Cameron Elliott

Hey @OliverKillane, you probably saw already, but I submitted a PR to your repo with these features: (simple, but maybe helpful) https://github.com/OliverKillane/divan/pull/1 1. Append to the output files rather than...

> * When appending to the same json, we need some way to identify benches by name (e.g. json is `{ "benchname1" : {..ben data..}, "benchname2" : ... }`, requires...

> I see @cameronelliott & I've approved. > > We still might want to add the following: > > * [ ] truncate file ... > * [ ] include...

Hey @nopeslide @OliverKillane, I tried running the Json output through a Json to Rust type creator. It works, but the types are based upon the names of the benchmarks. This...

I agree the straight json trees can have a lot of value. But! I also think flattened json can have a lot of value too, many developers use json to...

@OliverKillane thanks for doing that. I've [submited a PR to you](https://github.com/OliverKillane/divan/pull/2) to do the following: - make changes to enable the dozens or hundreds of json to type converter tools,...

To others reading along, I think the json flat output could use more review and possibly more tweaks. There is the option of star or snowflake like schemas, or just...

For those that are curious, attached is an example of the jsonflat format, based upon a pending PR to @OliverKillane. The attached file has [jsonlines format](https://jsonlines.org/), and so there is...

Two issues have revealed themselves using this PR by @OliverKillane with my mini-PRs for flat-json. (Also, thanks again to Oliver, for submitting this useful PR) * let me say up-front,...

One other nit that is worth mentioning, is the names on the internal data structures are not a great fit for some of the output: ``` "size": { "fastest": 0,...