fix: reset pruned sender numbers on stage drop

[hai-rise]

We may get over 100% stage progress if we do:

1. Prune some senders
2. Drop the sender recovery stage
3. Rerun the sender recovery stage

INFO ... stage=SenderRecovery checkpoint=4889589 target=19426586 stage_progress=101.19%
...
INFO ... stage=SenderRecovery checkpoint=16606018 target=19426586 stage_progress=177.06%

This is because we count pruned transactions towards progress:

reth/crates/stages/stages/src/stages/sender_recovery.rs

Lines 291 to 303 in f9b8dc4

	fn stage_checkpoint<Provider>(provider: &Provider) -> Result<EntitiesCheckpoint, StageError>
	where
	Provider: StatsReader + StaticFileProviderFactory + PruneCheckpointReader,
	{
	let pruned_entries = provider
	.get_prune_checkpoint(PruneSegment::SenderRecovery)?
	.and_then(|checkpoint| checkpoint.tx_number)
	.unwrap_or_default();
	Ok(EntitiesCheckpoint {
	// If `TransactionSenders` table was pruned, we will have a number of entries in it not
	// matching the actual number of processed transactions. To fix that, we add the
	// number of pruned `TransactionSenders` entries.
	processed: provider.count_entries::<tables::TransactionSenders>()? as u64 + pruned_entries,

This naive fix resets the pruned numbers when we drop the sender recovery stage. Hope it doesn't break something else 😅.

[shekhirin]

That's fair, and we need to do the same for other segments

[mattsse]

I see, ty

cc @joshieDo for vis, the other checkpoints we also need to update here @shekhirin ?

should we do this here or as followups?

[mattsse]

sorry forgot to pull the trigger on this one