MCPcopy Create free account

hub / github.com/PerpetualSoftware/pad / functions

Functions28,840 in github.com/PerpetualSoftware/pad

FunctionTestBUG3679_RefusedPasswordWriteKeepsResetLink
A fenced write that loses its fence changes nothing, so it must not spend the link either.
internal/store/bug3679_reset_links_on_password_change_test.go:78
FunctionTestBUG3680_AdminRegisterFromTheSPA
(t *testing.T)
internal/server/bug3680_cross_site_signin_doors_test.go:220
FunctionTestBUG3680_BootstrapRefusesCrossSiteForms
(t *testing.T)
internal/server/bug3680_cross_site_signin_doors_test.go:135
FunctionTestBUG3680_BootstrapTokenHeaderFromRemoteStillWorks
(t *testing.T)
internal/server/bug3680_cross_site_signin_doors_test.go:186
FunctionTestBUG3680_RegisterRefusesCrossSiteForms
(t *testing.T)
internal/server/bug3680_cross_site_signin_doors_test.go:105
FunctionTestBUG3680_ResetPasswordAcceptsAllowedCORSOrigin
The legitimate callers the guard must keep working beyond the same-origin and headless controls above (#2146): an allowed CORS origin, the X-Bootstrap
internal/server/bug3680_cross_site_signin_doors_test.go:162
FunctionTestBUG3680_ResetPasswordRefusesCrossSiteForms
(t *testing.T)
internal/server/bug3680_cross_site_signin_doors_test.go:66
FunctionTestBUG3686_SearchHonoursIncludeArchived
(t *testing.T)
internal/store/bug3686_list_search_test.go:48
FunctionTestBUG3686_SearchHonoursSort
(t *testing.T)
internal/store/bug3686_list_search_test.go:78
FunctionTestBUG3687_DashboardGuestNotStarvedByWorkspaceLevelRows
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:227
FunctionTestBUG3687_DashboardNotStarvedByRowsOfMissingItems
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:257
FunctionTestBUG3687_DashboardRecentActivityNotStarvedByHiddenRows
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:203
FunctionTestBUG3687_FeedBearerAdminIsScopedAndNotStarved
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:285
FunctionTestBUG3687_FeedItemGrantNarrowsItsCollection
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:165
FunctionTestBUG3687_FeedPagesCountOnlyVisibleRows
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:142
FunctionTestBUG3687_UnscopedDashboardNotStarvedByRowsOfMissingItems
(t *testing.T)
internal/server/bug3687_activity_visible_page_test.go:323
FunctionTestBUG3693_PrefixlessItemRenamedToARefKeepsItsLinks
An item with no ref of its own (its collection has no prefix) renamed to another item's ref: the cascade writes the collection-qualified title, which
internal/store/bug3693_reference_shaped_title_test.go:61
FunctionTestBUG3693_ReferenceShapedTitleWithNothingSafeToWriteIsLeft
No own ref and no slug to qualify with: the bracket is left as it is.
internal/links/bug3693_own_ref_test.go:46
FunctionTestBUG3693_ReferenceShapedTitleWritesTheOwnRef
BUG-3693: a reference-shaped new title is written as the renamed item's own reference, with the title as display text, and the writer and the projecti
internal/links/bug3693_own_ref_test.go:11
FunctionTestBUG3693_RenameToAReferenceShapedTitleKeepsItsLinks
BUG-3693: a rename whose new title is reference-shaped ("TASK-1") made the cascade write [[TASK-1]] into every linker, and the index and the renderer
internal/store/bug3693_reference_shaped_title_test.go:12
FunctionTestBUG3703_StoredFieldKeysStopsAtTheFirstAbsentKey
BUG-3703: StoredFieldKeys asked every undeclared filter key with its own scan of the list's scope, so N absent keys cost N scans, and the handler refu
internal/store/bug3703_stored_field_keys_test.go:13
FunctionTestBUG3703_TooManyFieldFiltersAreRefusedBeforeAnyProbe
BUG-3703: a list request's field filters are bounded before any store work; at the bound it is still answered.
