Repository navigation
Conversation
01.01 now runs on past the white fade: mash_to_move_hints presses A through the arrival scenes until the controls hint holds a second or the MAIN MISSION 01 banner shows on two looks. It is a scene machine over the shared ScreenMachine (CommonTools/ScreenMachine, with StepTrace), which runs states on ScreenObservations so it works on a console and on a scripted backend alike. The console backend reads two new detectors, the controls hint and the mission banner. 01.01's save is unverified.
# Conflicts: # SerialPrograms/Source/ComputerPrograms/ComputerPrograms_UnitTestRunner.cpp
…rough run_scene_machine match_screen_state, SceneScreen, scene_screen_*() and SceneTimeline were public only for tests. They move into the .cpps' anonymous namespaces. The timeline cases become runs that check the states the trace logs; the state-table cases (terminal / presses A) restated the table and are dropped.
… summed up every 30 s
The scene machine logged two lines per change of state, and 01.01's
arrival flipped between scene_moving and blind_press for 5 minutes
(console B, 2026-11-15): ~200 overlay lines. Now one line per change of
state or readings ("screen N: <state> at T s: <readings>"); looks that
only repeat, or keep two or three states taking turns, are counted and
summed up every 30 s and before the next line.
# Conflicts: # SerialPrograms/Source/VideoGames/PokemonLZA/Programs/AutoStory/PokemonLZA_AutoStoryGame_Tests.cpp
On console B (2026-11-15) 01.01's arrival went ad, loading, a fade from black, the rival's talk, Pancham, the MAIN MISSION 01 banner, a second black fade, then control ~7 s after the banner; on the Italian game the banner detector missed the banner. mash_to_move_hints takes "end_after_black_fades": the scene ends once that many black fades have lifted and the screen has settled 2 s, and the banner no longer ends it; text after the last fade waits for the next. 01.01 sets it to 2. A walk also takes the controls hint for control.
… hint length Both missed on console B 2026-11-15. The banner's icon is looked for across the header's left half (Italian's longer header put it at dx -96, outside the old +-60), and its bar is read at its ends so a long mission name cannot drop it. The controls hint's card sits ~20 px higher under a three-row hint (English and Italian 01.01); it is looked for from 60 px above to 20 px below. CI pins dayfine/CommandLineTests 2abde09: the English and Italian frames, and ArrivalPastMissionBanner 0198-0221 named with the hint they show.
The user's rule: the two black fades are the signal, the banner shows between them (logged, never required), and control is the controls hint or 5 s after the second fade, whichever comes first.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AKYzhrMnPK61XHWEch4a6E
Owner
Author
|
Superseded by #644 (one lean branch; split for review later) |
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 #577
Base
plza/slice1/e; the diff shows only this slice (2a).01.01 now runs the whole checkpoint: after the walk into the white fade,
mash_to_move_hintspresses A through the arrival scenes (title card, flyover, the tourist question's top row, the ad, Pancham) and ends on the third black fade (into the ad, out of the ad, after the MAIN MISSION 01 banner): the player is taken as in control on the controls hint or 5 s after that fade, whichever comes first, and the banner is only logged. Both detectors search a range rather than one spot, so they hold for other languages and hint line counts (checked on English and Italian frames). Its save isunverified. The scene machine runs on a new sharedCommonTools/ScreenMachine(states overScreenObservations, plusStepTrace), so it runs the same on the console and onFakeStoryBackend; the console backend gains the controls hint and mission banner detectors. Cut from the p4-02 version: the choice/battle goals, B mashes, mission-complete card, autosave-quiet end, interact prompt and party bar as control, the banner press-away, and the frame-based machine and screenshot trace.Tested with new suites for the screen machine (match, run, stall and limit failures, trace), the scene machine (timeline sequences, states, scripted runs) and 01.01 through the engine, plus the two detectors on synthetic frames and the CLT #18 arrival folders. Compiled locally (syntax only); Fork Draft CI builds and runs the tests.