Post-clustermempool package RBF: per-chunk processing

It points out that a child transaction could possess a high fee but a low feerate, which complicates its inclusion with a parent transaction in the same chunk. This situation could lead to increased expenses due to the mismatch in feerate. The text suggests that allowing such combinations is feasible without significant computational burdens, except for the linearization step, identified as step (10), which is necessary only after a collection of chunks acceptable for relay is assembled. This step is noted as being particularly resource-intensive, underscoring the need for efficient management of transactions to mitigate costs. The conversation implies a nuanced understanding of transaction handling where balancing between computational efficiency and cost-effectiveness is paramount.