MCPcopy Create free account

hub / github.com/PerpetualSoftware/pad / functions

Functions28,840 in github.com/PerpetualSoftware/pad

FunctionTestBUG3839_ThrottledEditOntoInvalidUTF8RebasesToTheFullBody
rebaseNewestVersionPatch (items.go): a throttled edit rebases the newest version's patch onto the new body, which is where the invalid text lands.
internal/store/bug3839_invalid_utf8_version_test.go:148
FunctionTestBUG3839_ThrottledEditOverARowGoDiffCannotApplyLeavesTheRow
rebaseNewestVersionPatch (items.go) applies the newest version's patch to the stored body before a throttled edit replaces it. When go-diff cannot app
internal/store/bug3839_invalid_utf8_version_test.go:426
FunctionTestBUG3839_UncheckedPatchCensusFindsEveryForm
The census's own instrument: each way a file can reach the unchecked builders is found, and the checked door and a same-named local are not.
internal/diff/bug3839_invalid_utf8_test.go:163
FunctionTestBUG3861_AlreadyMemberRefusalPrecedesRoleCheck
The last refusal the handler makes before the role check: inviting an existing member answers 409 conflict whatever the role says.
internal/server/bug3861_invite_role_test.go:97
FunctionTestBUG3861_InviteRoleMustBeKnown
BUG-3861: the invite door passed any role straight to CreateInvitation, so an unknown one reached the invitations CHECK constraint and answered 500. I
internal/server/bug3861_invite_role_test.go:16
FunctionTestBUG3861_SetupWindowRefusalPrecedesRoleCheck
The role check runs after every refusal the invite handler makes, so a bad role changes none of their answers. In the setup window (no users, no auth)
internal/server/bug3861_invite_role_test.go:81
FunctionTestBUG3864_A403ReadsTheSameOnBothTransports
BUG-3864, the classifier half: the same 403 body reaches the same envelope on both transports. Remote reads the body; stdio reads the marker the CLI r
internal/mcp/bug3864_permission_reasons_test.go:19
FunctionTestBUG3864_CrossWorkspaceRefusalsNameTheirGate
(t *testing.T)
internal/server/bug3864_forbidden_reason_test.go:58
FunctionTestBUG3864_InlineRoleGatesSayRoleOrNoAccess
The workspace-role gates that check the role inline rather than through requireMinRole (the bulk endpoint is one such gate). Driven through the router
internal/server/bug3864_forbidden_reason_test.go:106
FunctionTestBUG3864_RefusalsNameTheirCauseOnBothTransports
(t *testing.T)
cmd/pad/bug3864_permission_reasons_test.go:56
FunctionTestBUG3864_RequireMinRoleSaysRoleOnlyWhenARoleCanFixIt
A viewer can be given a higher role. A guest holds grants, not a role, and a caller with no workspace role (a platform admin who is not a member) has
internal/server/bug3864_forbidden_reason_test.go:80
FunctionTestBUG3864_TheConsentAllowListNamesItself
(t *testing.T)
internal/server/bug3864_forbidden_reason_test.go:40
FunctionTestBUG3864_TheRootWriterLeavesOtherRefusalsAlone
Other 403 codes keep their own writers. A permission_denied marker from the root would come after theirs and, last marker winning, replace it.
