> since one version’s forward and the other backward
Ah, I didn’t realize that at first read.
This analysis task is usually very heavy, I suppose. I tried `—txid` with some randomly picked recent txes, but the execution didn’t finish after quite some waiting (an hour or so). I’m currently running with `—all`, and the processed height so far is 31k after about 20min (using 1.4GB RAM). I guess the processing speed and RAM usage can become worse and worse as the size of ancestry keeps increasing. If the expected use case is to run it only once in quite a while when researchers need to extract the latest data, maybe the `—txid` option is not very useful and can be dropped.
Maybe it makes sense to allow incremental mode similar to `monero-blockchain-blackball`? Also, how about putting the analysis result into LMDB for easier handling?