Issues

Verify: BLPOP's family is inline-intercepted but absent from is_inline_intercepted, and the drift guard cannot see it

#946

· TinDang97 opened

on Sep 12, 2026

SADD under-counts its reply when one batch crosses set-max-intset-entries: replies 3 after adding 22

#944

· TinDang97 opened

on Sep 12, 2026

Beat Redis on INCR/SADD/LPUSH/HSET/ZADD: only 0.626 µs/op of the dispatch-path tax transfers to writes, so handler work is the primary lever

#942

· TinDang97 opened

on Sep 11, 2026

monoio write path starts the latency timer after the command has run: every write reports 0 µs, and SLOWLOG can never fire for a write

#941

· TinDang97 opened

on Sep 11, 2026

WRONGTYPE bumps the WATCH version: a failed, non-mutating SADD/HSET/LPUSH/ZADD aborts a transaction nothing wrote (#938 regression)

#940

· TinDang97 opened

on Sep 11, 2026

TXN.* and TEMPORAL.* are inline-intercepted but absent from is_inline_intercepted — moon#507 wait-set gap, and a fixed cluster slot

#937

· TinDang97 opened

on Sep 10, 2026

ZSCAN: COUNT is a hard result limit and iteration order is by member, not stored order

#934

· TinDang97 opened

on Sep 10, 2026

ZRANGE ... BYSCORE|BYLEX REV ignores the argument-order swap and answers empty

#933

· TinDang97 opened

on Sep 10, 2026

DEBUG OBJECT reports a key outside the caller's ~pattern (redis-parity, admin-gated)

#931

· TinDang97 opened

on Sep 10, 2026

FT.* is denied outright for key-restricted users — scope it per index PREFIX instead

#930

· TinDang97 opened

on Sep 10, 2026

SADD -13.9% and SPOP -8.3% since v0.8.9, measured on Linux — the loss is diffuse, not one commit

#923

· TinDang97 opened

on Sep 10, 2026

DEBUG RECLAMATION: every client-facing handler shadows the shard's manifest and WAL writer with None, so the # WAL section always reports the no-writer fallback

#918

· TinDang97 opened

on Sep 10, 2026

Navigation Menu

Uh oh!

All issues

Verify: BLPOP's family is inline-intercepted but absent from is_inline_intercepted, and the drift guard cannot see it

SADD under-counts its reply when one batch crosses set-max-intset-entries: replies 3 after adding 22

Beat Redis on INCR/SADD/LPUSH/HSET/ZADD: only 0.626 µs/op of the dispatch-path tax transfers to writes, so handler work is the primary lever

monoio write path starts the latency timer after the command has run: every write reports 0 µs, and SLOWLOG can never fire for a write

WRONGTYPE bumps the WATCH version: a failed, non-mutating SADD/HSET/LPUSH/ZADD aborts a transaction nothing wrote (#938 regression)

TXN.* and TEMPORAL.* are inline-intercepted but absent from is_inline_intercepted — moon#507 wait-set gap, and a fixed cluster slot

ZSCAN: COUNT is a hard result limit and iteration order is by member, not stored order

ZRANGE ... BYSCORE|BYLEX REV ignores the argument-order swap and answers empty

DEBUG OBJECT reports a key outside the caller's ~pattern (redis-parity, admin-gated)

FT.* is denied outright for key-restricted users — scope it per index PREFIX instead

SADD -13.9% and SPOP -8.3% since v0.8.9, measured on Linux — the loss is diffuse, not one commit

DEBUG RECLAMATION: every client-facing handler shadows the shard's manifest and WAL writer with None, so the # WAL section always reports the no-writer fallback

Uh oh!

Search results