fix: a cancelled or timed-out crawl stores what it fetched and ends

The crawl job's context ends on a cancel or at the job timeout. Storing
the pages, links and issues now runs on a context detached from it with
its own 10-minute bound, a timeout ends the crawl failed with a
time-limit error, a LoadLimits failure falls back to the defaults, and a
fetch cut short by the cancellation is recorded as canceled.

classes/crawl/crawler.go

@@ -247,6 +247,10 @@ func Crawl(ctx context.Context, f *Fetcher, opt Options) (*Result, error) {
 					return
 				}
 				p.Resp = f.Get(ctx, it.key)
+				if p.Resp.Err != "" && ctx.Err() != nil {
+					// The fetch was cut short by the cancellation.
+					p.Resp = &Response{Err: "canceled"}
+				}
 				p.FetchedAt = time.Now().UTC()
 				if p.Resp.Err == "" && p.Resp.Status == 200 && p.Resp.HTML() {
 					base, _ := url.Parse(f.Absolute(it.key))