Rudi Grinberg comments

Results 729 comments of


                                            Rudi Grinberg

Huge GC pressure when serving files with `Cohttp_lwt_unix.Server.respond_file`

We don't even need cstruct if we adopt core_kernel for every backend since `Bigsubstring` is analagous AFAIK.

Provide an example of a tls client

That definitely sounds like the way forward. I'd like to make cohttp usable without conduit anyway.

callv: graceful handling of remote close

@madroach is there a particular reason why you need pipelining? Pipelining is mostly a misfeature avoided by most http clients. I'd suggest you to do the same here.

simple parsing/printing of requests/responses

@marklrh Would this blocking IO depend on unix? IIRC @seliopou worked on this before. Let me recall what this would require.

simple parsing/printing of requests/responses

@marklrh then I think `block_io` is a bit of misleading name. If it isn't using unix how is it blocking? Something like `string_io` or `string` In any case, let's see...

create HTTP redirect example

@edwintorok Doesn't your conduit patch help mitigate this?

create HTTP redirect example

@rathereasy Could you give this a try with the latest version of conduit and see if it helps.

Make Uri optional when constructing requests

The above reference a possible type for targeting a cohttp client. `Uri u` would give us the current behavior.

Latency benchmarking cohttp shows queueing

@jlouis Thank you very much for doing this. This is hugely appreciated.

Blocking Unix implementation of IO module?

+1 this would be very useful