Brian Anderson
Brian Anderson
@batiati I have made the error-handling changes, reintroducing the free function "parse_value_size" and doing the error reporting there, while "ByteSize.parse" returns an error set. I've filed two followup issues around...
https://github.com/tigerbeetle/tigerbeetle/pull/1674 will likely conflict with this. I can wait until that lands before rebasing this one.
@batiati This is rebased over the conflicting PRs and I think all review requests are resolved.
> Sorry for too much back-and-forth, but I'm dogfooding this change and noticing things that may improve my own experience 😉 No worries. I'm happy to do as much rework...
Ah, don't approve this yet. My last patch introduced a new bug.
Ok, now it's ready for another review - with the split digits/units the unit comparison needs to be for equality, nod just "ends with".
``` async fn get_latest_block_number(&self) -> Result { let block_number = reqwest::get(format!("{}/{}", self.url, "blocks/tip/height")) .await? .text() .await?; Ok(u64::from_str(&block_number).expect("u64")) } ``` This code should error if the block can't parse.
In general it looks like we never check that status of HTTP `reqwest` responses: https://github.com/Aimeedeer/realtps/issues/72
I'll give it a look this weekend.
I see that the transactions are dominated by HCS, the Hedera Consensus Service, which I understand went live in the recent past. It's not clear to me what the nature...