(t *testing.T) {
+	t.Parallel()
+
+	cases := []struct {
+		name      string
+		freeBytes uint64
+	}{
+		{name: "100 MiB free", freeBytes: 104857600},
+		{name: "zero free", freeBytes: 0},
+		{name: "just below floor threshold", freeBytes: 699050665},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+
+			cache, _ := newEvictionTestCache(t, 1<<30)
+
+			got, err := cache.computeDefaultMaxBytes(t.Context(),
+				func(string) (uint64, error) { return tc.freeBytes, nil })
+			if err != nil {
+				t.Fatalf("computeDefaultMaxBytes returned error: %v", err)
+			}
+
+			if got != DefaultCacheMaxBytesFloor {
+				t.Errorf("computeDefaultMaxBytes = %d, want floor %d",
+					got, DefaultCacheMaxBytesFloor)
+			}
+		})
+	}
+}
+
+// TestComputeDefaultMaxBytesPropagatesProbeError verifies that a
+// failing free-space probe produces an error naming cache_max_bytes,
+// instead of a silently wrong default.
+func TestComputeDefaultMaxBytesPropagatesProbeError(t *testing.T) {
+	t.Parallel()
+
+	cache, _ := newEvictionTestCache(t, 1<<30)
+
+	_, err := cache.computeDefaultMaxBytes(t.Context(),
+		func(string) (uint64, error) { return 0, errTestStatfsFailed })
+	if err == nil {
+		t.Fatal("probe failure must produce an error, got nil")
+	}
+
+	t.Logf("got expected error: %v", err)
+
+	if !strings.Contains(err.Error(), "cache_max_bytes") {
+		t.Errorf("error %q does not name cache_max_bytes", err.Error())
+	}
+}
+
+// TestNewCacheComputesDefaultMaxBytesWhenAsked verifies that with
+// UseDefaultMaxBytes set, the cache's limit becomes the computed
+// default, and that the probe is pointed at <state_dir>/cache/, which
+// must be created first so statfs measures the right filesystem.
+func TestNewCacheComputesDefaultMaxBytesWhenAsked(t *testing.T) {
+	t.Parallel()
+
+	stateDir := t.TempDir()
+	wantCacheDir := filepath.Join(stateDir, "cache")
+
+	var probedPath string
+
+	// 4 GiB free -> 3 GiB default.
+	probe := func(path string) (uint64, error) {
+		probedPath = path
+
+		info, err := os.Stat(path)
+		if err != nil || !info.IsDir() {
+			t.Errorf("cache directory %q was not created before probing: info=%v err=%v",
+				path, info, err)
+		}
+
+		return 4294967296, nil
+	}
+
+	cache, err := newCache(evictionTestDB(t), CacheConfig{
+		StateDir:           stateDir,
+		UseDefaultMaxBytes: true,
+	}, probe)
+	if err != nil {
+		t.Fatalf("newCache returned error: %v", err)
+	}
+
+	if cache.config.MaxBytes != 3221225472 {
+		t.Errorf("MaxBytes = %d, want computed default 3221225472",
+			cache.config.MaxBytes)
+	}
+
+	if probedPath != wantCacheDir {
+		t.Errorf("free space probed at %q, want cache directory %q",
+			probedPath, wantCacheDir)
+	}
+}
+
+// TestNewCacheKeepsExplicitMaxBytes verifies that without
+// UseDefaultMaxBytes the cache keeps MaxBytes exactly as given and
+// never consults the free-space probe.
+func TestNewCacheKeepsExplicitMaxBytes(t *testing.T) {
+	t.Parallel()
+
+	probe := func(string) (uint64, error) {
+		t.Error("free-space probe must not be consulted for explicit values")
+
+		return 0, errTestProbeNotExpected
+	}
+
+	cache, err := newCache(evictionTestDB(t), CacheConfig{
+		StateDir: t.TempDir(),
+		MaxBytes: 1024,
+	}, probe)
+	if err != nil {
+		t.Fatalf("newCache returned error: %v", err)
+	}
+
+	if cache.config.MaxBytes != 1024 {
+		t.Errorf("MaxBytes = %d, want explicit 1024 (no floor, no recompute)",
+			cache.config.MaxBytes)
+	}
+}

internal/imgcache/eviction_internal_test.go

@@ -681,6 +681,11 @@ func TestEvictionRunsUnderWritePressure(t *testing.T) {
 	assertNoDanglingReferences(t, cache)
 }
 
