Michael Hixson

> We could have a table to keep track of each framework and the verification status. If contributors are not sure, then someone could help them with manual verification (I...

@billywhizz I think my verification strategy isn't good at detecting batching in single query and fortunes. I send each HTTP request manually with `curl`, but I bet that batching only...

Ah ok. You are sending many requests in parallel because you are running the full benchmark. I was sending a single multi-query request so that the query log and tcpdump...

@billywhizz That's certainly more readable than the tcpdump output I was looking at. Could you push the code to a separate repository so we could look at it?

@billywhizz ok no problem! Thanks for looking into it.

> There are a number of frameworks that seem to be failing the verification only on the update test and they all show the exact same results (10728 missing syncs)....

> so you are saying, for the frameworks that don't use a single bulk SQL statement to do all the updates, it is acceptable to batch together 20 Bind +...

@billywhizz Hmm I'm not sure. [This comment](https://github.com/TechEmpower/FrameworkBenchmarks/blob/73eb2abec4e5c594e40812250ebe061234fe9473/toolset/test_types/verifications.py#L410) in our toolset indicates that siege results in a reliable number of queries. We tolerate a range for some of the related counts,...

@jamiere-msft We'd like to publish results continuously without anyone needing to review each individual set of results. Currently, we publish results automatically at https://tfb-status.techempower.com/ and manually at https://www.techempower.com/benchmarks/. The amount...

> There is just one small issue right now - while a partial set of cached queries results (i.e. for a run that hasn't completed yet) is shown properly, the...