Jacob Brown

I am also curious about how one would do this with shio... There is a pattern where you put a core's handle in thread local storage, and then you can...

But that is the point: the `Client` *should* be put in thread-local storage (or made reusable in some other way), to take advantage of the internal connection pool, keep-alive (optionally),...

It's not applicable verbatim, but Rocket's security checks for paths might be a good starting point: https://github.com/SergioBenitez/Rocket/blob/master/lib/src/request/param.rs#L328.

@pinkforest Thanks for the response! Unfortunately I don't have the the time or familiarity with this project to work on a fix. Also, I fear the "damage" is already done,...

Is there any update on this?

Hi @kunjee17 , sorry for the delay in responding. That is a great question. It would definitely be possible to pass in pieces of the HTTP request to the handlers....

@kunjee17 I've made a first attempt at this here: #23. I don't think this will be too obtrusive to the non-HTTP use case. You would be responsible for copying identity/auth...

@kunjee17 Thanks for the additional information. I think #23 is now pretty close to how this will work. Getting data from the HttpRequest in a future exposed some pretty large...

It's mostly agnostic—you can put in bytes/request and get out a serializable response. The current "integrations" (actix-web and hyper) are at the HTTP layer, but these are pretty thin wrappers...

@fulmicoton Good to know; thanks for the issue! Is the issue `commit`-ing frequently, or not re-using the `IndexWriter`? It doesn't look like calling `commit` itself is too costly. For example,...