# Replay capture queue: recheck before every send, survive storage errors · dara/icarus

[View on GitCafe](https://git.cafe/dara/icarus/commit/9257881391c9fbcf1d894424bda49d540caab3b5)

Repository: [dara/icarus](https://git.cafe/dara/icarus)

Visibility: public

Requested revision: 9257881391c9fbcf1d894424bda49d540caab3b5

Requested commit: 9257881391c9fbcf1d894424bda49d540caab3b5

Commit: 9257881391c9fbcf1d894424bda49d540caab3b5

Tree: 780da252fa9766e20014ec7c1f837d87581a1a22

Author: Dara Adedeji

Committer: Dara Adedeji

## Message

```
Replay capture queue: recheck before every send, survive storage errors

Astra round 3:
- Item delivery that resumes after the page landed (a retry, a lost
  answer) now reads the strategy again first, so a strategy deleted or
  moved to another map meanwhile fails the capture instead of taking the
  positions; a page deleted meanwhile says so. A whole batch refused as
  not found marks the strategy lost.
- A failed write of a capture's progress no longer escapes delivery: the
  capture is kept in memory as stopped on this device, shown with Retry
  on the capture card and the library card, and blocks sign-out; other
  strategies' captures keep going.
- A capture goes into the open cloud folder only when it is this
  account's own; a shared folder, or one the server refuses, leaves it
  at the top level.
- A server that needs a newer Icarus is noted, so the app's update notice
  shows, and the card says to update Icarus to finish saving.
- The card shows a send whose answer was lost as unconfirmed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

```

## Parents

- [cfc9a2720f0802d79ccef6edd4c4171461de2e83](https://git.cafe/dara/icarus/commit/cfc9a2720f0802d79ccef6edd4c4171461de2e83?format=markdown)

[Source at this commit](https://git.cafe/dara/icarus/tree/9257881391c9fbcf1d894424bda49d540caab3b5?format=markdown)