internal/mcp/bug3864_permission_reasons_test.go:78
FunctionTestBUG3864_UnverifiedEmailIsItsOwnCodeOnBothTransports
The Cloud-only cause: an account whose email is unverified may read and may not write. The remote transport refuses in its own gate before the request
cmd/pad/bug3864_permission_reasons_test.go:116
FunctionTestBUG3870_AMemberAcceptingADeadInvitationIsAlreadyIn
BUG-3870: an invitee who is ALREADY a member, accepting an invitation whose inviter can no longer grant it, gets BUG-3281's idempotent success (marked
internal/store/bug3870_member_dead_invitation_test.go:51
FunctionTestBUG3870_AMemberWithADeadInvitationIsAlreadyIn
(t *testing.T)
internal/server/bug3870_member_dead_invitation_test.go:33
FunctionTestBUG3870_ANonMemberInviteeStillSeesTheDeadInvitation
Control: signed in as the invitee who is NOT a member, the preview still reports the dead invitation, so the true above is the membership and not the
internal/server/bug3870_member_dead_invitation_test.go:78
FunctionTestBUG3870_MemberAcceptAndInviteeRemoval_Postgres
BUG-3870 on Postgres: the invitee's membership is read share-locked after the inviter check refuses, so a removal of the invitee committing alongside
internal/store/bug3870_member_dead_invitation_test.go:128
FunctionTestBUG3879_ToolOutcomeReachesTheAuditRow
(t *testing.T)
cmd/pad/bug3879_mcp_audit_tool_outcome_test.go:19
FunctionTestBUG3879_UnverifiedEmailWriteAuditsAsDenied
BUG-3879: since BUG-3864 an unverified email's refused write is its own code, email_not_verified, and it is a refusal like the other gates, so the aud
cmd/pad/bug3879_mcp_audit_tool_outcome_test.go:167
FunctionTestBUG3882_DisableFailsWhenTheNewSessionCannotBeSaved
BUG-3882: the disable revokes this CLI's session before the save runs, so a failed save leaves the CLI signed out and must fail the command rather tha
cmd/pad/cmd_auth_2fa_test.go:296
FunctionTestBUG3882_DisableFailsWhenTheServerSendsNoNewSession
BUG-3882: a disable answered 200 with no replacement token has still revoked this CLI's session, so it fails rather than exiting 0 signed out.
cmd/pad/cmd_auth_2fa_test.go:416
FunctionTestBUG3976_FullSettingsWriteRepairsTheRow
Control: a full settings write repairs a corrupt row, and the typed member then works. The refusal leaves a way out.
internal/server/bug3976_settings_not_object_test.go:138
FunctionTestBUG3976_IncomingNonObjectSettingsWithTypedMember
Doors 2 and 3: incoming settings that are not an object never reach the merge. The request decoder refuses them (IDEA-1488); these rows pin the measur
internal/server/bug3976_settings_not_object_test.go:112
FunctionTestBUG3976_StoreWrapsErrSettingsNotObject
The sentinel at the store, for both members and both store doors (create and update), so the handlers' errors.Is has something to match.
internal/server/bug3976_settings_not_object_test.go:159
FunctionTestBUG3976_TypedMemberOverNonObjectStoredSettings
Door 1, the defect: a flags-only PATCH over corrupt stored settings.
internal/server/bug3976_settings_not_object_test.go:81
FunctionTestBUG3978_CLITokenKeepsItsWindowThroughRotation
A browser-approved CLI token (30-day slide) keeps its 30 days through a credential change.
internal/server/bug3978_rotation_ttl_test.go:69
FunctionTestBUG3978_LegacyRowRotatesToTheWebTTL
A legacy session with no sliding window (renew_ttl_seconds = 0) rotates to the web TTL, as every rotation did before.
internal/server/bug3978_rotation_ttl_test.go:140
FunctionTestBUG3978_LongWebSessionKeepsItsCookies
A web session with a 30-day window (a browser bootstrap from /setup mints one with cliSessionTTL) keeps 30 days through a rotation, and so do its sess
internal/server/bug3978_rotation_ttl_test.go:160
FunctionTestBUG3978_RotationDoesNotLengthenAWindow
A CLI session with a 7-day window (the shape `pad auth login -i` mints) keeps 7 days: a rotation changes no session's window, so it never lengthens on
internal/server/bug3978_rotation_ttl_test.go:92
FunctionTestBUG3978_WebSessionRotationUnchanged
A web session keeps its 7 days, and its cookie says so.
internal/server/bug3978_rotation_ttl_test.go:110
FunctionTestBUG3982_AccountDoorsStayOpen
What a pinned token must keep: who it is, its owner's token list, and revoking itself (the incident path).
