Justify the handlers tests' start limit with a measurement (closes #225)

The cost that pushed internal/handlers near its 90s package timeout, and test app starts past fx's 15s start timeout, was every start hashing the admin password at 64 MB; that went with #404. Measured since under the host's real load, the package takes 10.0s and the slowest test app start 0.19s, so neither limit changes. newTestApp now says what its start limit catches and the measurement behind it, and script/test notes that its figures predate that change and gives the current ones. No test changes what it asserts.

Model: opus-5-5

internal/handlers/handlers_test.go

@@ -78,6 +78,11 @@ func (r *recordingEvictor) Evicted() []string {
 	return out
 }
 
+// newTestApp builds the handlers with their real dependencies. Its
+// RequireStart fails the test when starting takes longer than fx's
+// default start timeout of 15s. That limit catches a start that hangs,
+// not a busy host: measured with make test on 2026-10-02 at host load
+// 40-52 on 48 cores, the slowest of this package's 182 starts took 0.19s.
 func newTestApp(
 	t *testing.T,
 	targets ...any,

script/test

@@ -2,9 +2,10 @@
 # script/test: run the test suite.
 #
 # -timeout is applied by `go test` per package, not to the run as a whole, so
-# it only has to clear the slowest single package. That is internal/handlers,
-# measured in a cache-defeated builder stage on the 48-core shared build host
-# (2026-08-18); load- and host-dependent, not invariants:
+# it only has to clear the slowest single package. When this budget was set
+# that was internal/handlers, measured in a cache-defeated builder stage on the
+# 48-core shared build host (2026-08-18); load- and host-dependent, not
+# invariants:
 #
 #   16.9s                  host load 5-20, GOMAXPROCS 48
 #   45.9s / 47.3s / 49.0s  three runs at deliberate host load 31-73
@@ -23,6 +24,11 @@
 # a condition CI runs under. If a CPU-limited runner ever puts a real run near
 # 67s, that is the datum to revisit the org figure with.
 #
+# Those figures predate tests hashing the admin password at 1 MB instead of
+# 64 MB (https://git.eeqj.de/sneak/webhooker/pulls/404). After that change, in
+# a cache-defeated build at host load 41-55 (2026-10-02), internal/handlers
+# took 10.0s and the slowest package was internal/database at 16.2s.
+#
 # -p 4 -parallel 8 keep the run under 2 GB of memory: at most four test
 # binaries build or run at once, each with at most eight parallel tests. Under
 # -race every test binary and every link costs a few hundred MB, so the