Add support for link existence awareness

[cdent]

It would be handy if the HTML produced from the markdown classed wikilinks as either existing or not existing.

To do this the plugin would need to have access to the current store and whatever the context of the current tiddler is (recipe or bag).

Since this could be a costly operation on some installations, it ought to be optional (via config?).

Why have it at all? Because knowing that a link needs to be fulfilled is part of the WikiWay and it would be nice to get some wiki-ness back into the game.

[cdent]

@FND: Since you're the (other) major user, do you care about this? The win is prettier links. The lose is some slowdown to access the store.

[FND]

I rarely notice the difference - but sometimes I do, and then it seems like an annoying omission.

So in the Little Big Details sense, it'd probably be a good thing to have.

[cdent]

I'm struggling to find the will to do this one. The current speed of the markdown is really nice. While it unlikely that existence checking will have much (if any) noticeable impact I don't want to touch it. For now will stall until it becomes especially annoying.

[cdent]

@pads and @colmjude, you guys have opinions on this?

[pads]

It seems like a lot of effort for some small prettyfication to me. Given our current stack of work, this is pretty low down the pecking order. I'd close this issue and re-open it if enough people shout about it.

[colmjude]

It is definitely something worth having. It is bad UX not to have it. Plus as the amount of content grows on TS, the number of people viewing content will grow, making this a larger issue. Agree with @pads that it isn't high on the priority stack at this point but is worth keeping on the backlog.

[FND]

FWIW, a temporary(!) workaround might be to do this client-side, using JavaScript to GET the respective URIs and setting a "missing" class on 404.

[pads]

That might result in many, many GETs for some pages which seems a bit messy? I think it's either do it properly now or some time in the future. I say - the future.

[FND]

Many GETs are not a bad thing though. One might also use HEAD if that's a concern.

[cdent]

@FND, I think writing the code for doing it client side is likely harder to get smooth and effective (nevermind properly distributed to the correct pages) than it is to just do the work server-side.

It's not a huge code change. I was thinking of it as a potential small win to do when it would be nice to just push some code rather than all the other stuff that takes too long to actual reach some measure of "done".