Investigate Resuming Downloads from Timed URLs

[Sewer56]

Spike

This is an open discussion topic, feel free to

Proposal

Currently, our system doesn't handle suspending (pausing) and unsuspending downloads for URLs that have an expiry date or time limit.

We need to investigate a strategy for resuming downloads from these timed URLs, whenever it is possible.

These URLs include NXM links served to free users and websites that generate temporary links.

Currently as is, we save the exact URL used to initiate the download upon suspending, attempting to restart it with that URL. In the case of NXM links, we preserve the nxm passed via IPC rather than any CDN URLs.

In any case, we can get failures when a user tries to resume the download after the URL has expired; and currently we don't present those errors to the user; leaving to inconvenience.

Impact

We need to ensure that our system can handle timed URLs appropriately.

Ideally, when a user attempts to resume a download from a known domain, the operation should be successful (if possible), or at least guide the user through getting a new link from the appropriate domain or download source.

In the case it is not possible to guide the user in getting a new URL, we should provide a clear error message at the very least if a resume operation fails due to an expired URL.

Supporting Information

Some questions include:

• How do we present an appropriate error message to the user?

  • To which level do we want to, or can guide user through getting new links.

• How do we handle unknown HTTP URLs?

  • Do we set rules by domain for well known domains?

• For NXM links, how can we implement resume for free users?

  • This includes the UX workflow towards getting a new link.
  • Talk to backend [if necessary]
  • Consider finding out how long free links last before expiring.

These questions, and potentially others, should be discussed.