MCPcopy Create free account

hub / github.com/PerpetualSoftware/pad / functions

Functions28,840 in github.com/PerpetualSoftware/pad

FunctionTestPushDispatch_LeaseExceedsSendTimeout
--- lease invariant (r6-Q1) ---
internal/server/task3939b_push_dispatch_test.go:101
FunctionTestPushDispatch_LoopRunsTicksAndStops
(t *testing.T)
internal/server/task3939b_push_dispatch_test.go:402
FunctionTestPushDispatch_MentionAndCommentRowsShareOnePush
A comment row and a mention row for one comment are two rows of one group: both settle sent, neither is dropped as a duplicate, and they share ONE pus
internal/server/task3939b_push_gates_test.go:282
FunctionTestPushDispatch_MessageShape
The message is content-free: no title, comment text or names, and its fields come from the right places.
internal/server/task3939b_push_gates_test.go:462
FunctionTestPushDispatch_NonSentResultsReleaseWithAnAttempt
(t *testing.T)
internal/server/task3939b_push_dispatch_test.go:59
FunctionTestPushDispatch_OneDeviceSentIsSent
A unit with one device sent and another retried is sent: re-pushing would duplicate to the device that already has it.
internal/server/task3939b_push_dispatch_test.go:82
FunctionTestPushDispatch_ParkedEntryAtTheCutoffSecondIsKept
TASK-4032. The drop rule compares against the same second-truncated cutoff the claim uses, so a row created exactly on the cutoff second is claimable
internal/server/task4032_push_record_failure_test.go:576
FunctionTestPushDispatch_ParkedEntryClearsWhenTheRecordAffectsNoRow
A parked row that was settled some other way (here skipped) records nothing when retried; the entry must still go, or it is retried forever.
internal/server/task4032_push_record_failure_test.go:412
FunctionTestPushDispatch_ParkedRowDoesNotRideWithASibling
A parked row on an item must not pull a sibling row into its own push.
internal/server/task4032_push_record_failure_test.go:362
FunctionTestPushDispatch_ParkedRowsCountTowardTheCap
Delivered-but-unrecorded rows count toward the user's 30/h cap.
internal/server/task4032_push_record_failure_test.go:385
FunctionTestPushDispatch_ParkedRowsCountTowardTheCapForASummary
(t *testing.T)
internal/server/task4032_push_record_failure_test.go:339
FunctionTestPushDispatch_ParkedRowsDoNotStarveTheClaimBatch
Parked rows are never claimed: with more of them than one claim batch, a fresh pending row is still claimed and sent.
internal/server/task4032_push_record_failure_test.go:290
FunctionTestPushDispatch_PartialRecordClearsTheChunksThatLanded
TASK-4032. A retry whose record fails in a later chunk has already committed the earlier ones; those ids must leave the parked set, or each counts twi
internal/server/task4032_push_record_failure_test.go:478
FunctionTestPushDispatch_PruneRunsAtMostHourly
(t *testing.T)
internal/server/task3939b2a_push_test.go:511
FunctionTestPushDispatch_RateCap
--- the cap ---
internal/server/task3939b2a_push_test.go:134
FunctionTestPushDispatch_RateCapAppliesToTheSummary
The cap is decided on the summary as a unit, after the collapse: 29 sent plus a 4-group summary is admitted whole, not capped per group.
internal/server/task3939b2a_push_test.go:179
FunctionTestPushDispatch_RateCapCountsRowsAdmittedThisTick
Rows admitted earlier in the same tick count against the cap.
internal/server/task3939b2a_push_test.go:165
FunctionTestPushDispatch_RateCapIsPerUser
(t *testing.T)
internal/server/task3939b2a_push_test.go:192
FunctionTestPushDispatch_RealMentionEnqueueThroughTheDispatcher
The mention leg runs the REAL enqueue: a comment @-mentioning bob is written through the store, and nothing seeds a row by hand. The dispatcher then p
internal/server/task3939b2b_notification_test.go:272
FunctionTestPushDispatch_RecordFailureBoundsResends
(t *testing.T)
internal/server/task4032_push_record_failure_test.go:55
FunctionTestPushDispatch_RecordFailureMixedChunkStillSettles
(t *testing.T)
internal/server/task4032_push_record_failure_test.go:145
FunctionTestPushDispatch_RecordFailureRecovers
(t *testing.T)
internal/server/task4032_push_record_failure_test.go:72
FunctionTestPushDispatch_RecordRetriedWithoutAGateway
With no gateway configured the tick still records what it delivered.
