fix(react): replacing route uses new route direction and animation

[aeharding]

Issue number: resolves #24260

---

What is the current behavior?

When replacing a route (router.push(newRoute, 'none', 'replace')) the RouterDirection from the route being replaced is being used (if it exists) instead of the new one the user specifies.

What is the new behavior?

User-specified RouteDirection is used, if it exists. If it doesn't it falls back to the RouteDirection of the route being replaced.

Does this introduce a breaking change?

• Yes
• No

Other information

Please see the following comment for why I think the current behavior is incorrect, and why this change is needed: #24260 (comment)

[liamdebeasi]

Thanks! Some more context for why this bug exists. When we are replacing, we need to copy over some of the previous state. For example, consider a history of A --> B. When replacing to C, we are replacing the state of B with the state of C. As result, the previous route for C is A (since B no longer exists).

Due to this requirement, we correctly copy information from currentRouteInfo first before falling back to routeInfo:

ionic-framework/packages/react-router/src/ReactRouter/IonRouter.tsx

Lines 174 to 189 in 37290df

	const currentRouteInfo = this.locationHistory.current();
	/**
	* If going from /home to /child, then replacing from
	* /child to /home, we don't want the route info to
	* say that /home was pushed by /home which is not correct.
	*/
	const currentPushedBy = currentRouteInfo?.pushedByRoute;
	const pushedByRoute =
	currentPushedBy !== undefined && currentPushedBy !== routeInfo.pathname
	? currentPushedBy
	: routeInfo.pushedByRoute;
	routeInfo.lastPathname = currentRouteInfo?.pathname || routeInfo.lastPathname;
	routeInfo.prevRouteLastPathname = currentRouteInfo?.lastPathname;
	routeInfo.pushedByRoute = pushedByRoute;

However, we should not do that for routeDirection and routeAnimation. The reason is that developers may have specified a custom animation or direction to use with this replace operation. We should prefer that first, and then fall back to the data in currentRouteInfo if the developer does not pass any information.

[aeharding]

Thanks @liamdebeasi, updated!

[liamdebeasi]

Good to go once the final change is made. I'm making a dev build now and will post this in the linked thread.

[thetaPC]

LGTM, works as intended.

[sean-perkins]

Thank you for fixing this!