Matthew Meszaros 57a26c4ee6 feat: attribute a received warmup email moved to spam on the evidence around it before charging anyone, never for mail that arrived in spam (warmup_received.landed_spam), holding each move 30 minutes in warmup_spam_moves and deciding provider on a cross-workspace correlation or a move straight after arrival with nobody there, owner on provider-reported read, unread or star activity in mailbox_owner_activity or a repeated uncorrelated pattern in a used mailbox, nobody otherwise, weighting a spam move as one strike, withdrawing owner verdicts a later correlation explains, clearing the strikes behind a hold an admin lifts or an appeal approves, migration 000233, docs and guide

2026-09-30 06:52:32 +02:00

..

evaluate_own_pool_test.go

fix: take the warmup health sweep from twelve round-trips per mailbox to two, list participants stalest first, stop at the deadline, and return the standing from the write (#502)

2026-09-14 03:14:15 -07:00

health_evaluation_live_test.go

feat: make warmup spam placement only slow a mailbox down (watch at 10%, half-volume throttle at 20%, never quarantine, block or appeal) and judge it only at Google, Microsoft and Yahoo over verified deliveries, keep small-host spam out of the health bands, warmup and cold ramps and the advisor while still showing it beside the headline rate on the drawer and Deliverability, draw small-host partners whose own filter junks warmup mail less often, send each quarantine or block webhook once plus health_changed, release placement-only quarantines in 000220, and update the warmup, deliverability, advisor and API docs

2026-09-26 22:26:58 -07:00

live_pools_test.go

fix: seed the two warmup pools on every instance under fixed ids and make one pool per type structural, since the baseline squash dropped the insert and a fresh self-hosted instance never warmed; move memberships onto the canonical pools, scope the standing mirror trigger to the columns it mirrors so a pool move keeps a retention window, commit the runtime and every seeder to the ids through MoveToPool, assert the pools at boot and in a warmup_pools_missing health check, and drop the guide's claim of cross-tier borrowing the health gate rejects (#493)

2026-09-13 21:37:17 -07:00

metrics_counts_live_test.go

feat: make warmup spam placement only slow a mailbox down (watch at 10%, half-volume throttle at 20%, never quarantine, block or appeal) and judge it only at Google, Microsoft and Yahoo over verified deliveries, keep small-host spam out of the health bands, warmup and cold ramps and the advisor while still showing it beside the headline rate on the drawer and Deliverability, draw small-host partners whose own filter junks warmup mail less often, send each quarantine or block webhook once plus health_changed, release placement-only quarantines in 000220, and update the warmup, deliverability, advisor and API docs

2026-09-26 22:26:58 -07:00

pool_membership_live_test.go

fix: retire the warmup spam score, a counter that grew with volume rather than misbehaviour and that no band could act on (#508)

2026-09-14 07:44:34 -07:00

service_test.go

feat: attribute a received warmup email moved to spam on the evidence around it before charging anyone, never for mail that arrived in spam (warmup_received.landed_spam), holding each move 30 minutes in warmup_spam_moves and deciding provider on a cross-workspace correlation or a move straight after arrival with nobody there, owner on provider-reported read, unread or star activity in mailbox_owner_activity or a repeated uncorrelated pattern in a used mailbox, nobody otherwise, weighting a spam move as one strike, withdrawing owner verdicts a later correlation explains, clearing the strikes behind a hold an admin lifts or an appeal approves, migration 000233, docs and guide

2026-09-30 06:52:32 +02:00

service.go

feat: attribute a received warmup email moved to spam on the evidence around it before charging anyone, never for mail that arrived in spam (warmup_received.landed_spam), holding each move 30 minutes in warmup_spam_moves and deciding provider on a cross-workspace correlation or a move straight after arrival with nobody there, owner on provider-reported read, unread or star activity in mailbox_owner_activity or a repeated uncorrelated pattern in a used mailbox, nobody otherwise, weighting a spam move as one strike, withdrawing owner verdicts a later correlation explains, clearing the strikes behind a hold an admin lifts or an appeal approves, migration 000233, docs and guide

2026-09-30 06:52:32 +02:00

sweep_round_trips_test.go

fix: retire the warmup spam score, a counter that grew with volume rather than misbehaviour and that no band could act on (#508)

2026-09-14 07:44:34 -07:00

withdraw_tampering_test.go

feat: never add a warmup strike from a recheck (confirm it, or withdraw it when the message is still in the mailbox or retention removed it since), revise a tampering hold before its strike is deleted so a retry completes it, compare-and-swap the revision on the hold's term, scope the ledger revision to the whole address, keep tampering events 37 days so a live hold's strikes survive pruning, answer an inconclusive search as unknown so IMAP rechecks stop, and return an IMAP transport failure as an error

2026-09-29 05:12:36 -07:00