internal/server/task4032_push_record_failure_test.go:95
FunctionTestPushDispatch_ReregisterKeepsPendingRowDeliverable
TASK-4027. A same-user re-register keeps the device's created_at, so a row created between the first registration and the re-register is still SENT: G
internal/server/task4027_push_reregister_test.go:12
FunctionTestPushDispatch_ResultMatrix
(t *testing.T)
internal/server/task3939b2a_push_test.go:234
FunctionTestPushDispatch_RetentionSevenDays
(t *testing.T)
internal/server/task3939b2a_push_test.go:473
FunctionTestPushDispatch_RetryDelayEqualsTheLease
A retry release holds the row for one lease (TASK-3939 B2a): not reclaimable before it has run out, reclaimable after.
internal/server/task3939b2a_push_test.go:292
FunctionTestPushDispatch_RetryUnrecordedClearsOnlyLandedIDs
TASK-4051: a retry removes committed chunks and keeps only the failed chunk's ids parked. Kills: M3b, M3c. The tick caller at push_dispatch.go:824 is
internal/server/task4051_push_settle_test.go:14
FunctionTestPushDispatch_RowDeletedMidSendIsNotAnError
3946's supersede can DELETE a pending row the dispatcher has claimed. The settle that finds it gone is a no-op, not an error and not a log line.
internal/server/task3939b_push_dispatch_test.go:349
FunctionTestPushDispatch_SameSecondReclaimIgnoresIDOrder
TASK-4031: two users registered one phone in the same second and the PREVIOUS registrant (bob) holds the larger id. Ownership follows the registration
internal/server/task4027_push_reregister_test.go:69
FunctionTestPushDispatch_SendsOnePerUnitAndSettlesSent
TASK-3939 part B1: the dispatch core. Gate cases live in task3939b_push_gates_test.go.
internal/server/task3939b_push_dispatch_test.go:21
FunctionTestPushDispatch_SharedTokenNeverCrossesUsers
Shared FCM token: push_devices is unique on (user, token), so two users can hold one phone's token. Only the newest LIVE registrant owns it.
internal/server/task3939b_push_gates_test.go:243
FunctionTestPushDispatch_ShutdownUnclaimsWithoutAnAttempt
--- shutdown ---
internal/server/task3939b_push_dispatch_test.go:369
FunctionTestPushDispatch_SlowSendNotReclaimedMidFlight
The Send runs under the tick's deadline (start + lease - margin), so it ends before the claim lapses; while it runs, and while the tick still holds th
internal/server/task3939b_push_dispatch_test.go:147
FunctionTestPushDispatch_SummaryAtFourGroupsNotThree
--- summary ---
internal/server/task3939b2a_push_test.go:58
FunctionTestPushDispatch_SummaryCountsOnlyNewRowsParkedStillCap
Parked rows are not part of the summary: it counts only the new rows. They still count toward the 30/h cap.
internal/server/task4032_push_record_failure_test.go:316
FunctionTestPushDispatch_SummaryCountsOnlyVisible
(t *testing.T)
internal/server/task3939b2a_push_test.go:106
FunctionTestPushDispatch_SummaryMessageShape
(t *testing.T)
internal/server/task3939b2a_push_test.go:75
FunctionTestPushDispatch_UnavailableThenDeadlineStillWarns
A tick that stops at the deadline after an unavailable chunk still warns once for it (TASK-3939).