internal/server/bug3703_list_filter_cap_test.go:12
FunctionTestBUG3704_EveryItemSortEndsOnTheID
BUG-3704: an offset page is only stable when the order is total, so every item-list sort ends on i.id. SQLite happens to return ties in rowid order, w
internal/store/bug3704_sort_tiebreak_test.go:12
FunctionTestBUG3704_ItemListsAreBoundedAndPaged
(t *testing.T)
internal/server/bug3704_list_bounds_test.go:27
FunctionTestBUG3704_ListAllItemsFollowsNextOffset
BUG-3704: a report that needs every row follows the server's next-offset until has-more is false, and a server that sends no paging state (older than
internal/cli/bug3704_list_all_items_test.go:18
FunctionTestBUG3704_PadItemListSaysWhenItIsTruncated
(t *testing.T)
internal/mcp/bug3704_item_list_truncation_test.go:32
FunctionTestBUG3704_PadItemListZeroLimitTakesTheDefault
A zero limit takes the default (it used to, through the CLI's own default; the probe row must not turn it into 1). A FRACTIONAL limit under 1 used to
internal/mcp/bug3704_item_list_truncation_test.go:76
FunctionTestBUG3704_PagingVisitsEveryRowOnce
Paging by next-offset visits every row exactly once, also on the ?search= path, which used to drop the offset, and across rows that tie on the sort.
internal/server/bug3704_list_bounds_test.go:71
FunctionTestBUG3706_TagFilterIsExact
(t *testing.T)
internal/store/bug3686_list_search_test.go:105
FunctionTestBUG3744_CursorlessSnapshotRefusedOnceACursorIsRecorded
A snapshot without a cursor cannot show what it covers: refused once the stored body has recorded one, written as before when none has.
internal/store/bug3744_snapshot_behind_content_test.go:137
FunctionTestBUG3744_DelayedOlderSnapshotIsRefused
The audit's scenario: cursors 10 and 11, the 11 flush lands, then the delayed 10 arrives.
internal/store/bug3744_snapshot_behind_content_test.go:53
FunctionTestBUG3744_OlderSnapshotRefusedWhenNewerWasBelowMax
The case the GC watermark cannot see: the newer flush is itself below MAX, so it never moves the watermark, and comparing against the watermark alone
internal/store/bug3744_snapshot_behind_content_test.go:79
FunctionTestBUG3744_SameOrNewerCursorIsWritten
Controls: the same cursor and a newer cursor are written.
internal/store/bug3744_snapshot_behind_content_test.go:118
FunctionTestBUG3744_SnapshotOlderThanAServerWriteIsRefused
A server-driven full write covers every row there is, so a snapshot from before it is refused too.
internal/store/bug3744_snapshot_behind_content_test.go:100
FunctionTestBUG3745_EveryVersionResolvesExactlyAfterThrottledWrites
The property: after any run of throttled writes, every version in the reverse chain still resolves to exactly the body it recorded.
internal/store/bug3745_throttled_version_test.go:81
FunctionTestBUG3745_FullBodyVersionCountsFollowTheThrottledWrite
A newest version stored in full needs no rebase, but its line counts still describe the change it covers, which the throttled write extended.
internal/store/bug3745_throttled_version_test.go:218
FunctionTestBUG3745_SystemRewritesKeepTheNewestVersionExact
The two system rewrites of items.content that write no version, the wiki title cascade and the attachment remap, also leave the newest reverse patch m
internal/store/bug3745_throttled_version_test.go:142
FunctionTestBUG3745_ThrottledWriteKeepsTheNewestVersionExact
BUG-3745: a throttled write (same actor and source within the window) replaces the body without a new version. The newest version is a reverse patch m
internal/store/bug3745_throttled_version_test.go:16
FunctionTestBUG3746_ContentReplacementAdmitsTheNewSeedAndRefusesTheOld
The fence must not OVER-fence (Heron, GO-blocking on #2134): a tab seeded from the replacement itself, at the new seq, is admitted normally, or every
internal/server/bug3746_prune_replace_fence_test.go:120
FunctionTestBUG3746_ContentReplacementMovesTheSeedBoundary
BUG-3746: a REST content replacement prunes the op-log and replaces items.content, but did not stamp the seed boundary a restore stamps. A tab that fe
internal/server/bug3746_prune_replace_fence_test.go:24
FunctionTestBUG3746_PatchReplacementAdmitsTheNewSeedAndRefusesTheOld
(t *testing.T)
internal/server/bug3746_prune_replace_fence_test.go:149
FunctionTestBUG3746_PatchReplacementMovesTheSeedBoundary
The patch route's twin. It prunes and replaces through its own PruneAndApply call (handlers_items_patch.go), so it is wired separately and needs its o
internal/server/bug3746_prune_replace_fence_test.go:69
FunctionTestBUG3747_ACompactionDoesNotCoverAnEarlierGenerationsCursor
(t *testing.T)
internal/store/bug3747_compaction_generation_test.go:37
FunctionTestBUG3747_ARecompactionKeepsItsGenerationsCursors
Compacting a generation again keeps the cursors the first snapshot covered: the new snapshot replaces the old one, and the generation's start carries.