+// TestEvictionRunsOnPeriodicSchedule writes three variant files straight
+// to disk, bypassing StoreVariant, so they have no accounting rows and no
+// write-pressure notification fires. Only a periodic reconciliation pass
+// can then adopt them, and only the eviction pass that follows it can
+// evict them.
 func TestEvictionRunsOnPeriodicSchedule(t *testing.T) {
 	t.Parallel()
 
@@ -688,13 +693,29 @@ func TestEvictionRunsOnPeriodicSchedule(t *testing.T) {
 
 	cache, _ := newEvictionTestCache(t, limit)
 
-	// Start the evictor while the cache is empty, then create tracked
-	// over-limit state WITHOUT going through the store methods, so no
-	// write-pressure notification fires and only the periodic ticker
-	// can trigger eviction.
+	// Hold the test database's only connection, so the startup pass
+	// waits for it after walking the still empty variant directory: the
+	// files written while it waits are first seen by a periodic pass.
+	conn, err := cache.db.Conn(t.Context())
+	if err != nil {
+		t.Fatalf("failed to take the database connection: %v", err)
+	}
+
+	defer func() { _ = conn.Close() }()
+
 	cache.StartEviction(100 * time.Millisecond)
 	defer func() { _ = cache.StopEviction(t.Context()) }()
 
+	deadline := time.Now().Add(5 * time.Second)
+
+	for cache.db.Stats().WaitCount == 0 {
+		if time.Now().After(deadline) {
+			t.Fatal("the startup pass never waited for the database")
+		}
+
+		time.Sleep(10 * time.Millisecond)
+	}
+
 	keys := []VariantKey{
 		testVariantKeyOne, testVariantKeyTwo, testVariantKeyThree,
 	}
@@ -703,25 +724,43 @@ func TestEvictionRunsOnPeriodicSchedule(t *testing.T) {
 	for i, key := range keys {
 		content := bytes.Repeat([]byte{fills[i]}, 1000)
 
-		_, err := cache.variants.Store(key, bytes.NewReader(content), "image/webp")
+		_, err = cache.variants.Store(key, bytes.NewReader(content), "image/webp")
 		if err != nil {
 			t.Fatalf("failed to store variant file: %v", err)
 		}
+	}
 
-		_, err = cache.db.ExecContext(t.Context(),
-			`INSERT INTO variant_content (cache_key, size_bytes, content_type)
-			 VALUES (?, ?, ?)`,
-			string(key), len(content), "image/webp",
-		)
-		if err != nil {
-			t.Fatalf("failed to insert variant accounting row: %v", err)
+	_ = conn.Close()
+
+	// Only one of the 1000-byte files fits under the limit: wait until
+	// the evictor has removed the other two.
+	stored := len(keys)
+	deadline = time.Now().Add(5 * time.Second)
+
+	for stored > 1 && time.Now().Before(deadline) {
+		time.Sleep(25 * time.Millisecond)
+
+		stored = 0
+
+		for _, key := range keys {
+			if cache.variants.Exists(key) {
+				stored++
+			}
 		}
 	}
 
-	usage := waitForUsageAtOrBelow(t, cache, limit, 5*time.Second)
+	if stored > 1 {
+		t.Fatalf("periodic schedule did not trigger eviction: %d of %d "+
+			"variant files still on disk, want at most 1", stored, len(keys))
+	}
+
+	usage, err := cache.UsageBytes(t.Context())
+	if err != nil {
+		t.Fatalf("UsageBytes failed: %v", err)
+	}
+
 	if usage > limit {
-		t.Errorf("periodic schedule did not trigger eviction: usage = %d, want <= %d",
-			usage, limit)
+		t.Errorf("usage after eviction = %d, want <= %d", usage, limit)
 	}
 
 	assertNoDanglingReferences(t, cache)
@@ -808,15 +847,28 @@ func TestPeriodicReconciliationAdoptsFileThatAppearsAfterStartup(t *testing.T) {
 
 	cache, _ := newEvictionTestCache(t, 1<<30)
 
