While calling things 'right to repair' is a nice easy wording for the lifecycle problem, it doesn't really show the complexity of the problem at hand.
There are plenty of things you can do as a manufacturer to make things repairable, but there are just as many things that you can't do without creating major issues in other areas.
Say you are a company that has a brand identity strongly linked to a certain aesthetic and a law now bans that because you are no longer allowed to glue on textiles on your product as a neat design choice. (yes, this has obvious/glaring problems on its own) What's the solution here? Ban a certain taste or preference for the certain many in the name of making it better for an uncertain few? This might also impact production processes in this example when you have a 'compromise'; i.e. you are required to sell spare parts because in the mind of policy makers "if they can mount that part in the factory, why couldn't they sell it separately".
Say you have a design that is now optimised for production and the method with which a fabric or textile part is added is by preparing a bunch of glue on the spot where the textile goes, then laying a sheet of textile on top of that, and then using a specifically shaped melting device to melt it into the glue, and another cutting device to neatly cut all the excess off, and all of this while under a specific amount of pressure, maybe some inert gas etc. This is how industrial production works, and one of the few ways it works at all (at scale). But that would not be possible since it isn't repairable enough according to some people in certain echo chambers. So now we would have to ban an entire class of production and design methods... and that's just one example.
The goal is relatively simple, and there will definitely be an impact that is worth attaining. But it's unlikely to be as simple as "provide us all the specs, datasheets and parts".
Just as unlikely that it could be "provide us with the private key to your PKI so we can sign our own firmware". (which ironically is practically banned by the FCC on all devices with wireless communication modules - technically it's just an implementation you can restrict to a data table in a driver for a PHY, but practically this is much cheaper as a manufacturer to just blanket-sign and be done with it, which isn't fun for us hackers, but isn't surprising at all.) And not signing or encrypting things is a whole class of problems on its own. We're basically screwed, and since we're often on shared systems (like telecommunication networks, or physical roads and buildings, or we share the fuels and oils across many consumers) it's not as simple as 'do whatever you want' either.
It's tough.