gh-154594: Fix deepcopy memo lookup when __deepcopy__ returns None

[tonghuaroot]

PR #138429 (gh-132657) replaced the memo-miss sentinel from _nil = [] to None for free-threading performance. But None is a valid deepcopy result, so memo.get(d, None) can't distinguish "not found" from "found None". This skips memoization for None-valued results, causing redundant __deepcopy__ calls.

Fix: use a private _MEMO_MISS = object() sentinel. Immortalized objects have no refcount contention, so the FT benefit is preserved.

• Issue: copy.deepcopy memo lookup fails when __deepcopy__ returns None #154594

[eendebakpt]

Two small comments, but I am leaning more towards the approach suggested in the corresponding issue.

[tonghuaroot]

Switched to if d in memo: return memo[d] as suggested. Drops the sentinel, and the test no longer touches memo.

Microbenchmark (5M lookups): d in memo is ~6 ns faster on a hit and ~22 ns on a miss than memo.get(d, None); full deepcopy is within noise.

[pochmann]

d in memo is ~6 ns faster on a hit and ~22 ns on a miss than memo.get(d, None)

Faster in both cases? Did you measure purely d in memo and memo.get(d, None)? Or did you include the other differences as well?

[tonghuaroot]

The earlier numbers included the surrounding statements. Measuring just the lookup, each returning the value (best of 5 runs of 10M):

                                  hit (present)   miss (absent)
memo.get(d, None)                    51 ns           60 ns
memo[d] if d in memo else None       55 ns           49 ns

So d in memo is about 4 ns slower on a hit and 11 ns faster on a miss. In deepcopy each object is looked up once, so misses are the common case, and full deepcopy timings are within noise either way.

[pochmann]

Ok if the surrounding stuff was properly included, then that's good. This new comparison for "just the lookup" is not, it's missing the is not None check in an if-context and the store/load of y.

[pochmann]

Anyway, I agree that misses are probably the common case, and I'm not surprised if full deepcopy timings are within noise either way, as deepcopy is rather slow.

[eendebakpt]

Two small nits, but this looks good to me.