-	const interval = 100 * time.Millisecond
+	// Hold the test database's only connection, so the startup pass
+	// waits for it after walking the still empty variant directory: the
+	// file written while it waits is first seen by a periodic pass.
+	conn, err := cache.db.Conn(t.Context())
+	if err != nil {
+		t.Fatalf("failed to take the database connection: %v", err)
+	}
 
-	cache.StartEviction(interval)
+	defer func() { _ = conn.Close() }()
+
+	cache.StartEviction(100 * time.Millisecond)
 	defer func() { _ = cache.StopEviction(t.Context()) }()
 
-	// Let startup reconciliation run and settle on an empty cache
-	// before introducing the untracked file, so the adoption we assert
-	// below can only be the work of a later, periodic pass.
-	time.Sleep(3 * interval)
+	deadline := time.Now().Add(5 * time.Second)
+
+	for cache.db.Stats().WaitCount == 0 {
+		if time.Now().After(deadline) {
+			t.Fatal("the startup pass never waited for the database")
+		}
+
+		time.Sleep(10 * time.Millisecond)
+	}
 
 	// Simulate a variant whose accounting insert failed after the
 	// process was already running and serving requests: the content
@@ -825,14 +877,16 @@ func TestPeriodicReconciliationAdoptsFileThatAppearsAfterStartup(t *testing.T) {
 	// insert had failed and only the file write had succeeded.
 	untracked := bytes.Repeat([]byte{0x41}, 900)
 
-	_, err := cache.variants.Store(
+	_, err = cache.variants.Store(
 		"aabbccdd0099", bytes.NewReader(untracked), "image/webp",
 	)
 	if err != nil {
 		t.Fatalf("failed to store untracked variant file: %v", err)
 	}
 
-	deadline := time.Now().Add(5 * time.Second)
+	_ = conn.Close()
+
+	deadline = time.Now().Add(5 * time.Second)
 
 	var usage int64
 

internal/middleware/middleware.go

@@ -35,13 +35,10 @@ const HSTSValue = "max-age=31536000; includeSubDomains"
 
 // ContentSecurityPolicyValue is the Content-Security-Policy header value.
 // default-src 'self' is the baseline and frame-ancestors 'none' is the primary
-// clickjacking control. 'unsafe-inline' is required in script-src and style-src
-// because the served templates carry inline onclick handlers (generator page)
-// and the bundled Tailwind asset injects a runtime <style> element; dropping it
-// needs template changes outside this issue's scope.
+// clickjacking control.
 const ContentSecurityPolicyValue = "default-src 'self'; " +
-	"script-src 'self' 'unsafe-inline'; " +
-	"style-src 'self' 'unsafe-inline'; " +
+	"script-src 'self'; " +
+	"style-src 'self'; " +
 	"object-src 'none'; " +
 	"base-uri 'self'; " +
 	"form-action 'self'; " +

internal/middleware/middleware_internal_test.go

@@ -325,6 +325,13 @@ func TestSecurityHeaders_PolicyHeaders(t *testing.T) {
 
 	handler.ServeHTTP(rec, req)
 
+	// The login and generator pages load their script and stylesheet from
+	// /static, so the policy allows no inline script or style.
+	csp := rec.Header().Get("Content-Security-Policy")
+	if strings.Contains(csp, "unsafe-inline") {
+		t.Errorf("Content-Security-Policy allows unsafe-inline: %q", csp)
+	}
+
 	tests := []struct {
 		header string
 		want   string
@@ -333,8 +340,8 @@ func TestSecurityHeaders_PolicyHeaders(t *testing.T) {
 		{
 			"Content-Security-Policy",
 			"default-src 'self'; " +
-				"script-src 'self' 'unsafe-inline'; " +
-				"style-src 'self' 'unsafe-inline'; " +
+				"script-src 'self'; " +
+				"style-src 'self'; " +
 				"object-src 'none'; " +
 				"base-uri 'self'; " +
 				"form-action 'self'; " +

internal/server/routes.go

@@ -53,7 +53,7 @@ func (s *Server) SetupRoutes() {
 	// Robots.txt
 	s.router.Get("/robots.txt", s.h.HandleRobotsTxt())
 
