#Algorithmic Techniques

This module contains cross-cutting competitive-programming techniques, including two pointers, sliding windows, Mo's algorithm, binary search on answer, ternary search, meet-in-the-middle, monotonic structures, and interval scheduling.

Packages are organized as one technique per directory. Each package contains a README.mbt.md tutorial with examples that are checked by moon test.

Run from the workspace root:

moon check
moon test
moon info
moon fmt

#Release notes

#0.18.0

Breaking — online_median: the public OnlineMedian::hi heap now stores each element of the larger half as its bitwise complement x.lnot() (= -x - 1) instead of its negation -x. Negation overflowed for Int::MIN_VALUE and corrupted every median query. Code that reads hi directly must decode with .lnot() instead of negation (after adding 1 and 3, hi.peek() is now Some(-4) rather than Some(-3)). The lower_median, upper_median and median methods are unaffected.
Overflow fixes (no API change): challenge_binary_search_answer (can_ship, min_capacity), sliding_window (longest_subarray_with_sum_le, min_subarray_with_sum_ge), challenge_two_pointers, challenge_meet_in_middle.

Site source Report issue Packages Build queue Skills Statistics

loop_invariants_techniques

#Algorithmic Techniques

#Release notes

#0.18.0