Hi Luke, Quoting Mondy, Luke (Data61, Kensington NSW) (2020-09-15 05:51:31) > Currently I use this curl one-liner to resolve an arbitrary date into a real snapshot date, though it's a bit of a hack: > SNAPSHOT_DATE=$(basename $(curl -Ls -o /dev/null -w %{url_effective} http://snapshot.debian.org/archive/debian/$(date -u +%Y%m%dT%H%I00Z)/) ) > > Is that something you can use? aha, the only times I interacted with the HTML pages was indeed to scrape the correct list of timestamps, so I never realized that the HTML pages would return a HTTP 301 with a Location header if I threw a "wrong" timestamp at it. Thanks, this fixes the bug for me. Small improvement on your line above: instead of making a GET request and downloading the full HTML, you could just make a HEAD request with the -I option and save a little bit of network bandwidth with that. Thanks! cheers, josch
Attachment:
signature.asc
Description: signature