-	// Static files (Tailwind CSS, etc.)
+	// The login and generator pages' stylesheet and script
 	s.router.Handle("/static/*", http.StripPrefix("/static/", static.Handler()))
 
 	// Login/generator UI. The form routes carry CSRF protection; the

internal/static/generator.js

@@ -0,0 +1,10 @@
+// Generator page: a click on the generated URL selects it, and the Copy
+// button copies it. Both are on the page only once a URL has been generated.
+const generatedURL = document.getElementById("generated-url");
+
+if (generatedURL) {
+    generatedURL.addEventListener("click", () => generatedURL.select());
+    document.getElementById("copy-url").addEventListener("click", () => {
+        navigator.clipboard.writeText(generatedURL.value);
+    });
+}

internal/static/static.go

@@ -7,7 +7,7 @@ import (
 	"net/http"
 )
 
-//go:embed *.js
+//go:embed *.css *.js
 var files embed.FS
 
 // FS returns the embedded filesystem containing static files.

internal/static/style.css

@@ -0,0 +1,190 @@
+/* The login and generator pages. */
+
+* {
+    box-sizing: border-box;
+}
+
+body {
+    margin: 0;
+    min-height: 100vh;
+    background: #f3f4f6;
+    font-family: system-ui, sans-serif;
+    line-height: 1.5;
+}
+
+h1 {
+    margin: 0;
+    font-size: 1.5rem;
+    line-height: 2rem;
+    font-weight: 700;
+    color: #1f2937;
+}
+
+label {
+    display: block;
+    margin-bottom: 0.25rem;
+    font-size: 0.875rem;
+    font-weight: 500;
+    color: #374151;
+}
+
+input,
+select {
+    width: 100%;
+    padding: 0.5rem 0.75rem;
+    border: 1px solid #d1d5db;
+    border-radius: 0.375rem;
+    box-shadow: 0 1px 2px rgb(0 0 0 / 5%);
+    font: inherit;
+}
+
+input:focus,
+select:focus {
+    outline: none;
+    border-color: #3b82f6;
+    box-shadow: 0 0 0 2px #3b82f6;
+}
+
+button {
+    width: 100%;
+    padding: 0.5rem 1rem;
+    border: none;
+    border-radius: 0.375rem;
+    background: #2563eb;
+    color: #fff;
+    font: inherit;
+    cursor: pointer;
+    transition: background-color 0.15s;
+}
+
+button:hover {
+    background: #1d4ed8;
+}
+
+button:focus {
+    outline: 2px solid #3b82f6;
+    outline-offset: 2px;
+}
+
+form > * + * {
+    margin-top: 1rem;
+}
+
+.card {
+    padding: 1.5rem;
+    border-radius: 0.5rem;
+    background: #fff;
+    box-shadow:
+        0 4px 6px -1px rgb(0 0 0 / 10%),
+        0 2px 4px -2px rgb(0 0 0 / 10%);
+}
+
+.error {
+    margin-bottom: 1rem;
+    padding: 0.75rem 1rem;
+    border: 1px solid #f87171;
+    border-radius: 0.25rem;
+    background: #fee2e2;
+    color: #b91c1c;
+}
+
+/* Login page: the card centred on the screen. */
+
+.login {
+    display: flex;
+    align-items: center;
+    justify-content: center;
+}
+
+.login .card {
+    width: 100%;
+    max-width: 28rem;
+    padding: 2rem;
+}
+
+.login h1 {
+    margin-bottom: 1.5rem;
+    text-align: center;
+}
+
+/* Generator page. */
+
+.page {
+    max-width: 42rem;
+    margin: 0 auto;
+    padding: 2rem 1rem;
+}
+
+header {
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    margin-bottom: 2rem;
+}
+
+header a {
+    font-size: 0.875rem;
+    color: #4b5563;
+}
+
+header a:hover {
+    color: #1f2937;
+}
+
+.result {
+    margin-bottom: 1.5rem;
+    padding: 1rem;
+    border: 1px solid #bbf7d0;
+    border-radius: 0.5rem;
+    background: #f0fdf4;
+}
+
+.result h2 {
+    margin: 0 0 0.5rem;
+    font-size: 0.875rem;
+    font-weight: 500;
+    color: #166534;
+}
+
+.result div {
+    display: flex;
+    gap: 0.5rem;
+}
+
+.result input {
+    flex: 1;
+    border-color: #86efac;
+    box-shadow: none;
+    font-family: ui-monospace, monospace;
+    font-size: 0.875rem;
+}
+
+.result button {
+    width: auto;
+    padding: 0.5rem 0.75rem;
+    background: #16a34a;
+    font-size: 0.875rem;
+}
+
+.result button:hover {
+    background: #15803d;
+}
+
+.result p {
+    margin: 0.5rem 0 0;
+    font-size: 0.75rem;
+    color: #16a34a;
+}
+
+.columns {
+    display: grid;
+    grid-template-columns: repeat(2, minmax(0, 1fr));
+    gap: 1rem;
+}
+
+.note {
+    margin-top: 1rem;
+    font-size: 0.75rem;
+    color: #6b7280;
+    text-align: center;
+}