internal/server/bug3982_pinned_token_test.go:358
FunctionTestBUG3982_CollabComparesThePinByID
Collab compares the pin by resolved ID behind the allow-list: with no allow-list on the request, only the ID compare refuses an item in B to a pin for
internal/server/bug3982_pinned_token_test.go:754
FunctionTestBUG3982_CrossWorkspaceComparesThePinByID
Cross-workspace authorization compares the pin by resolved ID behind its by-name check: a pin whose slug names B while its ID is A's passes the by-nam
internal/server/bug3982_pinned_token_test.go:739
FunctionTestBUG3982_CrossWorkspaceCopyRefused
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:211
FunctionTestBUG3982_DeletedPinFailsClosed
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:441
FunctionTestBUG3982_EveryGlobalRouteRefusesAPinnedPAT
Every route the router registers, public ones included, either sits inside the pin-confined workspace subrouter, is in pinnedAllowedGlobalRoutes, or r
internal/server/bug3982_pinned_token_test.go:288
FunctionTestBUG3982_InvitationPreviewTreatsAForeignPinAsSignedOut
The invitation preview tells an invitee who is already a member that the accept will work (BUG-3870). A token pinned to another workspace counts as si
internal/server/bug3982_pinned_token_test.go:638
FunctionTestBUG3982_MCPGlobalRouteToolRefusedForPin
A tool that maps to a global route (pad_workspace create is POST api/v1/workspaces) meets the route gate too: the dispatcher's synthesized request goe
cmd/pad/bug3982_mcp_pinned_test.go:87
FunctionTestBUG3982_MCPToolsConfinedToPin
(t *testing.T)
cmd/pad/bug3982_mcp_pinned_test.go:47
FunctionTestBUG3982_NamedGlobalRoutesRefuseAPinnedPAT
The routes a pinned token must be refused on, named independently of pinnedTokenRouteAllowed: the walk consults that function, so a route it wrongly a
internal/server/bug3982_pinned_token_test.go:331
FunctionTestBUG3982_PinnedTokenOnPostgres
The pin holds on PostgreSQL as it does on SQLite.
internal/server/bug3982_pinned_token_test.go:485
FunctionTestBUG3982_RESTReadRefusedOutsidePin
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:98
FunctionTestBUG3982_RESTWriteRefusedOutsidePin
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:109
FunctionTestBUG3982_RefResolver404sOtherWorkspace
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:191
FunctionTestBUG3982_RefResolverComparesThePinByID
The /-/r/ resolver compares the pin by resolved ID behind the by-name check, shaped like TestBUG3982_SSEComparesThePinByID: a pin whose slug names B w
internal/server/bug3982_pinned_token_test.go:700
FunctionTestBUG3982_SSEAndCollabRefuseOtherWorkspace
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:179
FunctionTestBUG3982_SSEComparesThePinByID
SSE compares the pin by resolved ID behind the by-name check, so the stream does not rest on slugs being unique either. Simulated as in TestBUG3982_Wo
internal/server/bug3982_pinned_token_test.go:671
FunctionTestBUG3982_SearchFanOutConfined
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:128
FunctionTestBUG3982_SearchRefusalIsNotAnExistenceOracle
A refused search names no more than the token already knew: a workspace that exists and one that does not get the same bytes.
internal/server/bug3982_pinned_token_test.go:151
FunctionTestBUG3982_ShareLinkTreatsAForeignPinAsSignedOut
An email-restricted share link reads the caller; a token pinned to another workspace counts as signed out there, so its user's address unlocks nothing
internal/server/bug3982_pinned_token_test.go:597
FunctionTestBUG3982_SlugDoorsRefuseIdenticallyBeforeResolution
On every slug-addressed door the pin is checked before the name is resolved (BUG-3982 r4): a workspace that does not exist, one the user is not in, an
internal/server/bug3982_pinned_token_test.go:523
FunctionTestBUG3982_TokenDoorsNarrowToThePin
The token doors a pinned token keeps are narrowed to the tokens pinned to its own workspace: it lists only those, and revoking any other id is the 404
internal/server/bug3982_pinned_token_test.go:381
FunctionTestBUG3982_UnpinnedCredentialsUnchanged
Tokens that carry no pin keep today's reach.
internal/server/bug3982_pinned_token_test.go:467
FunctionTestBUG3982_WorkspaceAccessComparesThePinByID
The pin is compared by workspace ID as well as by slug, so the check does not rest on slugs being unique. Simulated by a pin whose slug names B while
internal/server/bug3982_pinned_token_test.go:496
FunctionTestBUG3982_WorkspaceListAndExportConfined
(t *testing.T)
internal/server/bug3982_pinned_token_test.go:162
FunctionTestBUG3985_WhitespaceSettingsOnCreateActAsEmpty
Create keeps its repair: whitespace-only settings land as an object.
internal/server/bug3985_whitespace_settings_test.go:53
FunctionTestBUG3985_WhitespaceSettingsOnPatchActAsEmpty
The "" sentinel on PATCH replaces the settings with {} and carries the typed members (PLAN-3535, IDEA-3906). Whitespace-only settings must do the same
internal/server/bug3985_whitespace_settings_test.go:31
FunctionTestBUG3999_ChatSendIsNotWrapped
Chat's key is the caller's own: a failed send is not wrapped.
