htmx

nanomorph is 1.4k min/gz'd: https://bundlephobia.com/package/nanomorph@5.4.3

if it works well, I think that's well worth having ootb in htmx

one issue with the extension approach afik is that you can't do like morphdom + innerHTML

the hx-preserve approach works w/ nanomorph

I have a flyout component with a signout link, clicking that link does a boosted load

before the whole thing got replaced so the state of the flyout was blown away

but now it's preserved correctly so it stays open

ah the preserve approach is borked on back though

lol that uses the default swap which has the re-init issue

there's no way to use a custom swap style for history is there 🙂

https://github.com/bigskysoftware/htmx/blob/87e69fcd69a37a0d3ef9b904d5c22f13ba773c15/src/htmx.js#L1996

```js
            if (cached) {
                var fragment = makeFragment(cached.content);
                var historyElement = getHistoryElement();
                var settleInfo = makeSettleInfo(historyElement);
                swapInnerHTML(historyElement, fragment, settleInfo)
                settleImmediately(settleInfo.tasks);
                document.title = cached.title;
                window.scrollTo(0, cached.scroll);
                currentPathForHistory = path;
                triggerEvent(getDocument().body, "htmx:historyRestore", {path:path});
```

we could allow for a pluggable deal here, but there would almost certainly be cases where you are so far away from the current setup that it state won't be preserved

in this specific case the state is living in the html node

https://github.com/alpinejs/alpine/discussions/2924

it's a miracle that anything works, given the history API, to be frank

that's where the magic happens, the rest is just misdirection