fix goleveldb via dex 0.5.5 update

[chappjc]

With goleveldb at v1.0.1-0.20220614013038-64ee5596c38a:

2022-10-31 17:10:30.794 [INF] SKDB: Successfully loaded 707739 ticket pool diffs
2022-10-31 17:10:31.243 [ERR] SKDB: Error reading from database (missing key 707739 for new tickets).  Reinitializing.
2022-10-31 17:10:31.243 [DBG] SKDB: Initialized new stake db.
2022-10-31 17:10:31.251 [DBG] SKDB: badger: Got compaction priority: {level:0 score:1.73 dropPrefixes:[]}
2022-10-31 17:10:31.253 [ERR] DATD: Unable to create stake DB: bucket already exists
2022-10-31 17:10:31.253 [INF] DATD: Closing connection to dcrd.
2022-10-31 17:10:31.253 [INF] DATD: Bye!
2022-10-31 17:10:31.504 [ERR] DATD: StakeDatabase recovery failed: bucket already exists

This PR resolves the issue by updating to dcrdex v0.5.5, which has the goleveldb require reverted to 2ae1ddf74ef7

[chappjc]

This little gremlin took me a very long time to smoke out.

It was breaking dcrd's database/v3/ffldb code at stake.LoadBestNode > ticketdb.DbFetchNewTickets, at least.

And it was forcing this revision because of dcrdex -> go-ethereum 1.10.20, which had the bad revision required, thus hoisting the revision used by dcrd.

[chappjc]

Work as intended on affected host