Generate blobs for the state reconstruction test

The test in crates/l2/tests/state_reconstruct.rs replays a fixed set of blobs to reconstruct. The fixtures need to be regenerated whenever the genesis file changes, because a new genesis alters the hash of the very first block and, by extension, all descendant blocks. Our stored blobs encode parent pointers, so stale hashes make the fixtures unusable. If you ever need to regenerate those blobs, you need to change the files payload_builder.rs and l1_committer.rs and run the sequencer to capture fresh blobs.

Summary

1. Limit every block to ten transactions to make the batches predictable.
2. Store each blob locally whenever the committer submits a batch.
3. Run the dev stack long enough to capture six blobs and move them into fixtures/blobs/.

1. Cap block payloads at 10 transactions

Edit crates/l2/sequencer/block_producer/payload_builder.rs and, inside the fill_transactions loop, add the early exit that forces every L2 block to contain at most ten transactions:

#![allow(unused)]
fn main() {
if context.payload.body.transactions.len() >= 10 {
    println!("Reached max transactions per block limit");
    break;
}
}

That guarantees we have transactions in each block for at least 6 batches.

2. Persist every blob locally when the committer sends a batch

Still in the sequencer, open crates/l2/sequencer/l1_committer.rs:

• Add this helper function:

  #![allow(unused)]
  fn main() {
  fn store_blobs(blobs: Vec<Blob>, current_blob: u64) {
      let blob = blobs.first().unwrap();
      fs::write(format!("{current_blob}-1.blob"), blob).unwrap();
  }
  }

• At the end of send_commitment (after logging the transaction hash) dump the blob that was just submitted:

  #![allow(unused)]
  fn main() {
  // Rest of the code ...
  info!("Commitment sent: {commit_tx_hash:#x}");
  store_blobs(batch.blobs_bundle.blobs.clone(), batch.number);
  Ok(commit_tx_hash)  ```
  
  }

Running the node with the deposits of the rich accounts will create N-1.blob files (you can move them into fixtures/blobs/ afterwards).

3. Run the L2 and capture six blobs

Start the local L2 with a 20 seconds per commit so we have at least 6 batches with transactions:

COMPILE_CONTRACTS=true cargo run --release --bin ethrex --features l2,l2-sql -- l2 --dev --no-monitor --committer.commit-time 20000

In another terminal run:

make init-prover-exec

Once the sequencer has produced six batches you will see six files named 1-1.blob through 6-1.blob. Copy them into fixtures/blobs/ (overwriting the existing files).