John Keiser

They wouldn't even have to put their classes into the SIMDJSON_IMPLEMENTATION namespace. singleheader would still need tricks like this as it stands; but there are other ways to skin that...

I'd expect it to look like: ```c++ int i=0; for (auto point : doc["coordinates"]) { int j=0; for (double val : point) { matrix[i][j] = val; j++; } i++; }...

Yeah, using push back is necessary, though we could probably expose an upper bound on the array size if you are willing to risk overallocation... still, I"d test the push_back...

Holy wow, this is the first time (I think?) I've seen an interpreted language plugin that gets us into multiple GB/s! Nice.

If we changed chefspec so that it constructed a resource for its matcher instead of using a hash, would that solve the problem?

@TkTech assuming Python has optimizations for short-lived objects (which I imagine it does), one design I've been thinking about for On Demand python is, to forego the simdjson frontend entirely:...

That's actually how matchers for blocks work. The matcher is responsible for running it.

From the PR: we'd like to not do this globally; but it would be good to get a toggle in, at the least (and turn it on by default if...

@dullyouth https://github.com/chef/chef-provisioning-docker/pull/83 supports all docker options and . Mind trying that branch (`jk/more-options`) to see if it serves your needs? `machine_options docker_options: { volume: [ ] }` should let you...

Note: this is being run as pivotal.