internal/cli/bug3999_opkey_error_test.go:44
FunctionTestBUG3999_FailedCopyCarriesItsKey
BUG-3999: a failed write that sent an op_key carries it in an OpKeyError, without changing what errors.As / the retry classifiers see.
internal/cli/bug3999_opkey_error_test.go:12
FunctionTestBUG3999_FailedWritesPrintTheirKey
BUG-3999: any failed write that sent an op_key names it on stderr, so a retry can reuse it. A success prints nothing.
cmd/pad/bug3999_opkey_hint_test.go:17
FunctionTestBUG3999_FormatJSONFailedWriteKeepsStdoutParseable
BUG-3999: under --format json the key line goes to stderr only, so stdout stays empty or valid JSON.
cmd/pad/bug3999_opkey_hint_test.go:170
FunctionTestBUG3999_KeyedCreatePlanLimitKeepsItsStructuredMarker
BUG-3999: a keyed create comes back wrapped in *OpKeyError. The plan-limit branch asserted *cli.APIError directly and stopped matching, which lost the
cmd/pad/bug3999_opkey_hint_test.go:135
FunctionTestBUG3999_KeylessFailedWritePrintsNoHint
A failed write that sent NO op_key has no key to reuse, so it prints no hint (BUG-3999). The body is what the client marshalled; WrapOpKey is the tran
cmd/pad/bug3999_opkey_hint_test.go:111
FunctionTestBUG3999_NoKeyNoWrap
(t *testing.T)
internal/cli/bug3999_opkey_error_test.go:30
FunctionTestBUG3999_SkillNoLongerPromisesAutomaticRetrySafety
Both copies of the skill: the one embedded in the binary (pad agent guide, pad agent install) and the separately maintained one the Claude Code plugin
cmd/pad/bug3999_opkey_hint_test.go:88
FunctionTestBUG3999_SuccessAndKeylessPrintNothing
(t *testing.T)
cmd/pad/bug3999_opkey_hint_test.go:73
FunctionTestBUG4005_CreateOverHTTPParentlessOnTopChildAppends
BUG-4005: the HTTP create tells the store whether the new item is a child (the parent link is written after the insert), so a parentless item lands on
internal/server/bug4005_child_order_test.go:13
FunctionTestBUG4005_SharedCollectionBreaksSameSecondTiesByItemNumber
BUG-4005 r2: a shared collection breaks a same-second tie the way the board does, by item number, not by the random id. The two items tie on sort_orde
internal/server/bug4005_share_order_test.go:14
FunctionTestBUG4019_DisableAndDeleteDoorsRefuseOnAnUndecryptableSecret
T4: the no-password disable door and the account-delete door refuse a TOTP code on an undecryptable secret; 2FA stays on and the account stays.
internal/server/bug4019_totp_lazy_decrypt_test.go:210
FunctionTestBUG4019_EmptyStoredSecretRefusesACodeComputedFromEmpty
T5: an enabled row whose stored secret is "" refuses a code computed from "", and logs one warning naming the user (BUG-4019 r2).
internal/server/bug4019_totp_lazy_decrypt_test.go:250
FunctionTestBUG4019_EnableOnAnUndecryptableStoredSecretRefuses
T3: enabling 2FA over an undecryptable stored secret is a 409 and 2FA stays off.
internal/server/bug4019_totp_lazy_decrypt_test.go:183
FunctionTestBUG4019_LoginCodeOnAnUndecryptableSecretFailsClosedAndLogsOnce
T2: a TOTP code on an undecryptable secret mints no session, answers 500, logs exactly one record naming the user, and the log carries no secret.
internal/server/bug4019_totp_lazy_decrypt_test.go:142
FunctionTestBUG4019_ReadsOfAnUndecryptableUserSucceed
(t *testing.T)
internal/store/bug4019_totp_lazy_decrypt_test.go:45
FunctionTestBUG4019_RecoveryCodeRecoversAnUndecryptableSecret
T6: a recovery code still signs in over an undecryptable secret and is consumed; the user can then replace the secret and sign in with its codes.