internal/server/task3939b2a_push_test.go:417
FunctionTestPushDispatch_VisibilityErrorsLeakNothingToTheLog
The dispatcher shares computeWatchAccessVisibility's lookups but not its logging: that function logs the workspace id, user id and raw store error, an
internal/server/task3939b_push_gates_test.go:385
FunctionTestPushDispatch_ZeroMessageUnitIsSkipped
A unit none of whose devices can be given a message (NewMessage refuses an unknown kind, which the gates exclude, so the unit is built by hand) is set
internal/server/task3939b_push_gates_test.go:416
FunctionTestPushItemSendsResolvedAgentHeader
TestPushItemSendsResolvedAgentHeader (IDEA-2544 Phase 1, dispatcher review) pins that PushItem inherits X-Pad-Agent the same as every other mutating c
internal/cli/agent_identity_test.go:157
FunctionTestPushMessageBoundIsWhatTheWebComposerMirrors
The fixture's whitespace cases are only meaningful if they agree with the bound the handler actually enforces. Asserted THROUGH the endpoint at the bo
internal/server/push_message_collapse_test.go:125
FunctionTestPushMessageCollapseFixture
This file is one half of a cross-language contract test (PLAN-2558 S3). handlePushToItem measures a push message in runes AFTER whitespace collapse a
internal/server/push_message_collapse_test.go:54
FunctionTestPushMessageCollapseThroughTheHandler
The test above asserts the EXPRESSION handlePushToItem uses, copied into the test — which cannot notice the handler switching to a different normaliza
internal/server/push_message_collapse_test.go:80
FunctionTestPushOneShotPerTargetTime
(t *testing.T)
internal/store/commitment_move_test.go:210
FunctionTestPushRefusalCodesMatchTheWebAllowList
TestPushRefusalCodesMatchTheWebAllowList pins the two surfaces together. They answer the same question about the same endpoint, so drift means a push
cmd/pad/cmd_push_ambiguity_test.go:96
FunctionTestPushRefusedBeforePublishing
BUG-2699 — the CLI's half of "is this safe to re-run?". The push handler publishes BEFORE it writes its response, so an error is not proof that nothi
cmd/pad/cmd_push_ambiguity_test.go:28
FunctionTestPushSessionVisibility_MissingUserIsNotVisible
TestPushSessionVisibility_MissingUserIsNotVisible covers the resolver's "definitively gone" branch: a user who does not exist resolves to not-visible,
internal/server/handlers_push_delivery_symmetry_test.go:190
FunctionTestPushSessionVisibility_ResolvesAtMostTwice
TestPushSessionVisibility_ResolvesAtMostTwice pins the bound the fix rests on: however many sessions are counted, visibility is resolved at most once
internal/server/handlers_push_visibility_test.go:306
FunctionTestPushTables_Cascades
Kills: removing push_devices.session_id's cascade, user_notifications' user_id cascade, user_notification_settings' cascade, or the actor's SET NULL.
internal/store/notification_settings_test.go:113
FunctionTestPushTables_PurgeEmptiesTheQueue
Kills: dropping the workspace_id cascade.
internal/store/notification_settings_test.go:174
FunctionTestPushToItem_BroadcastCountsSessionsOnEveryInstance
TestPushToItem_BroadcastCountsSessionsOnEveryInstance covers the OTHER half BUG-2698 filed: delivered_sessions was a local count describing a global d
internal/server/handlers_push_cross_instance_test.go:148
FunctionTestPushToItem_BroadcastDeliveredSessionsCountsArmedOnly
TestPushToItem_BroadcastDeliveredSessionsCountsArmedOnly is PLAN-2613 S1's honesty requirement for the broadcast (untargeted) count: with TWO of the c
internal/server/handlers_push_test.go:521
FunctionTestPushToItem_BroadcastDeliveredSessionsCountsLiveSessions
TestPushToItem_BroadcastDeliveredSessionsCountsLiveSessions covers the broadcast (omitted target_session_id) mode with delivered_sessions wired up: it
internal/server/handlers_push_test.go:471
FunctionTestPushToItem_BroadcastDoesNotCountSessionsDeliveryWillDrop
TestPushToItem_BroadcastDoesNotCountSessionsDeliveryWillDrop is the BUG-2725 over-count regression. It FAILS against the unfixed code, which reported
internal/server/handlers_push_visibility_test.go:138
FunctionTestPushToItem_BroadcastStillCountsVisibleSessions
TestPushToItem_BroadcastStillCountsVisibleSessions is the control leg, and the reason the test above means anything. Same fixture, same admin, same w
internal/server/handlers_push_visibility_test.go:179
FunctionTestPushToItem_BroadcastWithUnreadablePresenceStillPublishes
TestPushToItem_BroadcastWithUnreadablePresenceStillPublishes: refusing here would manufacture an outage — presence never gated a broadcast, so the del
internal/server/handlers_push_presence_unreadable_test.go:74
FunctionTestPushToItem_ClosedBusIsRefused
TestPushToItem_ClosedBusIsRefused: ErrBusClosed proves nothing was published, so the caller gets the same 503 `unavailable` the nil-bus branch already
internal/server/handlers_push_publish_result_test.go:71
FunctionTestPushToItem_DeliversSelfAddressed
TestPushToItem_DeliversSelfAddressed mirrors TestWatchEventsStream_AddressedToYou_Assignment: a push must reach the SAME user's own connected stream w
internal/server/handlers_push_test.go:146
FunctionTestPushToItem_HandlerPassesRealVisibility
TestPushToItem_HandlerPassesRealVisibility is the CONVE-19 binding test. Every other test in this file could pass while the push HANDLER still called
internal/server/handlers_push_visibility_test.go:351
FunctionTestPushToItem_InvisibleItemDenied
TestPushToItem_InvisibleItemDenied covers the same resolve/visibility gate every other item-scoped handler enforces: pushing a nonexistent ref 404s.