internal/templates/generator.html

@@ -4,52 +4,47 @@
     <meta charset="UTF-8">
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
     <title>Pixa - URL Generator</title>
-    <script src="/static/tailwind.js"></script>
+    <link rel="stylesheet" href="/static/style.css">
 </head>
-<body class="bg-gray-100 min-h-screen">
-    <div class="max-w-2xl mx-auto py-8 px-4">
-        <div class="flex justify-between items-center mb-8">
-            <h1 class="text-2xl font-bold text-gray-800">Pixa URL Generator</h1>
-            <a href="/logout" class="text-sm text-gray-600 hover:text-gray-800 underline">
+<body>
+    <div class="page">
+        <header>
+            <h1>Pixa URL Generator</h1>
+            <a href="/logout">
                 Logout
             </a>
-        </div>
+        </header>
 
         {{if .GeneratedURL}}
-        <div class="bg-green-50 border border-green-200 rounded-lg p-4 mb-6">
-            <h2 class="text-sm font-medium text-green-800 mb-2">Generated URL</h2>
-            <div class="flex gap-2">
+        <div class="result">
+            <h2>Generated URL</h2>
+            <div>
                 <input
                     type="text"
                     readonly
                     value="{{.GeneratedURL}}"
                     id="generated-url"
-                    class="flex-1 px-3 py-2 bg-white border border-green-300 rounded-md text-sm font-mono"
-                    onclick="this.select()"
-                >
-                <button
-                    onclick="navigator.clipboard.writeText(document.getElementById('generated-url').value)"
-                    class="px-3 py-2 bg-green-600 text-white rounded-md hover:bg-green-700 text-sm"
                 >
+                <button id="copy-url">
                     Copy
                 </button>
             </div>
-            <p class="text-xs text-green-600 mt-2">
+            <p>
                 Expires: {{.ExpiresAt}}
             </p>
         </div>
         {{end}}
 
         {{if .Error}}
-        <div class="bg-red-100 border border-red-400 text-red-700 px-4 py-3 rounded mb-6">
+        <div class="error">
             {{.Error}}
         </div>
         {{end}}
 
-        <form method="POST" action="/generate" class="bg-white rounded-lg shadow-md p-6 space-y-4">
+        <form method="POST" action="/generate" class="card">
             {{ .CSRFField }}
             <div>
-                <label for="url" class="block text-sm font-medium text-gray-700 mb-1">
+                <label for="url">
                     Source URL
                 </label>
                 <input
@@ -59,13 +54,12 @@
                     required
                     placeholder="https://example.com/image.jpg"
                     value="{{.FormURL}}"
-                    class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
                 >
             </div>
 
