Support short links

[Nadrieril]

This adds support for short <1234567> links, as discussed in #59. In fact what this does is render any [foo](1234567) link as a zettel link. That can easily be changed if needed.

Some questions: where should I document this ? Do you use a Haskell formatter and if yes which one ?

[srid]

render any [foo](1234567) link as a zettel link

Can you post a screenshot? What happens to the "foo" in HTML view?

[srid]

Do you use a Haskell formatter and if yes which one ?

ormolu. See https://github.com/srid/neuron/blob/master/CONTRIBUTING.md

where should I document this ?

./guide zettelkasten.

[Nadrieril]

Can you post a screenshot? What happens to the "foo" in HTML view?

[foo](1234567), <1234567> (which is the same as [1234567](1234567)) and [1234567](z://) are all rendered exactly the same. The foo is dropped entirely. I imagined this would only be temporary: I would expect [foo](1234567) to be a link with text "foo".

[Nadrieril]

How can I make tests for this new feature ? I would imagine tests that test both the html output of a file and the graph output of a bunch of files, but I don't know how to write that myself.

[srid]

I think you should ignore the [foo](1234567) case. That can be handled in a different PR with a concrete spec.

This PR should handle the single case of <1234567> (same as [1234567](1234567)), making it behave the same as [1234567](z:).

How can I make tests for this new feature ?

I've been meaning to begin writing tests for link actions, but if you want to take stab, go ahead! Tests use hspec. Maybe add a test/Neuron/Zettelkasten/Link/ActionSpec.hs that tests linkActionFromLink by running it over each individual cases. See IDSpec.hs for reference implementation.

[srid]

I think you should ignore the [foo](1234567) case.

From the docs in the diff,

You may also annotate the link (ignored by neuron):

- [2008403](z://foo-bar). 	- [foo-bar](2008403).

I like the annotation feature in general. But if we finalize the spec, then we can't expect it to be a link with text "foo-bar" in the future (that would not be backwards compatible).

I suppose the user could specify that the link should be rendered normally using a query like this:

[foo-bar](2008403?linkTheme=simple)

[Nadrieril]

I like the annotation feature in general.

What's the use for annotations ? Is it to make it easier to know which zettel it corresponds to ? Then I'd say <1234567?comment=foo-bar>, makes a lot more sense semantically than ignoring the link text.
But more generally, for a thing that kinda functions as a wiki, it probably makes sense to sometimes use human-readable ids.

if you want to take stab, go ahead!

I'll let you do it, you know the codebase. I'll happily add tests to it once it's started though.

[srid]

Yes, annotations are mainly to act as an immediate visual hint when editing a zettel to know what a particular zettel ID refers to at first glance without having to open it.

I prefer something like <1234567#foo-bar> (i.e., the "anchor" is ignored) for annotations. Although anchors could potentially be used to link to a particular section of the zettel. But this can happen in different PR I guess.

But more generally, for a thing that kinda functions as a wiki, it probably makes sense to sometimes use human-readable ids.

The proposal at #70 currently stands at supporting alphanumeric IDs. Perhaps we could add extra characters, like -, to it - so that human-readable ids like "foo-bar.md" is supported?

I'll let you do it, you know the codebase. I'll happily add tests to it once it's started though.

Alright.

[srid]

Added tests and merged in 0aac0fa