The following tests fail on 11.4 under macOS as of git sha 058510a62ff9056223685d2bc544c6bff13d226f. They are organized by the kind of failure manifested in the tests, and include a very brief description of the test failure.

Case Sensitivity Failures
compat/oracle.sp-inout - test output diff case sensitivity different (expected "PKG2.func" but observed "pkg2.func")
atomic.drop_db_long_names - test diff shows different text case used than what was expected "tABBBB..." expected vs "tabbb.." observed
parts.partition_mgm_lc2_innodb - show create table output includes COLLATE=latin1_swedish_ci when the result file does not. This is because of case insensitive filesystem on macos, see https://mariadb.slack.com/archives/C03ER9QC3/p1709672581130439
parts.partition_mgm_lc2_memory - show create table output includes COLLATE=latin1_swedish_ci when the result file does not. This is because of case insensitive filesystem on macos, see https://mariadb.slack.com/archives/C03ER9QC3/p1709672581130439
parts.partition_mgm_lc2_myisam - show create table output includes COLLATE=latin1_swedish_ci when the result file does not. This is because of case insensitive filesystem on macos, see https://mariadb.slack.com/archives/C03ER9QC3/p1709672581130439

Crashes
unit.my_json_writer - signal 11 segfault https://github.com/MariaDB/server/pull/3681
unit.dynstring - signal 11 segfault https://github.com/MariaDB/server/pull/3681
unit.json_normalize - signal 11 segfault https://github.com/MariaDB/server/pull/3206
parts.alter_table - crashes in MDL_key::mdl_key_init(MDL_key::enum_mdl_namespace, char const*, char const*), assertion at line 438 fails.

Lost Allocated Memory
unit.innodb_sync - 2072 bytes allocated lost https://github.com/MariaDB/server/pull/3683
unit.pfs_noop - 2072 bytes allocated lost https://github.com/MariaDB/server/pull/3294
unit.pfs_instr-oom - 64 bytes allocated lost CONC-750

Unstable
main.connect client bug, CONC-703 , workaround https://github.com/MariaDB/server/commit/f2eda615798d6d879a913a44c22d5dd0fc7ed8c9 -
main.connect2 client bug, CONC-703 , workaround above

Superficial Diffs
plugins.multiauth - expected "client_ed25519 could not be loaded: no such file" but on macos it gives more detailed error info, which fails the test CONC-749
mroonga/storage.index_desc - expected "...Unknown error 1005..." but observed on macos "...Unknown error: 1005..." https://github.com/MariaDB/server/pull/3679
innodb_gis.rtree_debug - expected "...Unknown error 1000..." but on macos observed "...Unknown error: 1000..." https://github.com/MariaDB/server/pull/3679

Hangs
rpl.rpl_semi_sync_shutdown_await_ack - hangs
spider/bugfix.mdev_27240 - hangs

Nonexistent Table Errors
parts.debug_innodb_fail - ER_TABLE_EXISTS_ERROR
parts.debug_innodb_crash - ER_TABLE_EXISTS_ERROR
parts.partition_special_innodb - ER_TABLE_EXISTS_ERROR

Other
rpl.rpl_perfschema_applier_status_by_worker - worker_idle_time is zero when it should be >= 1
innodb.undo_space_dblwr - diff says that we have unknown storage engine innodb (but we're running in mtr...) now working both in 10.5 2ab10fbec2dc086, and in main f5821aaf7731e1f74b
period.i_s_notembedded - select query returns a table name t when the test doesn't expect
main.repair_symlink-5543 - repair error diff