fix: boost allocate commands

[strahe]

This PR fixes two issues with boost allocation:

1. When the number of entries in --piece-file exceeds the default batch size of 500, the datacap amount calculation is incorrect, leading to erroneous messages.

2. When sending multiple messages, the nonce value remains the same because the datastore in SignAndPushToMpool does not persist(is not shared), so only the first message can be successfully sent.

After:

Tested in my calibnet environment.

[LexLuthr]

Can you please also fix the datacap calculation for extend-claim command as well?

[strahe]

Can you please also fix the datacap calculation for extend-claim command as well?

The logic for calculating datacap in extend-claim should be correct.

boost/cmd/boost/util/util.go

Lines 370 to 376 in 84f46db

	// Calculate Datacap for this batch
	dcap := big.NewInt(0)
	for _, k := range batch {
	dc := newClaims[k]
	dcap.Add(dcap.Int, dc.Int)
	}