chore: use heed for LMDB in benchmark

[shekhirin]

just bench failed on my machine (M1 Pro, 32GB RAM)

➜  redb git:(master) ✗ just bench
...
lmdb-rkv: Bulk loaded 1000000 items in 1395ms
lmdb-rkv: Wrote 100 individual items in 11ms
lmdb-rkv: Wrote 100 x 1000 items in 3364ms
lmdb-rkv: len() in 0ms
lmdb-rkv: Random read 1000000 items in 1063ms
lmdb-rkv: Random read 1000000 items in 1045ms
error: bench failed, to rerun pass `--bench lmdb_benchmark`

Caused by:
  process didn't exit successfully: `/Users/shekhirin/Projects/redb/target/release/deps/lmdb_benchmark-9bace23009390253 --bench` (signal: 11, SIGSEGV: invalid memory reference)
error: Recipe `bench` failed on line 34 with exit code 101

So I decided to try out heed as an alternative LMDB wrapper library. It's more maintained and subjectively has higher code quality. Also, it works 😄

➜  redb git:(lmdb-heed) ✗ just bench
...
heed: Bulk loaded 1000000 items in 1501ms
heed: Wrote 100 individual items in 7ms
heed: Wrote 100 x 1000 items in 2988ms
heed: len() in 0ms
heed: Random read 1000000 items in 1038ms
heed: Random read 1000000 items in 1008ms
heed: Random range read 10000000 elements in 1544ms
heed: Random range read 10000000 elements in 1548ms
heed: Random read (4 threads) 1000000 items in 263ms
heed: Random read (8 threads) 1000000 items in 144ms
heed: Random read (16 threads) 1000000 items in 145ms
heed: Random read (32 threads) 1000000 items in 122ms
heed: Removed 500000 items in 1460ms
...

[cberner]

Merged. Thanks!