Where they encounter this resistance, the imperial army falls back. Not always quickly or promptly, since this is admittedly less a principled strategy on the the part of their leadership than a belated acknowledgment of reality, but while it's possible to get soldiers to charge hard points over the bodies of their comrades you have to spend a bit working them up to it.  Instead the surviving leadership do their best to keep their units from disintegrating while they draw upon the extensive imperial urban warfare experience (forged, in large part, by fighting themselves in the empire's various civil wars) to link up with nearby units while riders fetch reinforcements. 

Mages are of course ideal for this sort of work, but the legions never have half as many of those as they'd like and losing too many of them tends to make their replacements sign up with your rivals instead, so instead that means waiting on artillery and other siege equipment. It's fortunate for their wait times that purposeful delays aren't really a concern - If the locals are going through this much trouble to keep them out, it's almost certainly somewhere that they really want to get in, and while that doesn't guarantee loot the odds are good enough that nobody's going to be dragging their feet too much. 

And in turn, absent the numbers to actually take on the main force of the alien invasion, dath ilani soldiers will withdraw in the face of any forces sufficient to dislodge them. They're here to protect people, not the city itself, and in general killing enemy soldiers is not actually a very valuable terminal goal here. They'd rather kill people than let them kill others but it would be really much better if neither was happening right now, so delay is the name of the game here.

Command would also rather prefer as few guns and bullets fall into unfriendly hands as possible, and an unsuccessful defense would be pretty bad for that goal. It seems unlikely that whatever set of constraints or alternatephysics resulted in their counterparts having an air force and wormholes but still relying on jabbing people with metal sticks to kill them is particularly stable, and many ways for it to destabilize would be very bad. The issue is probably not just needing a proof of concept but it would be really bad to be wrong about that, especially since enough kinetic energy to keep a wyvern in the sky in defiance of aerodynamics is enough kinetic energy to propel rocks very fast indeed.

Huh, victory! Nothing here looks especially valuable, but they can pry up everything plausible not nailed down anyway to be sure. They'll have to make sure to spin it as "the enemy fled before us"  instead of "we took too long and they escaped," but that shouldn't be too hard to make happen.

It's about three quarters of an hour into his invasion, and legate Cattaneo is feeling cautiously optimistic about the way things are going. It's certainly not an unmixed set of positives, especially not considering the alarming reports starting to filter in second hand about the dragon corps running into trouble, but it could be going much worse. The decision to be particularly cautious with the early scouts he sent through is seeming particularly prescient now given how total the surprise of their attack seems to be as a result, and he's also quite pleased with the amount of loot that looks to be on the table. Much of the city's contents are strange and foreign to his eyes, but coming from another world is sufficiently exotic to sell even worthless trinkets, and some of the more impressive works have risen even to his attention. The glasswork in particular has his people marveling, and he can see why - the great sheets they put in windows would bring pride to any imperial artisan, and the glass ornaments that festoon much of the local clothing are of a particularly fine quality. With any luck it won't be long before they find some of the locals responsible for doing said glasswork, and returning with such artisans would be a true feather in his cap for years to come. It would be better if they can find whatever passes for a local treasury, but even without it there's more than enough loot at hand to keep the majority of his men content.

Of course, it's not all positive news. For all that the willingness of the locals to flee rather than fight back does wonders in reducing casualties, it also makes it difficult to do much in the way of capturing slaves, and combined with his lack of local interpreters means he's no closer to identifying the local elite or rulers. Unless that changes they'll likely be able to smuggle most of their more portable wealth out from under his nose, and it seems difficult to do anything about. And as for a more long term occupation... it might narrowly be possible to keep a sufficiently cowed population of this size under military rule, especially if he could coopt the local institutions to do it, but doing that and defending the place against whatever army marches to retake it is an obvious nonstarter that would get his forces defeated in detail, and attempting it without even being able to communicate with the locals would be madness. Add in that there's some concerning news from the dragon corps as well; he has no idea what trouble they've run into, but something that can bloody their noses is a threat he has to respect at least a little. Perhaps a day to finish establishing their position before starting the process of withdrawing his forces, rather than risk snatching defeat from the jaws of victory if thrice his number in barbarians would arrive next week. 

By this point in the attack, the combined military-emergency services crisis coordination teams have started to piece together the basics of what's going on. The how and why are still incredibly baffling to everyone involved, since even the most parsimonious explanations require a large number of things they had thought they knew about the universe to be instead laughably false, but they do have a wealth of data to work with here. Around the globe, hundreds of teams are hard at work sorting through tens of thousands of eyewitness reports and camera feeds and satellite imagery and other data streams in an attempt to make sense of the madness, and of course they have a direct line to all their own forces at all times. 

From what they can tell so far, the enemy's muster seems to be at least 10,000 strong, with 80th percentile guesses pegging them between 12,000 and 13,000 thus far deployed in dath ilan. Ludicrously small for any proper invasion, even if more continue to spew through the bizarre building, but also far too many for them to simply be dath ilani in some sort of disguise, to say nothing of the fact that such an act would require some truly excellent disguise work. While the majority of the attacking forces seem to be human or animal, it's become apparent that the flying creatures aren't the only sort of bioengineering at play here. Other than their tusks and apparent resilience the greenish-grey skinned sorts might have been able to pass for a human, with some commenters comparing the work to Sparashki, but the same cannot be said of the much larger set of humanoids deployed alongside the primitive invader artillery. The heavyset creatures average over 12 feet tall, and appear to have exchanged some of their manual dexterity for an absurd amount of brute strength and surprising durability, including the ability to manhandle absurdly heavy equipment and ammunition with ease. From all observations the flying creatures are much more impressive, but it takes a fair bit of doing to get something strong enough to compare favorably to a human in power armor. 