-            <div class="grid grid-cols-2 gap-4">
+            <div class="columns">
                 <div>
-                    <label for="width" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="width">
                         Width
                     </label>
                     <input
@@ -76,11 +70,10 @@
                         max="8192"
                         value="{{if .FormWidth}}{{.FormWidth}}{{else}}0{{end}}"
                         placeholder="0 = original"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
                     >
                 </div>
                 <div>
-                    <label for="height" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="height">
                         Height
                     </label>
                     <input
@@ -91,21 +84,16 @@
                         max="8192"
                         value="{{if .FormHeight}}{{.FormHeight}}{{else}}0{{end}}"
                         placeholder="0 = original"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
                     >
                 </div>
             </div>
 
-            <div class="grid grid-cols-2 gap-4">
+            <div class="columns">
                 <div>
-                    <label for="format" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="format">
                         Format
                     </label>
-                    <select
-                        id="format"
-                        name="format"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
-                    >
+                    <select id="format" name="format">
                         <option value="orig" {{if eq .FormFormat "orig"}}selected{{end}}>Original</option>
                         <option value="jpeg" {{if eq .FormFormat "jpeg"}}selected{{end}}>JPEG</option>
                         <option value="png" {{if eq .FormFormat "png"}}selected{{end}}>PNG</option>
@@ -115,14 +103,10 @@
                     </select>
                 </div>
                 <div>
-                    <label for="quality" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="quality">
                         Quality
                     </label>
-                    <select
-                        id="quality"
-                        name="quality"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
-                    >
+                    <select id="quality" name="quality">
                         <option value="25" {{if eq .FormQuality "25"}}selected{{end}}>Potato</option>
                         <option value="50" {{if eq .FormQuality "50"}}selected{{end}}>Low</option>
                         <option value="70" {{if eq .FormQuality "70"}}selected{{end}}>Medium</option>
@@ -132,16 +116,12 @@
                 </div>
             </div>
 
-            <div class="grid grid-cols-2 gap-4">
+            <div class="columns">
                 <div>
-                    <label for="fit" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="fit">
                         Fit Mode
                     </label>
-                    <select
-                        id="fit"
-                        name="fit"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
-                    >
+                    <select id="fit" name="fit">
                         <option value="cover" {{if eq .FormFit "cover"}}selected{{end}}>Cover</option>
                         <option value="contain" {{if eq .FormFit "contain"}}selected{{end}}>Contain</option>
                         <option value="fill" {{if eq .FormFit "fill"}}selected{{end}}>Fill</option>
@@ -150,14 +130,10 @@
                     </select>
                 </div>
                 <div>
-                    <label for="ttl" class="block text-sm font-medium text-gray-700 mb-1">
+                    <label for="ttl">
                         Expires In
                     </label>
-                    <select
-                        id="ttl"
-                        name="ttl"
-                        class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
-                    >
+                    <select id="ttl" name="ttl">
                         <option value="0" {{if or (eq .FormTTL "0") (eq .FormTTL "")}}selected{{end}}>Never</option>
                         <option value="60" {{if eq .FormTTL "60"}}selected{{end}}>1 minute</option>
                         <option value="3600" {{if eq .FormTTL "3600"}}selected{{end}}>1 hour</option>
@@ -169,17 +145,15 @@
                 </div>
             </div>
 
-            <button
-                type="submit"
-                class="w-full bg-blue-600 text-white py-2 px-4 rounded-md hover:bg-blue-700 focus:outline-none focus:ring-2 focus:ring-blue-500 focus:ring-offset-2 transition-colors"
-            >
+            <button type="submit">
                 Generate Encrypted URL
             </button>
         </form>
 
-        <p class="text-xs text-gray-500 mt-4 text-center">
+        <p class="note">
             Generated URLs are encrypted and cannot be modified. They will expire at the specified time.
         </p>
     </div>
+    <script src="/static/generator.js"></script>
 </body>
 </html>

internal/templates/login.html

@@ -4,22 +4,22 @@
     <meta charset="UTF-8">
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
     <title>Pixa - Login</title>
-    <script src="/static/tailwind.js"></script>
+    <link rel="stylesheet" href="/static/style.css">
 </head>
-<body class="bg-gray-100 min-h-screen flex items-center justify-center">
-    <div class="bg-white p-8 rounded-lg shadow-md w-full max-w-md">
-        <h1 class="text-2xl font-bold text-gray-800 mb-6 text-center">Pixa Image Proxy</h1>
+<body class="login">
+    <div class="card">
+        <h1>Pixa Image Proxy</h1>
 
