Safer behavior around remote HTTP assets
We’ve seen an incident where the engine became unresponsive shortly after a remote HTTP asset (e.g. a badge or image URL) failed or hung. It’s unclear whether the mod was involved, but to be safe we’re changing how StuckAsset treats remote http/https assets.
What’s new
- skipHttpHttps (default: on) – The mod no longer cancels or retries jobs for
http://orhttps://URLs. Stuck detection and retries stay focused on local/session transfers, and we avoid touching the HTTP fetch path that can wedge the engine when a remote URL is dead or unresponsive. - maxRetriesForHttp (default: 0) – If you turn off
skipHttpHttps, plainhttp://URLs are still limited to 0 retries by default (configurable), since plain HTTP is more likely to hang.
Behavior for local:// (session) assets is unchanged. If you relied on the mod to clean up stuck remote HTTP downloads, you can set skipHttpHttps to false in the mod config (with the above tradeoffs in mind).