RFC: default domains folder

[scripting]

The proposed feature

If a request arrives for which there is no sub-folder of the domains folder to route it to, if there is a sub-foldler named default, the request is routed to that folder.

Why?

It makes it easy to set up PagePark without having to worry about where the request is designated to go.

This is particularly difficult when setting up a server on Glitch, where the assigned domain might be something like horse-diego-haircut.glitch.me. It's hard to instruct the user to set this up when your instructions can't tell them exactly what to type. With this feature the instructions can say: Create a file named domains/default/index.md

Concerns

As always the first concern is breakage.

[andrewshell]

This seems like a good idea.

[LeadDreamer]

Isn't this "default" almost by definition a 404? That is the standard for "page not found".... Unless, of course, the intent is a site that always appears to be working, even if the content is missing...

[andrewshell]

The idea is that it wouldn't show up for all missing files, just if there isn't a domain configured for what you're requesting. If you've created a domain folder for a particular domain it would work exactly as it does now.

[LeadDreamer]

While reasonable, I'd personally at least like the default to deliver SOME information indicating what domain was requested, and why it isn't available... don't want it to fail, but don't want it to be TOO polite and quiet about it... From a web-standards point of view, I would expect SOME form of not-200 response...

[scripting]

@LeadDreamer & @andrewshell -- this is an interesting discussion with good points on both sides.

First observation -- PagePark is different from other servers, in that there is currently no default directory.

Apache has one, if you install an Apache server, you get a standard page, with information about Apache.

Same with WebStar and Frontier, the two other servers I'm familiar with.

PagePark does not have a default folder. So when you do a fresh install and visit the server in a browser you get the very unhelpful "The file was not found." Not a great first-time experience for users.

We're now trying to make the setup process smoother, and less error-prone, hence the RFC for the default folder.

As Andrew points out, it is not a replacement for the 404 page.

[scripting]

I implemented the feature.

https://github.com/scripting/pagePark#v0820-7521-by-dw

BTW, it overcomes any objection that it interferes with 404 pages, because the user has to create the default folder.

For now, a newly installed PagePark server does not have a default folder, as with previous versions.

[LeadDreamer]

Given this, and the steps @andrewshell describes in setting up his Glitch version, it now seems reasonable to expect pagePark to setup a default folder (unless one already exists), and to populate it with an OPML file that is fairly technical - a consistent page that identifies it as pagePark, and includes links to this repository, and other FAQ's and How-To's (including how to change/delete the default). As soon as pagePark is installed, the user then has easy access to information to take the next steps...