internal/server/handlers_push_test.go:183
FunctionTestPushToItem_MessageAtCapSurvivesIntact
TestPushToItem_MessageAtCapSurvivesIntact proves a message whose COLLAPSED length is exactly at maxPushMessageLen is accepted and delivered byte-for-b
internal/server/handlers_push_test.go:97
FunctionTestPushToItem_OmittedTargetSessionIDMatchesPreS5RequestShape
TestPushToItem_OmittedTargetSessionIDMatchesPreS5RequestShape is the wire-level compatibility leg: a request body in EXACTLY the pre-S5 shape — no tar
internal/server/handlers_push_test.go:577
FunctionTestPushToItem_ReadablePresenceStillReportsACount
TestPushToItem_ReadablePresenceStillReportsACount is the positive control for both. Without it, a handler that reported null for every push — or refus
internal/server/handlers_push_presence_unreadable_test.go:121
FunctionTestPushToItem_RejectsOverLongMessage
TestPushToItem_RejectsOverLongMessage covers the 400-over-cap validation: a message whose COLLAPSED length exceeds maxPushMessageLen is rejected rathe
internal/server/handlers_push_test.go:80
FunctionTestPushToItem_RequiresMessage
TestPushToItem_RequiresMessage covers the 400-on-empty(-after-trim) validation: an absent message and a whitespace-only one are both rejected, so a bl
internal/server/handlers_push_test.go:19
FunctionTestPushToItem_ResponseCarriesWorkspaceSlug
TestPushToItem_ResponseCarriesWorkspaceSlug covers the response shape: a successful push's JSON response must carry ref/workspace/pushed/message so `p
internal/server/handlers_push_test.go:43
FunctionTestPushToItem_SucceedsWhenPublishAccepted
TestPushToItem_SucceedsWhenPublishAccepted is the positive control for both tests above. Without it, a handler that refused every push would pass them
internal/server/handlers_push_publish_result_test.go:171
FunctionTestPushToItem_TargetSessionIDAtCapAccepted
TestPushToItem_TargetSessionIDAtCapAccepted is the at-boundary counterpart: a target_session_id exactly at the cap must NOT be rejected (it is still a
internal/server/handlers_push_test.go:446
FunctionTestPushToItem_TargetSessionIDOverLengthRejected
TestPushToItem_TargetSessionIDOverLengthRejected covers the 400-over-cap guard: target_session_id is opaque and unvalidated by format, but not unbound
internal/server/handlers_push_test.go:425
FunctionTestPushToItem_TargetedAtInvisibleSessionSkipsPublish
TestPushToItem_TargetedAtInvisibleSessionSkipsPublish is the worse half of BUG-2725 and the one that lost an instruction. Pre-fix, the targeted gate
internal/server/handlers_push_visibility_test.go:221
FunctionTestPushToItem_TargetedAtSessionOnAnotherInstance
TestPushToItem_TargetedAtSessionOnAnotherInstance is the defect: with a SHARED registry, replica A must publish a push aimed at a session that only re
internal/server/handlers_push_cross_instance_test.go:65
FunctionTestPushToItem_TargetedAtSessionOnAnotherInstance_MemoryRegistryDrops
TestPushToItem_TargetedAtSessionOnAnotherInstance_MemoryRegistryDrops is the NEGATIVE CONTROL, and it is what makes the test above mean something: the
internal/server/handlers_push_cross_instance_test.go:119
FunctionTestPushToItem_TargetedAtVisibleSessionPublishes
TestPushToItem_TargetedAtVisibleSessionPublishes is the control for the test above: the identical flow with a cookie-opened target must actually publi
internal/server/handlers_push_visibility_test.go:263
FunctionTestPushToItem_TargetedSessionOfAnotherUserTreatedAsVanished
TestPushToItem_TargetedSessionOfAnotherUserTreatedAsVanished pins the self-addressed boundary (TASK-2588): a target_session_id that names a REAL, curr
internal/server/handlers_push_test.go:358
FunctionTestPushToItem_TargetedSessionReceivesOnly
