Race condition ni bug ambapo matokeo yanategemea muda usiotabirika wa operations zinazoendana zikifikia shared state. Kesi ya kawaida ni read-modify-write ambayo siyo atomic.
counter += 1 inaonekana atomic lakini kwa kweli ni hatua tatu: read, add, write. Threads mbili zinaweza kuingiliana hivyo update inapotea.
counter = 0
Thread A: read 0 ─────────── write 1
Thread B: read 0 ─ write 1
Result: 1 (should be 2 — one increment lost)
counter = 0
def inc():
global counter
for _ in range(100_000):
counter += 1 # NOT atomic: read, +1, write
# Run two threads, expect 200_000...
# ...but you often get less — a lost update
Suluhisho ni kufanya sehemu hiyo kuwa atomic kwa lock:
lock = threading.Lock()
def inc():
global counter
for _ in range(100_000):
with lock: # only one thread in here at a time
counter += 1
Dalili ndiyo inayofanya races kuwa hatari: matokeo yasiyo sahihi ya mara kwa mara, yasiyoweza kuzalishwa upya, yanayopita katika tests na kushindwa chini ya mzigo. Zinajificha hadi muda unapobadilika — mashine ya haraka zaidi, traffic zaidi, core mpya. Kutambua shared mutable state kama chanzo cha msingi ni hatua ya kwanza; wahoji wanataka kukuona ukigundua read-modify-write isiyolindwa.
Maktaba ya maswali ya mahojiano ya IT yenye majibu ya kina — kutoka Junior hadi Senior.
Changia