p2: fence document tags, atomic fleetctx, harden workspace

Neutralise </document and <document in page text. Counters are
atomic.Uint64. Workspace names are [A-Za-z0-9_-] and PathEscaped;
foreign workspace_id rows are dropped. Wayback closest.url must be
https://web.archive.org. archive.today snapshot_at is Memento-Datetime
or null.
This commit is contained in:
2026-09-23 00:01:04 -07:00
parent 2f9d3d97e6
commit ff87a04f94
8 changed files with 149 additions and 17 deletions
+16
View File
@@ -69,6 +69,22 @@ func TestWaybackAvailable(t *testing.T) {
}
}
func TestWaybackRejectsForeignHost(t *testing.T) {
avail, _ := json.Marshal(map[string]any{
"archived_snapshots": map[string]any{
"closest": map[string]any{"available": true, "url": "https://evil.example/x", "timestamp": "20240101120000"},
},
})
f := &fakeFetch{
body: map[string][]byte{"https://archive.org/wayback/available": avail},
ct: map[string]string{"https://archive.org/wayback/available": "application/json"},
}
_, err := Try(context.Background(), f, "https://example.com/x")
if err == nil {
t.Fatal("want reject")
}
}
func TestWaybackUnavailableGoesToArchiveToday(t *testing.T) {
avail, _ := json.Marshal(map[string]any{"archived_snapshots": map[string]any{}})
page := []byte("<html><head><title>Saved</title></head><body>" + strings.Repeat("ok ", 200) + "</body></html>")