internal/store/bug3747_compaction_generation_test.go:112
FunctionTestBUG3747_ASnapshotWithoutAFloorCoversNothing
An item compacted before migration 145 has a snapshot but no floor. Nothing says which generation its cursors came from, so none is covered: those tab
internal/store/bug3747_compaction_generation_test.go:137
FunctionTestBUG3747_AnOldGenerationsCursorIsRefreshedAfterTheNextIsCompacted
BUG-3747, through Join: a tab whose cursor came from a deleted op-log generation is refreshed, even after the NEXT generation is compacted past its cu
internal/server/bug3747_compaction_generation_test.go:16
FunctionTestBUG3747_TheFloorCarriesPastALateLowerRow
A row allocated a lower id that commits after a snapshot (Postgres) sits below the snapshot at the next compaction, so the previous snapshot is not th
internal/store/bug3747_compaction_generation_test.go:160
FunctionTestBUG3749_FailedAppendInBracketIsNotPersisted
(t *testing.T)
internal/collab/bug3749_applier_ack_durable_test.go:84
FunctionTestBUG3749_FailedAppendInLegacyRoundTripIsNotPersisted
A legacy (un-bracketed) round-trip cannot attribute the failure to its frame, and still must not report a failed append as stored.
internal/collab/bug3749_applier_ack_durable_test.go:99
FunctionTestBUG3749_RestoreFinalizesAFailedBracketAsNotDurable
The freeze-finalize path: a restore that freezes the room while a bracket holds a failed append must not finalize it as not-persisted. That outcome re
internal/collab/bug3749_applier_ack_durable_test.go:127
FunctionTestBUG3749_StoredAppendAndNoOpDiffStillSucceed
Controls: a stored frame and a no-op diff (no frame at all) still succeed.
internal/collab/bug3749_applier_ack_durable_test.go:107
FunctionTestBUG3750_RenumberingKeepsStoredReferencesOnTheirItems
BUG-3750: converting legacy per-collection numbering to per-workspace numbering changed item numbers and nothing else, so a stored [[IDEA-1]] whose it
internal/store/bug3750_renumber_refs_test.go:15
FunctionTestBUG3750_RewriteRefs
(t *testing.T)
internal/links/bug3750_rewrite_refs_test.go:8
FunctionTestBUG3750_StalePrefixReferenceKeepsItsFallbackTarget
A reference whose prefix no collection has (the item moved after it was written) resolved, before the conversion, through the number fallback: to the
internal/store/bug3750_renumber_refs_test.go:105
FunctionTestBUG3751_ClassificationStopsWhereverTheDeadlineLands
Classification is bounded by the context wherever the deadline lands (BUG-3751): for every version application during which it can expire, the item is
internal/store/version_tie_diagnostic_test.go:200
FunctionTestBUG3751_DiagnosticCannotWrite
The diagnostic cannot write, by construction: on Postgres its transaction is READ ONLY at the server, and on SQLite its connection is query_only. A wr
internal/store/version_tie_diagnostic_test.go:136
FunctionTestBUG3751_Migration054OrdersSameSecondTiesByID
BUG-3751: Postgres migration 054 backfills version_seq for rows that predate it. Same-second versions are ordered by id, never by ctid, which an UPDAT
internal/store/bug3751_migration_054_test.go:16
FunctionTestBUG3751_OperatorRunRequiresADeadline
The operator entry point refuses to run without a run deadline: it is the only bound on opening the transaction (BUG-3751).
internal/store/version_tie_diagnostic_test.go:288
FunctionTestBUG3751_StatementTimeoutNeverRoundsToZero
(t *testing.T)
internal/store/version_tie_diagnostic_test.go:272
FunctionTestBUG3751_VersionTieDiagnosticCounts
BUG-3751: the read-only tie-order diagnostic counts what it should on a seeded history. Each version replaces the whole body with unrelated text, so a
internal/store/version_tie_diagnostic_test.go:19
FunctionTestBUG3752_PGMigration004MigratesAPhasePlanDocument
BUG-3752: Postgres migration 004 renamed doc_type 'phase-plan' to 'plan' with an UPDATE before dropping 001's CHECK, which rejects 'plan', so any data
internal/store/bug3752_pg_migration_004_test.go:13
FunctionTestBUG3768_RestrictedAfterStagingCannotConfirm
The confirm doors check the restriction at confirm time, not only at preview: an owner restricted after staging cannot finish.
