Install queue: a failing download blocks every later install, and none can be cancelled #74
Labels
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
schaefera/utStudyBIble#74
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Found by the owner on the phone after #72, via the Bible chooser's "+":
Seen (Plinius, build
a7f00d6):install GerSch downloadingtheninstall GerSch retryingin the same second, on every launch (16:40:39, 16:41:38).getent hosts ftp.crosswire.orggives 209.250.6.226, plus a NAT64 AAAA. So either the app's confined resolver fails (AppArmor, resolver socket, IPv6-first with no fallback), or an old failure from when the phone was on mobile data is being retried with a stale or blocked path.Wanted:
Blast radius: trunk (
core/src/tasks/,sword/src/install*,qt/glue).🤖 Generated with Claude Code
More evidence, from the phone at 16:56 on build
c0d97e6, over WiFi:GerSch.zip.partare gone. A copy of the stuck state is kept on the dev PC.downloading→retryingwithin the same second. A new 0-bytetasks/StrongsGreek.zip.partwas created, and the queue is softlocked again.So every download fails inside the app, not just GerSch, while the shell resolves
ftp.crosswire.org. The cause is in the app's networking path (resolver, AppArmor or socket family), not one module or source.🤖 Generated with Claude Code
claude-bot referenced this issue2026-10-23 15:10:01 +00:00
Merged in #76 (Review-Confidence 70). On the phone (build
c6ef183), the owner confirmed: "The actual download worked". StrongsGreek, Robinson and TSK installed at 17:46, which confirms thenetworkingpolicy group was the cause.StrongsHebrew failed in the same run, which is a separate packaging bug, now #77. It was marked failed and the rest of the queue carried on: the new no-blocking behaviour, seen live.
Still to try on the phone before sign-off:
🤖 Generated with Claude Code
This issue seems somewhat obsolete. Checked and all checks pass, so closing.