wip: bisect nums instead of ranges

There are fewer ranges than numbers, so that already gives a speedup.
Additionally, we can find multiple fresh products from a single range,
which would otherwise require n searches through the ranges.

Author: Pencilcaseman

src/year2025/day05.rs

@@ -32,22 +32,35 @@ pub fn parse(input: &str) -> Input {
     (merged, nums)
 }
 
+// pub fn part1((ranges, nums): &Input) -> usize {
+//     nums.iter()
+//         .filter(|&&n| {
+//             ranges
+//                 .binary_search_by(|r| {
+//                     if n < r.start {
+//                         std::cmp::Ordering::Greater
+//                     } else if n > r.end {
+//                         std::cmp::Ordering::Less
+//                     } else {
+//                         std::cmp::Ordering::Equal
+//                     }
+//                 })
+//                 .is_ok()
+//         })
+//         .count()
+// }
+
 pub fn part1((ranges, nums): &Input) -> usize {
-    nums.iter()
-        .filter(|&&n| {
-            ranges
-                .binary_search_by(|r| {
-                    if n < r.start {
-                        std::cmp::Ordering::Greater
-                    } else if n > r.end {
-                        std::cmp::Ordering::Less
-                    } else {
-                        std::cmp::Ordering::Equal
-                    }
-                })
-                .is_ok()
+    println!("{} {}", ranges.len(), nums.len());
+    ranges
+        .iter()
+        .map(|r| {
+            let start = nums.binary_search(&r.start).unwrap_or_else(|e| e);
+            let end = nums.binary_search(&r.end).unwrap_or_else(|e| e);
+
+            end - start
         })
-        .count()
+        .sum()
 }
 
 pub fn part2((ranges, _): &Input) -> u64 {