internal/server/bug3768_restricted_owner_apps_test.go:118
FunctionTestBUG3768_RestrictedOwnerCannotGrantAnApp
(t *testing.T)
internal/server/bug3768_restricted_owner_apps_test.go:49
FunctionTestBUG3769_EveryDoorSlugifiesTheWorkspaceSlug
BUG-3769: a workspace slug is slugified whichever door supplied it. An import name override, an import bundle's own slug and a create body's slug all
internal/store/bug3769_workspace_slug_test.go:12
FunctionTestBUG3769_ImportNameOverrideGetsAUsableSlug
BUG-3769, through the HTTP import door: ?name= with a space answered 201 with that raw name as the slug, and the next request to the workspace's own e
internal/server/bug3769_import_name_slug_test.go:14
FunctionTestBUG3769_MigrateKeepsALegacyWorkspaceSlug
BUG-3769: every import door canonicalizes a workspace slug, but `pad db migrate-to-pg` copies EXISTING workspaces, whose slugs are already in URLs, bo
cmd/pad/bug3769_migrate_slug_test.go:14
FunctionTestBUG3770_AcceptWaitsOnAConcurrentInviterLoss_Postgres
BUG-3770 on Postgres: the inviter's role is read in the same share-locked select as its scope, so a demotion or a removal committing alongside an acce
internal/store/bug3770_inviter_authority_test.go:133
FunctionTestBUG3770_AnInviterStillAnOwnerIsAcceptedAtEveryDoor
Control, per door: an inviter who is still an owner is offered and accepted as before, so the refusal above is the loss and not the fixture.
internal/server/bug3770_inviter_authority_test.go:141
FunctionTestBUG3770_EveryAcceptDoorRefusesAnInviterWhoLostTheAuthority
(t *testing.T)
internal/server/bug3770_inviter_authority_test.go:56
FunctionTestBUG3770_InvitationDiesWithItsInvitersAuthority
BUG-3770: a pending invitation is accepted only while its inviter still holds the authority to invite, read at accept time. An inviter removed (there
internal/store/bug3770_inviter_authority_test.go:16
FunctionTestBUG3771_LivePeerSurvivesThePongWait
A live peer that keeps reading answers the pings and stays connected well past the pong wait: the keepalive reaps only the silent.
internal/collab/bug3771_ws_deadlines_test.go:89
FunctionTestBUG3771_SilentPeerIsReaped
BUG-3771: a peer that stops reading after it is anchored never answers a ping, so its read deadline lapses and Join returns.
internal/collab/bug3771_ws_deadlines_test.go:75
FunctionTestBUG3773WatchPayloadSlugIsMemoisedPerTick
BUG-3773: a stream resolves each workspace's slug once per reval tick, not once per notification. The memo is observed through the payload: a workspac
internal/server/bug3773_watch_slug_cache_test.go:16
FunctionTestBUG3776ForeignEndpointLinkRowsAreIgnored
BUG-3776: a link row is scoped by its own workspace_id, but the readers that enrich or count it joined its endpoint items without asking where THEY li
internal/store/bug3776_link_endpoint_workspace_test.go:23
FunctionTestBUG3789_AnOffsetPastTheEndIsAnEmptyArray
An offset past the end (#2153) is an empty page that serialises as [] (not null), with has-more false, on both bare-array lists.
internal/server/bug3789_bounded_lists_test.go:374
FunctionTestBUG3789_ChildrenAreBoundedAndPaged
(t *testing.T)
internal/server/bug3789_bounded_lists_test.go:78
FunctionTestBUG3789_ChildrenWindowIsOverVisibleRows
A restricted member's children window counts only children they may see, and has_children reflects only visible grandchildren.
internal/server/bug3789_bounded_lists_test.go:121
FunctionTestBUG3789_TiesComeBackInATotalOrder
Rows that tie on every leading sort key come back in a total order: item number for children (creation order, BUG-4005), id for stars. The links and s
internal/server/bug3789_bounded_lists_test.go:317
FunctionTestBUG3790_PadItemStarredIsBoundToTheTruncatingAction
CONVE-19: the tool's starred action IS the truncating action, not a bare pass-through to the CLI.
internal/mcp/bug3790_starred_truncation_test.go:96
FunctionTestBUG3790_RemoteStarredAsksOnePastTheCapAndSaysTruncation
Remote: the HTTP dispatcher sends ?limit=301 to /starred.
internal/mcp/bug3790_starred_truncation_test.go:61
FunctionTestBUG3790_StarredIsBoundedPagedAndLoadsBodiesForThePage
(t *testing.T)
internal/server/bug3789_bounded_lists_test.go:192
FunctionTestBUG3790_StdioStarredAsksOnePastTheCapAndSaysTruncation
Stdio: the CLI is asked for --limit 301.
