LazyScholar

I could do it. Although the first PR is a very minor fix. So it can wait (unless you want to keep PRs for issues strictly separated).

> > I could do it. Although the first PR is a very minor fix. So it can wait (unless you want to keep PRs for issues strictly separated). >...

I deleted some dabbling comments i made while experimenting with other solutions. I settled with those two commits: - f6d17a0 : asset registration by `asset_path::String`, `asset_list::Vector{HTMLAsset}` fields passed by the...

> Thanks for iterating on this! This exposes a completely new API for plugin writers, so it's worth taking some time to get this right. I agree. > * I...

> Do you have any thoughts or see any issues with this design? I think that would be perfect. So registering the `HTMLAsset` will push them into a vector. Which...

Changes on 4407a4c: - moved the checks of `HTMLAsset` to `asset()` and added checks for new `uri`&`src` fields - added documentation on `HTMLAsset` and `asset()` - now the plugin assets...

Still pondering on how to write a short overview on "How to Plug-in" for the docs. I am afraid it is no easy task. Looking at how `DocumenterCitations.jl` does the...

> > Adding a additional field like index::AbstractFloat (defaulting to 0.0) allowing to sort the assets. > > I am not a huge fan of this idea -- it doesn't...

> I added a test in #1663 which is now broken, because absolute paths are now okay. With that in mind, could you add tests that: > > 1. Test...

@mortenpi is the current implementation okay so far? Possible changes would be: - place a absolute path check in `HTML` struct constructor - split the function `asset()` function into a...