Consider shipping patternlab-node with connect task

[bmuenzenmeyer]

As tridnguyen pointed out on twitter:

@bmuenzenmeyer hey is there a reason why there’s no connect task to serve static files for patternlab-node?

— Tri Nguyen (@tridnguyen) July 15, 2014

<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>

There may not be a good reason. He goes on to state that out of the box, pl ships a means to compile templates to generate files, but no way to view them in a "production" setting.

A question to @dmolsen and @bradfrost and @benedfit, do you see any downside to including mini-webservers with patternlab?

Pros

• A decision made for end-users is a good one if it reduces barrier to entry
• More precise analog to actual web server than filesystem

Cons

• A decision made for end-users is something not everyone may want (see pattern-only discussions)
• Shipping more complexity

It's fairly easy for a savvy designer/developer to add or remove grunt tasks for connect, so the decision really falls, in my mind, to what's best for those new to the space.

Also - is this something the spec should address?

[dmolsen]

If it makes sense for the platform then 👍 to including it with the node version. The few node projects I've seen have included that kind of thing. While the PL front-end should work sans web server I think most people just expect to use a web server w/ absolute paths so shipping/starting it is fine. Rails introduced that as a common thing and even PHP 5.4 ships with a built-in web server now.

I don't think this is a spec MUST or SHOULD so I probably won't address it. I'd like to preserve the web server-less aspect of PL. At least for a little longer ;)