This is clinging on technicality. You may have not explicitly told to player what to think, but when you start to talk about their wish not being in accordance of genre expectation and thus will be adjudicated differently than they intended then you implicitly have. And in practice this really does not differ from what I do. In all my years of playing RPGs I don't remember ever having to really pull a rank as GM or GM doing that to me. It doesn't ever go that far, because the sort of discussion you suggest is enough for the player to get the hint.
Well, for starters, we've not been talking about telling a player what to think, but instead what a player is allowed to assert their character thinks. If we talking about controlling player thought, that's far beyond the realm of the game and into politics, so I'll pass while saying "nope!"
If that was a typo, then what you're saying here is that by pointing out that an action will fail, I'm de facto controlling what a character thinks. That's interesting, if wholly wrong. It's interesting in that is explains a good bit of your thinking, and further reinforces my previous observation that you see any restriction or constrain as condoning all restrictions and constraints, up to those that you don't personally like. So, if there's a restriction on how I will adjudicate an action (which is a GM side restriction, by the way), you see this as forcing player's to have their characters think things so a general restriction on such is equivalent. This fails logically in a few ways. First is the extrapolation from a specific thing to a general thing -- providing ad argumentum that the specific thing is indeed a restriction on character thoughts imposed by the GM. So, even if we accept that, your argument logically fails because you can't draw a line between this instance and a general authorization or agreement to your position.
But, it isn't a similar incident. The adjudication is in the same vein as any other automatical failure the GM adjudicates due to the fictional framing. If my adjudication that F-14s don't exist in my D&D game so a wish for them will fail is controlling player thought, so is my adjudication that you can't magic missile the darkness. These are both because the agreed to rules prohibit these actions -- in the latter because darkness isn't a valid target and besides you can't see it to target it. The former because of the agreed setting rules. Neither is a restriction on PC thought -- the PC is free to wish for an F-14 all they want, as they are free to believe they can magic missile the darkness. By not restriction PC thought, I'm allowing for cases where the PC can think whatever they want at any time, and then controlling the game via my authority to adjudicate.
This last is really the key part you keep missing -- my game is not without constraints, but those constraints exist within the authorities provided by the rules of the game. Those state that the player determines what the PC thinks and tries and that the GM adjudicates the outcomes. So, I don't worry about "metagaming" because the rules say I don't have the authority to. I worry about adjudication and setting, because those things the rules tell me to worry about. I can avoid any problems with "metagaming" with the authorities I have over adjudication and setting, and so do that without worry about the player's authorities. This doesn't mean that anything goes -- a position you seem to enjoy taking that unless the GM controls improper PC thoughts (improper being decided by the GM) then the game will spiral into craziness, like the GM suddenly being powerless to stop rampaging players with textbooks. This despite zero evidence to support it and much evidence to refute it. What's really amusing is that you then claim the defense that you don't even need to use the authorities you've claimed because nothing happens in your game to require it while insisting that such things must happen in games that don't assume this same unused authority.
I, very recently, thought as you did -- that anti-"metagaming" rules where necessary to preserve the game and achieve the kind of play I wanted. I haven't really changed my mind about the kind of play I want, but I have realized that my anti-"metagaming" rules were a detriment to it, not a salve. The real impediment to my games was how I was running them -- the problem was me. I fixed that, and, in the process, realized that anti-"metagaming" rules were unnecessary and actually harmful to my games, so they went as well. This was super easy because the rules don't require them -- they're all just table rules.