refactor(archive) tab split (phase 4 pt 3) : delete + maintenance helpers

[ebuzerdrmz44]

Description

Continues the ArchiveTab breakup started in #2519 (pt 1) and #2525 (pt 2, diff + rename). This peels the remaining destructive/maintenance actions out of the god-object into the same collaborator-class pattern under views/archive/:

• archive/archive_delete.py ArchiveDelete (delete_action / delete_result)
• archive/archive_maintenance.py ArchiveMaintenance (check, compact, prune actions + results, save_prune_setting, toggle_compact_button_visibility)

archive_tab.py shrinks by 130 lines and now just wires these helpers alongside the pt 2 ArchiveDiff / ArchiveRename.

No behavior change .

Notes

• Stacks on refactor(archive): peel diff and rename into views/archive/ helpers (Phase 4 pt 2) #2525. Until that merges, this PR's diff includes pt 2's commit; review only the top commit here. Kept as draft for that reason.
• The helpers call the tab's _set_status / _toggle_all_buttons. De-underscoring that tab/helper boundary is a deliberate later cleanup, kept consistent with pt 2 rather than widened here.
• Test call sites updated to the new module paths (tests/unit/test_archives.py, tests/integration/test_archives.py).

[m3nu]

Reviewed the pt-3 commit (93a4c93) in depth — it's a faithful, line-by-line move of delete/check/compact/prune into ArchiveDelete/ArchiveMaintenance, matching the extract/mount collaborator pattern. Selection goes through selected_archives() / ArchiveRole, so it's proxy-safe; helpers are strongly referenced on the tab, so the signal lifecycle is sound; all test call sites are updated with no stale patches.

Two nits before merge:

1. The description says toggle_compact_button_visibility moved into ArchiveMaintenance — it actually stays on the tab, which is correct (application.py calls archiveTab.toggle_compact_button_visibility()), but please fix the description.
2. The application.py string-literal cleanup hunk is unrelated to this PR. Harmless, but out of scope.

Once #2525 is squash-merged, rebase onto master (the diff collapses to the pt-3 commit) and un-draft — then this is good to go. Optional but cheap: a smoke test each for check_action/compact_action/prune_action, which currently have no direct coverage (pre-existing gap, not something this PR introduced).

The merge-base changed after approval.

[ebuzerdrmz44]

@m3nu rebased onto master now that #2525 is in and addressed the review nits. The old approval got auto-dismissed when I pushed, so this should just need a quick re-confirm, not a full re-review.

Also heads up: #2523 is waiting on this one. It deletes ArchiveSortProxyModel and touches the same import block in archive_tab.py, so we can rebase it once this lands instead of fighting the same conflict twice.

[m3nu]

Re-reviewed the current head after the rebase onto master. The pt-3 move is faithful and correct — approving.

What I verified:

• ArchiveDelete / ArchiveMaintenance follow the stateless-collaborator pattern from archive_extract.py / archive_mount.py, and both are held as strong attributes on the tab, so the job.result.connect(...) lifetimes are sound.
• Selection goes through self.tab.selected_archives() → ArchiveRole, so the proxy mapToSource footgun is handled.
• Every call site is updated: the bDelete / compactButton / bPrune / bCheck connections, the archiveitem_contextmenu button_connection_pairs list, the prune-interval valueChanged connections, and both test files. No stale references left.
• test_populate_does_not_overwrite_prune_keep_within (the #2493 regression) exercises the real signal path rather than patching the method, so it still guards after the move.
• toggle_compact_button_visibility correctly stayed on the tab — application.py calls archiveTab.toggle_compact_button_visibility() directly.

Two things from my earlier review are still open on this head — please fix before you hit merge:

1. The PR description still lists toggle_compact_button_visibility as moving into ArchiveMaintenance. It didn't, and that's the right call. Worth correcting so future archaeology isn't misled.
2. The application.py string-literal hunk (line 303, implicit concat → single literal) is still in the diff. Runtime-identical and harmless to the .ts catalogs, but unrelated to the tab split. Drop it or split it out.

Neither blocks, hence the approval rather than another round.

Optional, pre-existing gap: check_action / compact_action / prune_action still have no direct coverage. A smoke test each, mirroring test_archive_delete, would be cheap insurance now that they live in a new module.

Merge order note: #2523 conflicts on the archive_tab.py import block, so it rebases after this lands.