Taken holistically it seems obvious that someone is jobbing here, because even with no skills bar biotechnology it would be trivial to do far more damage with a plague. That's bad news for dath ilan's ability to win a serious confrontation with whatever civilization is behind this (and has gotten some very serious people to work ramping up their efforts elsewhere), but encouraging news about this confrontation if their opponents aren't seriously trying to win. One of the major orders of the day is trying to figure out if whatever self-imposed limitations their foes are operating under should be getting them to adopt reciprocal terms, but the whole situation is illegible enough it'll be hard to say with confidence if they succeed in identifying it. For now, the order of the day remains the same  - to mobilize troops and material to respond to this incursion with the minimum viable collateral damage rather than attempt to bomb the gateway into rubble. There are over fifty thousand soldiers already en route by train, and by this time tomorrow they can quintuple that without compromising their ability to respond to another such incursion in a different city.

Right, so your divinations have managed to spot the cavalry, the infantry, the wyverns, the orcs, the siege artillery, and the ogres. In pretty good detail, too; if you take into account the casualties you're aware of and he isn't, in some ways you have a significantly better grasp on the disposition of most of the Legate's forces than they do.

But what about the wizards?

...What's a wizard?

You know what? That's a good question.

I'm just going to let that next part be a surprise, then.

That's certainly not ominous!

After almost an hour of marching and looting interspersed with light fighting, Saderan forces had managed to occupy nearly 6 square kilometers of Schelling Point to varying degrees of success. Absent a clear sense of the relative importance of different areas the imperial soldiers had elected to spread out in all directions at once, but an inability to effectively project forces across the river (courtesy of a handful of drone strikes and the prioritization of this goal by local security groups) and the vagaries of local elevation had left the zone of control looking more like a vague blobby shape than a neat circle, with a number of radial spokes not properly occupied by anyone where Saderan cavalry had ridden ahead of their fellows and then not stopped to secure their gains. 

The extent to which this included large buildings in the area  was also hardly uniform; near the center of the zone around the gate, imperial detachments had been diligent about going door to door and uprooting any potentially problematic staging grounds, but as they traveled outside the immediate supervision of their superiors this diligence gave way to opportunism; any soldier left to their own devices could be relied upon to prioritize any building that offered the prospect of loot over any attempt at completionism, and if the officers would in turn prioritize locked doors for being plausibly important locations they were hardly immune to the allure of wealth themselves. Open storefronts typically took the worst of it, with anything that sold clothing or shiny baubles getting a particularly thorough ransacking, but anything with a shiny facade or large amounts of glass had to take its chances.

Small and lightweight objects, including a number of pried out decorative gemstones, found their way into pockets and packs, while anything to large for that while still of sufficient value to take made their way into the hands of porters and from there onto an efficient network of wagons to head back through the gate. In the traditional imperial manner the lion's share of the resulting wealth would be split up among the legate and his command staff, but enough would go towards celebratory alcohol, entertainment, and cash payouts to keep the soldiers happy and everyone's interests pointing in the same direction. Falmart might lack the specialized language of dath ilan for describing incentive alignment, but early imperial conquerors had made the advantages of an army that actually wanted to win obvious. 

(This increased efficiency of looting had, of course, made every resultant internal conflict far more devastating to the defeated cities, but it also gave the besieging generals a potent tool in encouraging defections and surrenders from the defenders. The soldiers expected it from their leadership, particularly after a hard siege, and the only ways for locals to limit the devastation of a loss were a quick surrender or a deal with the person directing the wagons of stolen goods; when the extraction was coordinated from up aboce, hiding your most precious possessions and hunkering down to wait it out became a far less practical as an option.)

Almost nowhere else, however, were they ever actually stopped. The designers of Schelling Point had made it under the assumption that most long distance transportation would take place via underground automobile and train tunnels, but significant effort had been involved at every step to ensure the resulting construction would be both safely and pleasantly walkable. It's not true that literally nobody would have moved in otherwise, but if you wanted people to live in a city that couldn't even accomplish that much the haircut you'd have to take to prices to balance things out would be substantial. The same spacious boulevards meant to allow residents to walk unimpeded without the significant risk of crowding also permit soldiers to travel freely without breaking ranks. Moreover, even a were a successful blockade of the streets to be enacted, it would only go so far; ensuring damage to roads didn't actually prevent traffic was considered to be a positive factor in planning, to say nothing of the possibility of simply routing through a building or subway line. Absent vastly more time and manpower, it wasn't a remotely feasible possibility.

It's kind of weird how few people there are, though. Obviously it's hardly rare for civilians to flee or hide in the face of a sack, but typically that meant a lot more tightly crowded streets thronged with people attempting to move or carts of hastily gathered valuables, all the more convenient to loot out of if you managed to catch them in time. The city's still big enough that there's more than enough slaves to take back in chains, even if you're being picky, but surely the buildings in question could fit way more than that, and the initial scouting reports certainly agreed with that assumption. Presumably an evacuation given the demeanor of most people they do encounter, but it's oddly successful for one of those if it was really organized on short notice.

...What kind of evacuations are you even used to? Obviously there aren't emergency drills for evacuations from murderous aliens with pointy sticks, but there are still evacuation rehearsals in case of any other kind of unforeseen disaster, and even absent those measures that sort of behavior is obviously foolish. You don't have to be a genius to realize that clogging up the streets in a way that prevents everyone from getting out is just strictly worse than an evacuation where only some people get out. There's still crowding on the subway platforms, but at least it doesn't significantly trade off volume with travel speed there.