Rectangle Fit Queries

Process the rows of operations from left to right. Each row has one of two forms:
[0, a, b]: create and save a rectangle of size a × b.
[1, a, b]: determine whether every rectangle saved by earlier operations can fit inside a box of size a × b.
Test each saved rectangle separately; the rectangles do not need to fit in the box at the same time. You may rotate a rectangle by 90 degrees.
Return one boolean for each query operation, in query order.

Function
solution(operations: int[][]) → boolean[]

Examples
Example 1
operations = [[1,1,1]]
return = [true]
No rectangles have been saved, so every saved rectangle vacuously fits and the answer is true.
Example 2
operations = [[0,1,3],[0,4,2],[1,3,4],[1,3,2]]
return = [true,false]
Both saved rectangles fit the 3 × 4 box after choosing the appropriate orientation. The 4 × 2 rectangle cannot fit the later 3 × 2 box, so the answers are [true,false].

Constraints
operations contains at least one row.
operations[i].length = 3.
operations[i][0] is either 0 or 1.
operations[i][1] and operations[i][2] are positive integers.

Normalize every rectangle so its short side is first and its long side is second. Keep two values: maxShort, the largest short side seen, and maxLong, the largest long side seen. For a query, normalize the box the same way into boxShort and boxLong. Every rectangle fits if and only if maxShort <= boxShort and maxLong <= boxLong. That works because each rectangle is checked on its own, and sorted sides are the best orientation for a fit. With no saved rectangles, initialize both maxes to 0 so the answer is true. The common pitfall is storing all rectangles and rescanning on each query, which is slow. Another is forgetting to normalize the box, so a 4 x 3 box fails wrongly. Each operation is O(1). If your head goes blank mid-OA, StealthCoder can hand you this solution from the screen without the proctor seeing it.

StealthCoder is the hedge for the one pattern you didn't drill. It runs invisibly during the screen share.