Open matt2xu opened 8 years ago
There are a few reasons that I can recall:
#[unstable(..)]
.nickel
repo only seems to give people the impression that they can't build their own extensions out of tree, having something out of tree can be an example of how to build things without having to relying on internal features. nickel_cookies
. Hopefully this is less of an issue when incremental compilation lands.cookie/secure
which transitively depends on openssl
, which some people don't have an easy time building (windows mostly). We did initially plan to rework things so nickel could still re-export cookies or other major crates through feature flags, letting us develop in downstream crates but package them as a 'batteries included' package, but we haven't really had any push towards that as nobody has really been pushing for that integration (perhaps it's worth looking at again though!). cc @cburgdorf @SimonPersson
and I had to use the git version as I did not find it on crates.io.
Pushing to crates.io is an omission that I will fix, I think the api should be fine to push an initial version up.
Unless I get objections, I will be pulling nickel-cookies into nickel as a sub-crate.
Given the focus of Nickel which is to provide an Expressjs-inspired web framework(*), I would expect that cookies are handled out-of-the-box. Instead, I just found out that I need to use this crate; and I had to use the git version as I did not find it on crates.io.
Why are cookies treated as a separate crate, when it is central to many Web applications?
(*) which I find much superior to Iron's "everything is a plugin", because in Iron's case this creates an unfortunate feeling of "where are my batteries" for everything you want to do, even the most mundane tasks like serving static files and routing requests...