Always include CSS in revealjs slides

[sagikazarmark]

I want to include a certain piece of CSS in all of my slides (regardless of the theme), but still want to be able to use custom CSS on a per slide basis if I want to.

Currently if I set a default value for theme with a custom css file, it gets overridden as soon as I pick another theme in a slide (which is kind of expected).

I'd like to be able to include a CSS file in addition to the theme.

A good enough workaround for now is using include-in-header, but that can still be overridden.

[cderv]

Currently if I set a default value for theme with a custom css file, it gets overridden as soon as I pick another theme in a slide (which is kind of expected).

Can you share a reproducible example of what you are currently doing ? That would help us understand correctly all the context. In particular, how to you assign another theme to a slide ? Thank you

[sagikazarmark]

Sure!

Here is what I have in my _quarto.yml file:

format:
  revealjs: 
    theme: [default, custom.scss]

Then in any slide, if I want to use another theme:

---
format:
  revealjs: 
    theme: sky
---

I want to be able to include my custom.scss in all of my slides without having to remember to include it when I don't use the default value.

What I'm currently doing is I have include-in-header defined in my _quarto.yml to include custom CSS. I don't use include-in-header at the moment anywhere in my slides which is why I can rely on the default.

[cderv]

Oh I see. This is indeed not possible as theme in main YAML will be replaced by the one in local file, and currently custom theme needs to be passed at same layer as theme list.

It would require another mechanism to allow that 🤔

I'll open an issue to think about it further

[sagikazarmark]

Appreciate it, thank you. It might probably make sense to think about this in a broader context.

For example: I'm pretty sure that the same would be useful for websites as well (not sure how the config works currently, but I'M guessing it's similar to reveal).

I actually like the fact that there is an overriding mechanism, so I definitely agree that this should be some other, global configuration.

[mcanouil]

Closing in favour of #4193