Repository navigation
Conversation
…battle engine StoryData/frlg-league-plan.json is PokemonSimulation #115's plan.json unchanged. The loader reads its new rows: a foe's sequence (Lapras: Strength, Bite, Bite) and heal block (Machamp, Dragonite: below 80 HP a heal of at least 50), kill_order, between.heal "one", the item caps, the reset rules (Roar/Whirlwind; Bruno's early reset after Onix #1) and the pass figures. The planned policy ports #115's rows (foe_heal, sequence, kill_order off with a sequence), the one-heal top-up between fights and the spare-only Super Potion count; league_early_reset() and league_try_passes() are the final plan's reset and accept rules. The parity dump is regenerated for #115 (PokemonSimulation #116) with our attacks per foe slot; every mismatch prints its decision line. A fake battle engine feeds PlannedBattle the readers' per-turn inputs from a scripted foe side and checks ten League scenarios.
PokemonFRLG_AutoStoryLeaguePlanData.h had 11 public names, over code health's 8: the foe's plan (SlugCounts, its setup goals and heal block) moves to PokemonFRLG_AutoStoryLeagueFoePlan.h, and parse_league_plan() into the .cpp, which alone calls it.
…e Full Heals Parity: the HP box shows BRN and FRZ as one icon, and the program took it for a burn, so a frozen Blastoise went to the cure row where the sim's sleep_cure takes it (chain 7, Lorelei turn 7). BattleTrack now keeps our side's last "frozen solid" or "was burned" text, and the icon is a freeze after the former. The Dragonair fake battle started at +0 Sp. Atk, where Blizzard's low rolls leave Dragonair in Lance's Full Restore range and kill_order rightly opens with Strength; it now carries the plan's +2 from Gyarados. The approved shopping test still had 2 Full Heals at 20.01a; the plan buys 4 (user, 2026-11-16).
Owner
Author
|
Superseded by #652 (frlg/lean), which continues from this branch's tip; the branch stays. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #613 (based on frlg/check-levels, not #618: the League plan C++ lives only on check-levels, and #618 hasn't merged its 20.0x plan rewrite yet). Data from dayfine/PokemonSimulation#115 and dayfine/PokemonSimulation#116.
The C++ League policy was a port of the sim's #111 plan, so it was missing what the final plan (#115) added: foe move sequences, per-foe heal blocks, the one-heal rule between fights, item caps, and the reset and pass rules. This ships #115's plan.json unchanged as StoryData, extends the loader, ports those rows into the planned policy and its between-fight top-up, and adds league_early_reset() / league_try_passes(). The parity dump is regenerated against #115 (now with our attack count per foe slot), and every mismatch is printed with its decision line. A test-side fake battle engine feeds PlannedBattle the per-turn inputs the readers give it and plays ten League scenarios. The touched files compile locally; this branch has no headless test target and the app isn't run locally, so Fork Draft CI runs the parity and fake-battle tests.