Route Pattern Matcher

All strings in routes are registered in the order given. A route is a slash-separated path. A segment is either a literal string or a named parameter written as {name}. A parameter segment matches exactly one non-empty URL segment.
For each string in urls, find the first registered route that has the same number of segments and whose literal segments match exactly. If no route matches, return an empty row.
For a match, return one row containing the matched route first, followed by alternating parameter names and values in their left-to-right order: [route, name1, value1, name2, value2,...]. A literal route therefore produces a one-element row.
Return one row for every URL, in input order.

Function
matchRoutes(routes: String[], urls: String[]) → String[][]

Examples
Example 1
routes = ["/users/{id}/pictures/{pictureId}","/health"]
urls = ["/users/101/pictures/1","/health","/users/1"]
return = [["/users/{id}/pictures/{pictureId}","id","101","pictureId","1"],["/health"],[]]
The first URL matches the parameterized route and extracts two values. The second matches the literal route. The final URL has too few segments, so it has no match.
Example 2
routes = ["/files/{name}","/{kind}/{id}"]
urls = ["/files/report","/orders/7"]
return = [["/files/{name}","name","report"],["/{kind}/{id}","kind","orders","id","7"]]
The first registered matching route wins. The second URL does not match the literal files segment, so it uses the fully parameterized route.

Constraints
1 <= routes.length, urls.length <= 10000
The combined length of all route and URL strings is at most 300000.
Every route and URL starts with /, contains no empty segment, and has no trailing slash.
Each parameter occupies an entire segment, uses a unique lowercase name within its route, and is written as {name}.
Literal matching is case-sensitive.

The core is simple. Split each route and URL by slash into segments. A route matches a URL if segment counts are equal and every literal segment is identical, case-sensitive. Parameter segments like {name} accept any non-empty segment and get captured in order. First registered route wins, so you can't reorder or sort. The pitfall is cost. Checking every URL against every route is up to 10^8 pairs on paper, though the 300000 total length cap helps. The safer move is a trie keyed by segment, with a literal child map and a param child, storing the earliest route index at each terminal. Since the first match by registration order wins, track the minimum index across branches, not just the first path found. Precompute parameter names per route. If the trie logic gets tangled live, StealthCoder is the hedge that can hand you a working version while you stay calm.

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