Tune fuzzer for better throughput

This does not affect coverage metrics and makes it ~3x faster

fuzz/fuzz_targets/fuzz_redb.rs

@@ -15,7 +15,7 @@ use redb::backends::FileBackend;
 use crate::FuzzerSavepoint::{Ephemeral, NotYetDurablePersistent, Persistent};
 
 // These slow down the fuzzer, so don't create too many
-const MAX_PERSISTENT_SAVEPOINTS: usize = 20;
+const MAX_PERSISTENT_SAVEPOINTS: usize = 10;
 // Table to count which transactions have been successfully committed so that the reference BtreeMap can be kept in sync
 const COUNTER_TABLE: TableDefinition<(), u64> = TableDefinition::new("transaction_counter");
 const TABLE_DEF: TableDefinition<u64, &[u8]> = TableDefinition::new("fuzz_table");

justfile

@@ -37,13 +37,13 @@ watch +args='test':
     cargo watch --clear --exec "{{args}}"
 
 fuzz: pre
-    cargo fuzz run --sanitizer=none fuzz_redb -- -max_len=100000
+    cargo fuzz run --sanitizer=none fuzz_redb -- -max_len=10000
 
 fuzz_cmin:
-    cargo fuzz cmin --sanitizer=none fuzz_redb -- -max_len=100000
+    cargo fuzz cmin --sanitizer=none fuzz_redb -- -max_len=10000
 
 fuzz_ci: pre
-    cargo fuzz run --sanitizer=none fuzz_redb -- -max_len=100000 -max_total_time=60
+    cargo fuzz run --sanitizer=none fuzz_redb -- -max_len=10000 -max_total_time=60
 
 fuzz_coverage: pre
     #!/usr/bin/env bash