Vasilis Papavasileiou

The problem comes from Ocamlnet. We can only possibly use Ocamlnet `4.1.2`, and that one is explicitly incompatible with Ocaml `4.04+flambda`. The easiest solution would be to stick to the...

It is not really our network infrastructure. We only use peripheral parts of it, e.g., the PCRE bindings and the URL manipulation module. Indeed, we are in the process of...

Would a link in the description of the `File` module help? "If you need to modify the content type, please see `File_ct`", or something to that effect. We could put...

If we add an option `--force-eliom` (or `--ignore-extension`, or something to that effect) to the current `eliomc` / `js_of_eliom`, would that be of any help?

Unfortunately, I much doubt that we can do it. To begin with, `React.S.t` is not marked covariant. Basically all our wrappers use mutation all over the place.

I understand your concerns. But at the same time I think types like these are useful for forcing us to do things in a uniform way (and document the uniformity)....

Thanks for doing this. I haven't read the last-chunk part carefully, but the rest of the code looks OK to me. The overhead (hashtable, stacks, release closures) should be insignificant...

The major part of #54 was a `cohttp` backend, with two separate (but related) efforts towards that goal (#64 #124). If a volunteer wants to pick up that work, that...

That's probably OK. The `http://` or `https://` is optional, and `q7fPbXDbpCLdP5acC7WfCTmeYr2QFnglLhjgCPrYfd` looks like a valid hostname without a port and path (both optional). Do you have an example that is...

We need to intervene in the `Url` module (sooner rather than later) to get rid of the Ocamlnet dependency (`Netstring_pcre`). It may well be the case that `ocaml-uri` is all...