bug(replays): Correctly extract DOM event HTML that is captured during FullSnapshot (#41555)

When a FullSnapshot happens we should look ahead in the list of
breadcrumbs for the next crumb, and then see if the `nodeId` for that
next event is already in the page. If it is in the page, we'll cache
that html.
Why? Well, say that DOM node is a button that gets removed when you
click it. If we waited and tried to find the HTML after the click event
happened, the HTML will already be gone! So we grab the element before
it's gone, so we can match it up to the event no matter what.

See https://github.com/getsentry/sentry/pull/40421 for more descrption
of the problem. That PR implemented the bulk of the fix, but it didn't
handle the case where the DOM node is created during the initial
FullSnapshot.

Fixes #40715

static/app/utils/replays/hooks/useExtractedCrumbHtml.tsx

@@ -149,7 +149,9 @@ class BreadcrumbReferencesPlugin {
   }
 
   handler(event: eventWithTime, _isSync: boolean, {replayer}: {replayer: Replayer}) {
-    if (event.type === EventType.IncrementalSnapshot) {
+    if (event.type === EventType.FullSnapshot) {
+      this.extractNextCrumb({replayer});
+    } else if (event.type === EventType.IncrementalSnapshot) {
       this.extractCurrentCrumb(event, {replayer});
       this.extractNextCrumb({replayer});
     }