internal/mcp/bug3790_starred_truncation_test.go:39
FunctionTestBUG3791_ChangesIsBounded
(t *testing.T)
internal/server/bug3791_changes_cursor_test.go:71
FunctionTestBUG3791_CursorIsValidated
(t *testing.T)
internal/server/bug3791_changes_cursor_test.go:296
FunctionTestBUG3791_WalkSeesEveryChangeIncludingMidWalkWrites
A walk visits every change exactly once, reports deletions, repeats the first page's server time, and a change committed between pages, to a row an ea
internal/server/bug3791_changes_cursor_test.go:97
FunctionTestBUG3791_WalkWindowIsOverVisibleRows
Visibility is applied before the window, not after: a restricted member's walk returns only the rows they may see, and the has-more probe does not cou
internal/server/bug3791_changes_cursor_test.go:214
FunctionTestBUG3792_SharedCollectionIsBoundedAndPaged
(t *testing.T)
internal/server/bug3789_bounded_lists_test.go:247
FunctionTestBUG3793_CompletedWorkPagesToTheCutoff
BUG-3793: the completed-work report (changelog, standup) reads newest-first pages until one reaches past the cutoff. It read one page as every match,
cmd/pad/bug3793_reconcile_pages_test.go:85
FunctionTestBUG3793_GitHubStatusScansEveryPage
BUG-3793 (CONVE-19): pad github status scans every page of each collection.
cmd/pad/bug3793_reconcile_pages_test.go:48
FunctionTestBUG3793_ListAllCollectionItemsFollowsNextOffset
BUG-3793: ListAllCollectionItems follows the collection list's next-offset to the end, dropping any caller limit, as ListAllItems does for the workspa
internal/cli/bug3793_list_all_collection_items_test.go:18
FunctionTestBUG3793_ReconcileReadsEveryPage
BUG-3793 (CONVE-19): pad reconcile reads EVERY page of the workspace list. It used to ask for one 500-row page and reconcile only that, silently.
cmd/pad/bug3793_reconcile_pages_test.go:18
FunctionTestBUG3793_RemoteItemResourceSaysWhenItIsTruncated
(t *testing.T)
internal/mcp/bug3793_item_resource_truncation_test.go:69
FunctionTestBUG3793_StdioItemResourceSaysWhenItIsTruncated
(t *testing.T)
internal/mcp/bug3793_item_resource_truncation_test.go:57
FunctionTestBUG3796_ArchivedGrantKeepsTheAppCeiling
A granted item is no longer tested against the visible-collection list (BUG-3796), so for an app request the grant list's own ceiling is the only one
internal/server/bug3796_changes_archived_grant_test.go:185
FunctionTestBUG3796_ArchivedGrantedItemReachesGranteeThroughChanges
BUG-3796: a grant-only caller (a restricted member or a guest whose only access to an item is an item grant) must receive that item's archive through
internal/server/bug3796_changes_archived_grant_test.go:22
FunctionTestBUG3797_ASignInWithoutASubjectIsRefusedForEveryProvider
BUG-3797: an OAuth sign-in without the provider account's subject is refused, whatever the account's binding state. BUG-3671 refused it only once a su
internal/server/bug3797_oauth_subject_required_test.go:15
FunctionTestBUG3802_JoinRacingAPruneAndReplaceWaitsAndIsFenced
BUG-3802 (BUG-3746 residual): a prune-and-replace write stamps the in-memory seed boundary (SetLastRestoreSeq) inside PruneAndApply's applyFn, and Joi
internal/collab/bug3802_join_races_prune_test.go:23
FunctionTestBUG3810_FailedStatFailsTheWriteAndLeavesTheFile
BUG-3810: a failed stat used to read as "not a regular file", which skips the truncate, so a shorter body left the old tail behind. It now fails the w
internal/cli/credentials_bug3810_test.go:13
FunctionTestBUG3811_ADisabledAdminDoesNotCount
The behaviour change the PR states: a DISABLED admin is not the admin who remains, so the last active one cannot be demoted, disabled or deleted while
internal/store/bug3811_last_admin_race_test.go:165
FunctionTestBUG3811_ConcurrentRemovalsKeepAnActiveAdmin
The racy pairs. The interleaving is made deterministic by holding the FIRST removal of admin A open in a transaction (the door's real transaction body
internal/store/bug3811_last_admin_race_test.go:61
← previousnext →17,101–17,200 of 28,840, ranked by callers