renderFromStorage methodNow, when rendering a document encoded as storage, you can usa this shorthand function.
Before:
render(
registry,
await documentToParsed(
registry,
makeJDDContext(ctx),
documentContainerFromStorage(
page.get("content") as RawJDDocument
)
),
makeJDDContext(ctx)
)
You can just call
renderFromStorage(
registry,
page.get("content") as RawJDDocument,
makeJDDContext(ctx),
);
The insert_nbsp function is more fault-tolerant now - it will not throw an error if given a null value instead of a string
The documentToParsed now will not throw an error when encountering unknown component. The unknown component will now be skipped and a warning printed to the console. An optional handle_unknown_component argument lets you customize this behavior and throw an error if you want.
Add a SingleReference argument type
Added support for adding hardcoded soft hyphens to work around Chrome’s lack of hyphenation dictionary for Polish language 
It is a breaking change, as now calling render_markdown requires you to specify the language of the text within.
Only apply soft hyphens to text parts of rendered markdown, to avoid mangling link URLs
Change the API of the renderHTML function. It now takes a classes argument that is a list of strings and you have to put those classes into the main html element of the component in order for some advanced JDD editor features like scrollto to work.
Add jdd-component class to each rendered component using the api from v.0.4.19. This is used to create a “click component to reveal its params” feature in the JDD editor
Add renderEarlyAssetsFromStorage - an easier alternative to renderEarlyAssets where you don’t have to parse the JDD document that comes from storage. So rendering early assets is now:
renderEarlyAssetsFromStorage(
registry,
(entry.get("content") as RawJDDocument) || [],
makeJDDContext(ctx)
)
Component.storageToParsed now does not throw an error when given values that don’t correspond to an existing argument
This is a breaking change. The toHTML method of the Component class now takes a single object argument, instead of four arguments. This makes the function declaration more concise, as unused args can be simply skipped instead of aliased as _, __, etc.
Before:
async toHTML(
{
title,
content,
images,
}: ExtractStructuredComponentArgumentsParsed<typeof component_arguments>,
{ render_markdown, render_image }: JDDContext
): Promise<Readable> {
After:
async toHTML({
args: { title, content, images },
jdd_context: { render_markdown, render_image },
}: ComponentToHTMLArgs<typeof component_arguments>): Promise<Readable> {