There is literally no point of playing in real time of the four hours of the characters doing nothing until the dragon attack happens and skipping it is not railroading by any sensible definition of the term.
Agreed. But that doesn't mean that I didn't railroad them.
Fundamentally, it's railroading whether the players see the rails or not. So taking your case, I say something like, "Does anyone want to do anything before nightfall?" and I know (but the players don't) that a dragon attacks in the midafternoon.
What I'm really saying then is, "Can I leave you in your current position until the dragon attacks in the mid-afternoon?" Now of course, the players don't know that I'm saying that. They don't know if something significant is going to happen or not. But if I let them play out all that "boring time" I'm avoiding any chance that they will do something to prepare for, avoid, or prevent the dragon attack.
Because if I let them play it out they might ride out of town looking for something, or they might decide to split the party, or one of them might go up into a watch tower and possibly spot the dragon approaching, or they might wander over to a powerful NPC and by their presence change how they NPC behaves during the planned dragon attack. The fact that the players don't know or haven't considered whether I am negating all of these options that they have doesn't mean I'm not negating them. It just means that I have decided that either I don't think there is much chance they will do anything useful, or more subtly that I have decided there is a chance they might do something significant and I would rather they don't.
This is why hand waves and time skips are such a powerful railroading technique. Players generally are on board and generally don't realize their agency has been limited, especially if I get consent. And like you they naively think, "Well, he interrupted me when the dragon attacked, he hasn't taken any of my choices away."
That's just not the case. Now, I'm not saying that I deliberately try to take away player agency very often, but I am saying that there are times it is justified - such as to keep the pace up when the game is stalling. A lose of player agency is worth avoiding playing out four hours of uninteresting low stake activity, even at the risk of preventing actual high-stake decisions that are the result of "random" action by the players.
