component markers

[amitu]

Imagine we allow components to be attached to marker characters, eg

-- import: foo

-- ftd.text: hello 💃world💃
$marker:💃$: foo.bold

Which is equivalent to:

-- import: foo

-- ftd.text: hello {foo.bold: world}

[siddhantk232]

I think that the syntax of this can be improved by https://github.com/orgs/fastn-stack/discussions/884

[amitu]

@siddhantk232 I think you linked to wrong discussion here.

[siddhantk232]

oh no I copied the wrong link. Lemme see if I can find that discussion

[siddhantk232]

fixed the link

[siddhantk232]

We could use TOML's dotted keys here to get the following syntax:

-- import: foo

-- ftd.text: hello 💃world💃
marker: $foo.bold
marker.key: 💃

Equivalent to:

-- import: foo

-- ftd.text: hello {-- foo.bold: world}

[siddhantk232]

I am not sure. Let's suppose marker is a record:

-- record marker:
string key:
component with:

and ftd.text is able to take marker somehow:

-- ftd.text: some text 💃bold💃 😃emphasised😃
markers: { ;; notice the plural `markers`. It's a list
  .key: 💃
  .with: foo.bold ;; f-script so $ is omitted in $foo.bold
}
markers: {
  .key: 😃
  .with: foo.em
} ;; these two header values will be UNIONED together to create a `marker list markers`

We created a record instance (of record marker) on the fly and passed it to header markers of ftd.text component. The other version where record instance was created separately would have been:

-- marker list my-markers:

-- marker:
key: <>
with: <>

-- marker:
key: <>
with: <>

-- end: my-markers

-- ftd.text: some text 💃bold💃 😃emphasised😃
markers: $my-markers

Can we say that: "if the header is a list type then multiple declarations of that header in component call will be a union operation on the value of said headers?"

definition of ftd.text:

-- component text:
marker list markers:

-- end: text

[siddhantk232]

I don't like this. This has to be a compile time transformation (or parse time transformation) so marker header can't really exist on the ftd.text component.

[amitu]

Can we say that: "if the header is a list type then multiple declarations of that header in component call will be a union operation on the value of said headers?"

We did that in ftd 0.2 I think. Not sure if we removed it for reasons, or if it got lost during 0.3 somehow.

Do you think it is a good idea? It does save one reason to use {}.

[siddhantk232]

Let's see:

This is when union of both header declarations will happen

-- ftd.text: some text 💃bold💃 😃emphasised😃
markers: {
  .key: 💃
  .with: foo.bold
}
markers: {
  .key: 😃
  .with: foo.em
}

no union but we have ability to create variables on the fly (a list of marker in this case).

-- ftd.text: some text 💃bold💃 😃emphasised😃
markers: { ;; notice the plural `markers`. It's a list of `marker`
  -- marker:
  key: 💃
  with: foo.bold ;; f-script so $ is omitted in $foo.bold
 
 -- marker:
  key: <>
  with: <>
}

It's not a big deal to allow union of same headers but that is a behaviour that we will have to communicate to the users. I feel like the second code snippet is better as it's more explicit.

[amitu]

In the first snippet, I am not sure the .key: syntax is necessary. We could have just used key:

[amitu]

I am also happy to give up on inline styling related concerns.

These are the motivation for them. But we have plenty of inline styling available in markdown, strong, emphasis, strikethrough, highlight, super script, subscript. What if we "appropriate" those, and allow people to say highlight in this text node will be done by changing the color and strong means circle the word etc.