internal/server/bug4019_totp_lazy_decrypt_test.go:341
FunctionTestBUG4019_SealedEmptySecretRefusesACodeComputedFromEmpty
T5b: an enabled row whose stored value is non-empty but DECRYPTS to "" (a hand-sealed empty plaintext; encrypt never emits one) also refuses a code co
internal/server/bug4019_totp_lazy_decrypt_test.go:296
FunctionTestBUG4019_TOTPSecretForVerify
(t *testing.T)
internal/store/bug4019_totp_lazy_decrypt_test.go:134
FunctionTestBUG4024_AutoPageNeverPrintsTheCappedGuessOnAnOlderServer
An older server sends no paging state, and the unpaged notice falls back to "len == limit" as its guess. On the auto-page path there is no limit, so a
cmd/pad/bug4024_item_list_json_autopage_test.go:244
FunctionTestBUG4024_CallerGitHubStatusScansEveryPage
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:404
FunctionTestBUG4024_CallerReconcileReadsEveryPage
--- the existing ListAll* callers: normal multi-page input is unchanged --- (reconcile and github status are also covered by bug3793_reconcile_pages_t
cmd/pad/bug4024_item_list_json_autopage_test.go:392
FunctionTestBUG4024_CallerStandupReadsEveryInProgressPage
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:423
FunctionTestBUG4024_CollectionAliasFallbackStillPagesToCompletion
The alias fallback still works on the auto-page path: an older server that does not resolve `task` 404s it, and the retry as `tasks` pages.
cmd/pad/bug4024_item_list_json_autopage_test.go:366
FunctionTestBUG4024_CollectionListPagesAndKeepsTheUndeclaredWarning
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:349
FunctionTestBUG4024_ExplicitLimitStaysCappedWithTheNotice
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:187
FunctionTestBUG4024_JSONListAcrossTheDefaultPageBoundary
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:175
FunctionTestBUG4024_JSONListPagesToCompletionWithoutALimit
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:152
FunctionTestBUG4024_LimitZeroMeansTheDefaultCapNotUnlimited
What --limit 0 means today (BUG-4024), pinned: it is "the default", so the CLI clamps it to defaultItemListLimit (200) and sends limit=200, one page,
cmd/pad/bug4024_item_list_json_autopage_test.go:208
FunctionTestBUG4024_NonAdvancingOffsetIsAnErrorNotAShortArray
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:275
FunctionTestBUG4024_OffsetWithoutALimitPagesFromThatOffsetToTheEnd
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:259
FunctionTestBUG4024_OlderServerWithoutPagingHeadersIsOneRequest
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:226
FunctionTestBUG4024_PageCapIsANonZeroExitNamingTheCapNeverAPartialArray
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:299
FunctionTestBUG4024_PagerAcceptsAListOfExactlyTheCap
(t *testing.T)
internal/cli/bug4024_list_all_pager_test.go:175
FunctionTestBUG4024_PagerErrorsAtThePageCapNamingIt
(t *testing.T)
internal/cli/bug4024_list_all_pager_test.go:162
FunctionTestBUG4024_PagerErrorsOnANonAdvancingOffset
(t *testing.T)
internal/cli/bug4024_list_all_pager_test.go:107
FunctionTestBUG4024_PagerKeepsTheFirstPagesUndeclaredKeys
(t *testing.T)
internal/cli/bug4024_list_all_pager_test.go:88
FunctionTestBUG4024_PagerRefusesAMalformedHasMorePage
A has-more page must be full and contiguous; every other shape is an error with no items, even when a later page says has-more=false.
internal/cli/bug4024_list_all_pager_test.go:124
FunctionTestBUG4024_PagerStartsAtTheGivenOffset
(t *testing.T)
internal/cli/bug4024_list_all_pager_test.go:47
FunctionTestBUG4024_PagerTreatsANegativeStartOffsetAsZero
A negative start offset (`--offset -5`) is sent as no offset, so the server reads from 0, as it does for any bad offset. The pager must count from 0 t
internal/cli/bug4024_list_all_pager_test.go:73
FunctionTestBUG4024_RealServerJSONListReturnsEveryRowOnBothBackends
--- a real server over the real store, on both backends --- The fake server above proves the CLI follows the headers it is told about. This proves the
cmd/pad/bug4024_item_list_json_autopage_test.go:461
FunctionTestBUG4024_SkippingNextOffsetIsAnErrorNotAShortArray
(t *testing.T)
cmd/pad/bug4024_item_list_json_autopage_test.go:287
← previousnext →17,301–17,400 of 28,840, ranked by callers