         {{if .Error}}
-        <div class="bg-red-100 border border-red-400 text-red-700 px-4 py-3 rounded mb-4">
+        <div class="error">
             {{.Error}}
         </div>
         {{end}}
 
-        <form method="POST" action="/" class="space-y-4">
+        <form method="POST" action="/">
             {{ .CSRFField }}
             <div>
-                <label for="key" class="block text-sm font-medium text-gray-700 mb-1">
+                <label for="key">
                     Signing Key
                 </label>
                 <input
@@ -28,15 +28,11 @@
                     name="key"
                     required
                     autocomplete="current-password"
-                    class="w-full px-3 py-2 border border-gray-300 rounded-md shadow-sm focus:outline-none focus:ring-2 focus:ring-blue-500 focus:border-blue-500"
                     placeholder="Enter your signing key"
                 >
             </div>
 
-            <button
-                type="submit"
-                class="w-full bg-blue-600 text-white py-2 px-4 rounded-md hover:bg-blue-700 focus:outline-none focus:ring-2 focus:ring-blue-500 focus:ring-offset-2 transition-colors"
-            >
+            <button type="submit">
                 Login
             </button>
         </form>

scripts/manual-test.sh

@@ -1,147 +0,0 @@
-#!/bin/bash
-#
-# Manual test script for pixa server
-# Requires: server running on localhost:8080
-#
-set -e
-
-BASE_URL="${BASE_URL:-http://localhost:8080}"
-SIGNING_KEY="${SIGNING_KEY:-test-signing-key-for-development-only}"
-TEST_IMAGE_URL="https://s3.sneak.cloud/sneak-public/2021/2021-04-18.untitled.a7r4.07723.jpg"
-COOKIE_JAR=$(mktemp)
-
-cleanup() {
-    rm -f "$COOKIE_JAR"
-}
-trap cleanup EXIT
-
-pass() {
-    echo "✓ PASS: $1"
-}
-
-fail() {
-    echo "✗ FAIL: $1"
-    exit 1
-}
-
-echo "=== Pixa Manual Test Suite ==="
-echo "Base URL: $BASE_URL"
-echo ""
-
-# Test 1: Healthcheck
-echo "--- Test 1: Healthcheck endpoint ---"
-HEALTH=$(curl -sf "$BASE_URL/.well-known/healthcheck.json")
-if echo "$HEALTH" | grep -q '"status"'; then
-    pass "Healthcheck returns status"
-else
-    fail "Healthcheck did not return expected response"
-fi
-
-# Test 2: Login page displays
-echo "--- Test 2: Login page (GET /) ---"
-LOGIN_PAGE=$(curl -sf "$BASE_URL/")
-if echo "$LOGIN_PAGE" | grep -qi "password\|login\|sign"; then
-    pass "Login page displays password form"
-else
-    fail "Login page did not display expected content"
-fi
-
-# Test 3: Wrong password shows error
-echo "--- Test 3: Login with wrong password ---"
-WRONG_LOGIN=$(curl -sf -X POST "$BASE_URL/" -d "key=wrong-key" -c "$COOKIE_JAR")
-if echo "$WRONG_LOGIN" | grep -qi "invalid\|error\|incorrect\|wrong"; then
-    pass "Wrong password shows error message"
-else
-    fail "Wrong password did not show error"
-fi
-
-# Test 4: Correct password redirects to generator
-echo "--- Test 4: Login with correct signing key ---"
-curl -sf -X POST "$BASE_URL/" -d "key=$SIGNING_KEY" -c "$COOKIE_JAR" -b "$COOKIE_JAR" -L -o /dev/null
