docs: explain checkpoint file

README.md

@@ -4,6 +4,10 @@ Extract logs expected by [firehose-starknet](https://github.com/starknet-graph/f
 
 This tool exists because it could be slow to sync a Starknet node from scratch, which is currently the only way to bootstrap a [firehose-starknet](https://github.com/starknet-graph/firehose-starknet) deployment. This tool makes calls to the JSON-RPC endpoint offered by a trusted synced node, and emits the exact same format as an instrumented client node would do.
 
+## Checkpoint file
+
+This tool keeps track of the sync progress through a _checkpoint_ file, specified by the `--checkpoint` command line option or `CHECKPOINT_FILE` environment variable. The content of the checkpoint file is updated every `50` blocks, which means it's likely that some blocks might be sycned more than once when the tool restarts. This is fine as Firehose handles duplicated blocks without issue.
+
 ## License
 
 Licensed under either of

src/main.rs

@@ -135,6 +135,7 @@ async fn run(
         };
 
         if checkpoint.latest_blocks.len() >= MAX_BLOCK_HISTORY {
+            // TODO: use ring buffer instead as this is extremely inefficient
             checkpoint.latest_blocks.remove(0);
         }
         checkpoint.latest_blocks.push(new_block_info);