Update libuv to 1.0.2

[akoeplinger]

Instead of installing the rc2 version of libuv, we should maybe update to the latest 1.0.2 release.
Additionally, they are now on libuv/libuv instead of joyent/libuv: https://github.com/libuv/libuv/releases/tag/v1.0.2

[Eilon]

@lodejard for this. We presumably need to update Kestrel as well, which is ultimately what uses libuv.

[ahmetb]

Is it confirmed that 1.0.2 has the required fix? Perhaps debian:wheezy's package index is updated to install this from apt-get? Can we please check those? cc: @muojp

[akoeplinger]

Is it confirmed that 1.0.2 has the required fix?

Which fix?

Perhaps debian:wheezy's package index is updated to install this from apt-get?

Not yet (and I doubt it'll happen soon).

[muojp]

Switching to 1.0.2 and new libuv home sound reasonable because there will be no further releases held on old GitHub repo ( https://github.com/joyent/libuv/ ).

I tested KRE-Mono 1.0.0-rc1-10831 with

1. latest Kestrel(1.0.0-rc1-10848 from NuGet) + libuv-1.0.0-rc2
2. latest Kestrel(1.0.0-rc1-10848 from NuGet) + libuv-1.0.2
3. latest Kestrel(1.0.0-rc1-10848 from NuGet) + libuv-1.1.0(a Christmas present by libuv team 🎅 )

using helloworldweb project in https://github.com/davidfowl/HelloWorldVNext and putting some load by apache-bench and all of them worked fine (note: I found KRE-Mono 1.0.0-beta1 doesn't work well with latest Kestrel. I'll file this as another issue -> was my mistake).

So, I don't think there are urgent reasons (e.g. show-stopping breaking changes) to switch to newer libuv. Don't see enough reason to stay on older releases either. Of course we can't say there won't be breaking changes in future Kestrel releases, newer is better.

By the way,

Perhaps debian:wheezy's package index is updated to install this from apt-get?

Not yet (and I doubt it'll happen soon).

+1. libuv in (even) sid is 0.10 (ref: https://packages.debian.org/search?keywords=libuv ). 0.10 series are still maintained by libuv team but well, they are somehow different from 1.x.x.

[ahmetb]

Closing to move on at #22.