TestPushToItem_TargetedSessionReceivesOnly is PLAN-2558 S5's (TASK-2588) core positive case: with TWO of the caller's own sessions connected, a push n
internal/server/handlers_push_test.go:199
FunctionTestPushToItem_TargetedUnarmedSessionTreatedAsMiss
TestPushToItem_TargetedUnarmedSessionTreatedAsMiss covers PLAN-2613 S1's honesty requirement for a REAL, currently-connected target that simply never
internal/server/handlers_push_test.go:307
FunctionTestPushToItem_TargetedVanishedSessionMisses
TestPushToItem_TargetedVanishedSessionMisses is S5's honest-miss case: a target_session_id that names no live session (mistyped, or expired) gets a 20
internal/server/handlers_push_test.go:253
FunctionTestPushToItem_TargetedWithUnreadablePresenceIsRefused
TestPushToItem_TargetedWithUnreadablePresenceIsRefused: the handler cannot tell whether the named session exists, so it must not guess in either direc
internal/server/handlers_push_presence_unreadable_test.go:44
FunctionTestPushToItem_UnavailableWithoutBus
TestPushToItem_UnavailableWithoutBus covers the 503-on-nil-bus guard: unlike a durable watch, a push has no persistence to fall back on, so a missing
internal/server/handlers_push_test.go:130
FunctionTestPushToItem_UnconfirmedPublishIsNotReportedAsSent
TestPushToItem_UnconfirmedPublishIsNotReportedAsSent: any NON-closed publish error means the notification may or may not have gone out (a go-redis ret
internal/server/handlers_push_publish_result_test.go:121
FunctionTestQuarterlyShorthand
(t *testing.T)
internal/schedule/shorthand_test.go:75
FunctionTestQuestionValidation
--------------------------------------------------------------------------- Question validation ------------------------------------------------------
internal/decision/decision_test.go:679
FunctionTestQuoteExcerptLeadingWhitespace
(t *testing.T)
internal/store/conversations_test.go:450
FunctionTestRBAC_EditorAllowedItemMutations
(t *testing.T)
internal/server/handlers_rbac_test.go:158
FunctionTestRBAC_EditorBlockedFromAgentRoleMutations
(t *testing.T)
internal/server/handlers_rbac_test.go:341
FunctionTestRBAC_EditorBlockedFromOwnerOperations
(t *testing.T)
internal/server/handlers_rbac_test.go:190
FunctionTestRBAC_OwnerAllowedEverything
(t *testing.T)
internal/server/handlers_rbac_test.go:220
FunctionTestRBAC_SearchScopedToUserWorkspaces
(t *testing.T)
internal/server/handlers_rbac_test.go:387
FunctionTestRBAC_ViewerBlockedFromCommentCreation
(t *testing.T)
internal/server/handlers_rbac_test.go:290
FunctionTestRBAC_ViewerBlockedFromDocumentMutations
(t *testing.T)
internal/server/handlers_rbac_test.go:251
FunctionTestRBAC_ViewerBlockedFromItemLinkMutations
(t *testing.T)
internal/server/handlers_rbac_test.go:353
FunctionTestRBAC_ViewerBlockedFromItemMutations
(t *testing.T)
internal/server/handlers_rbac_test.go:116
FunctionTestRBAC_ViewerCanReadEverything
(t *testing.T)
internal/server/handlers_rbac_test.go:313
FunctionTestRDateAllDayMismatchRefused
(t *testing.T)
internal/schedule/nextbetween_test.go:321
FunctionTestRDateExDateWithCount
(t *testing.T)
internal/schedule/conformance_test.go:263
FunctionTestRESTCallsDoNotSpendTheCollabDialBucket
And the reverse: a user who has spent their API burst can still dial.
internal/server/handlers_collab_bug1308_test.go:97
FunctionTestRFC5545ErratumHourlyUntil
TestRFC5545ErratumHourlyUntil is the expected-different row: the example as RFC 5545 prints it (erratum 3883) must give 09:00 and 12:00 EDT, and the c
internal/schedule/conformance_test.go:210
FunctionTestRFC5545RowCount
(t *testing.T)
internal/schedule/conformance_test.go:222
← previousnext →22,901–23,000 of 28,840, ranked by callers