-GENERATOR_PAGE=$(curl -sf "$BASE_URL/" -b "$COOKIE_JAR")
-if echo "$GENERATOR_PAGE" | grep -qi "generate\|url\|source\|logout"; then
-    pass "Correct password shows generator page"
-else
-    fail "Generator page not displayed after login"
-fi
-
-# Test 5: Generate encrypted URL
-echo "--- Test 5: Generate encrypted URL ---"
-GEN_RESULT=$(curl -sf -X POST "$BASE_URL/generate" -b "$COOKIE_JAR" \
-    -d "url=$TEST_IMAGE_URL" \
-    -d "width=800" \
-    -d "height=600" \
-    -d "format=jpeg" \
-    -d "quality=85" \
-    -d "fit=cover" \
-    -d "ttl=3600")
-if echo "$GEN_RESULT" | grep -q "/v1/e/"; then
-    pass "Encrypted URL generated"
-    # Extract the encrypted URL
-    ENC_URL=$(echo "$GEN_RESULT" | grep -o '/v1/e/[^"<]*' | head -1)
-    echo "  Generated URL: $ENC_URL"
-else
-    fail "Failed to generate encrypted URL"
-fi
-
-# Test 6: Fetch image via encrypted URL
-echo "--- Test 6: Fetch image via encrypted URL ---"
-if [ -n "$ENC_URL" ]; then
-    HTTP_CODE=$(curl -sf -o /dev/null -w "%{http_code}" "$BASE_URL$ENC_URL")
-    if [ "$HTTP_CODE" = "200" ]; then
-        pass "Encrypted URL returns image (HTTP 200)"
-    else
-        fail "Encrypted URL returned HTTP $HTTP_CODE"
-    fi
-else
-    fail "No encrypted URL to test"
-fi
-
-# Test 7: Fetch image via allowlisted host (direct proxy)
-echo "--- Test 7: Fetch image via direct proxy (allowlisted host) ---"
-# URL format: /v1/image/<host>/<path>/<WxH>.<format>
-PROXY_PATH="/v1/image/s3.sneak.cloud/sneak-public/2021/2021-04-18.untitled.a7r4.07723.jpg/400x300.jpeg"
-HTTP_CODE=$(curl -sf -o /dev/null -w "%{http_code}" "$BASE_URL$PROXY_PATH")
-if [ "$HTTP_CODE" = "200" ]; then
-    pass "Direct proxy returns image (HTTP 200)"
-else
-    fail "Direct proxy returned HTTP $HTTP_CODE"
-fi
-
-# Test 8: Logout
-echo "--- Test 8: Logout ---"
-curl -sf "$BASE_URL/logout" -b "$COOKIE_JAR" -c "$COOKIE_JAR" -L -o /dev/null
-AFTER_LOGOUT=$(curl -sf "$BASE_URL/" -b "$COOKIE_JAR")
-if echo "$AFTER_LOGOUT" | grep -qi "password\|login"; then
-    pass "Logout redirects to login page"
-else
-    fail "Logout did not redirect to login"
-fi
-
-# Test 9: Generate short-TTL URL and verify expiration
-echo "--- Test 9: Expired URL returns 410 ---"
-# Login again
-curl -sf -X POST "$BASE_URL/" -d "key=$SIGNING_KEY" -c "$COOKIE_JAR" -b "$COOKIE_JAR" -L -o /dev/null
-# Generate URL with 1 second TTL
-GEN_RESULT=$(curl -sf -X POST "$BASE_URL/generate" -b "$COOKIE_JAR" \
-    -d "url=$TEST_IMAGE_URL" \
-    -d "width=100" \
-    -d "height=100" \
-    -d "format=jpeg" \
-    -d "ttl=1")
-SHORT_URL=$(echo "$GEN_RESULT" | grep -o '/v1/e/[^"<]*' | head -1)
-if [ -n "$SHORT_URL" ]; then
-    echo "  Waiting 2 seconds for URL to expire..."
-    sleep 2
-    HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" "$BASE_URL$SHORT_URL")
-    if [ "$HTTP_CODE" = "410" ]; then
-        pass "Expired URL returns 410 Gone"
-    else
-        fail "Expired URL returned HTTP $HTTP_CODE (expected 410)"
-    fi
-else
-    fail "Could not generate short-TTL URL"
-fi
-
-echo ""
-echo "=== All tests passed! ==="