done (split in multiple diffs)

comment nit

fix a warning

improve behavior when cancelling a reindexing: remove alarming log, wipe the index and don't set the metadata flag to true when shutdown is requested during the process

as_member_ref -> as_ref, factor repeated code into a member_ser_from_member method

add comments about the double switch Some(cf_member_hash) and mem_data.is_member_hash_index_enabled
mention in the summary that the mempool indexing is not part of this diff

I still need to figure out the mempool thing and maybe rust tests

rebase, feedback

feedback:

• allow shutdown during scripthash reindexing
• rename scripthash -> member_hash in the group_history.rs context
• let the hash member method deal with endianness to minimize risks of getting in wrong in callsite (hash_member() -> Sha256 -> ser_hash_mebmer -> [u8; 32])
• &[0xff; 32]

It is always possible to add more libs later if needed

Looks good. I'm not really on top of chronik plugins yet so not sure about the needed packages in Setup.local.

the bytes dependency for chronik-http is no longer needed

add more tests, more endpoints (unconfirmed_txs, history)

fix style warnings, better handle invalid scripthash payloads

still a WIP

fix stuff

use an enum, move code to query/group_history

that was supposed to be published as a draft

There is a test that does not do what it claims to do.
Otherwise looks good.

remove useless return, fix comment nit, use MainNet testing setup for new test (needed after changing HasCheckPoint in parent diff)

use GetRequireHeight in HasCheckpoint

check last checkpoint height to decide whether checkpoints should be validated
This requires a unit test to now use the MainNet params.

-,

rebase

this need some work to raise other limits first. Clipping broadcast_max at 1000 makes the situation actually worse than our current limit (INVENTORY_BROADCAST_MAX_PER_MB * 32 = 1120)

fix the while condition, add a ceiling to the max value (1000)