uzbl + webkitnix == ?

[mathstuf]

I'm an uzbl[1] developer which uses GTK port currently. I was wondering what you think the suitability of the Nix port would be for uzbl. Basically, uzbl just exposes the GTK's port with a command-like API, a bare-bones UI, and a socket for communications.

We're currently looking at breaking backwards compatibility to migrate over to the overhaul on my fork[2] which cleans up the code quite a bit, syncs with lots of features that are missing in the current releases, and cleans up uzbl's API to allow some more controls over things (mainly the async I/O support).

I think Nix could be interesting if it breaks uzbl off from an X dependency which would then potentially allow a uzbl-gtk, uzbl-qt, uzbl-efl, uzbl-ncurses, etc. UI layer on top of it.

Nix could benefit as well since uzbl gives a really nice lightweight sandbox to test out new APIs with a real browser. I've been able to test new APIs pretty quickly since uzbl itself doesn't care about how they work as much as that they do (the authenticate-signal in the GTK port was, AFAIK, first actually used by uzbl (though just my fork)).

[1]https://github.com/Dieterbe/uzbl
[2]https://github.com/mathstuf/uzbl/tree/next

[hugopl]

Nice to hear you plan to use Nix on uzbl, however I didn't get why this was filed as an issue here instead of a e-mail on Nix ML.

[mathstuf]

On Thu, Sep 26, 2013 at 12:04:03 -0700, Hugo Parente Lima wrote:

Nice to hear you plan to use Nix on uzbl, however I didn't get why
this was filed as an issue here instead of a e-mail on Nix ML.

Well, there aren't "plans" per sé, it's more of a discovery phase. I
don't recall exactly why I put it here and not on the ML. I think my
thought process was for there to be a single point-of-reference for any
issues opened.

I'll repost to the ML on Monday (I'm away for the weekend).