Skip to content

AutoStory 2D core: tile navigation package - #360

Merged
dayfine merged 10 commits into
autostory/enginefrom
nav/grid-nav
Nov 13, 2026
Merged

dayfine merged 10 commits into
autostory/enginefrom
nav/grid-nav

Conversation

@dayfine

@dayfine dayfine commented Nov 11, 2026 •

Copy link
Copy Markdown
Owner

Depends on #332 (base autostory/engine; the diff shows only this change).

The AutoStory core is now split so PLZA builds on the engine without the tile-grid code, and this is the first of three 2D core PRs that BDSP and FRLG build on. It brings back #114's navigation package unchanged by reverting 0859830: the leg planner and walker, map data, map tiles and features, route log, runtime map, wall recovery, their tests and test maps, the TestMaps embed, devtools/runtime-map-tests and the runtime-map-reader workflow. Pokemon_Grid stayed in core because the engine backend and PLZA use Direction. The tree at nav/grid-story-steps is the old autostory/engine tip (b6c802c) plus only the SceneJudges test split in #361, which was built and tested as part of that tip.

🤖 Generated with Claude Code

https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss

@dayfine

dayfine commented Nov 12, 2026

Copy link
Copy Markdown
Owner Author

API review (claude-sonnet-5-5) at ea9b142: LGTM_with_nits

No bug or hard-rule findings. This is a revert that restores the navigation package: no Qt, std::filesystem, std::thread or std::mutex, and control-flow formatting is consistent. Remaining items are comment content, positional aggregate init in tests, an unused include, and stale README counts.

  • [style] SerialPrograms/Source/Pokemon/Navigation/Pokemon_LegWalker.h:72 The WatchAlong comment cites a lab run ("FRLG 00.27, console B, 2026-11-09 01:41, the warp to CeruleanCity"). A comment should describe what the screen does, not one run's evidence. Fix: Keep the why: a warp's fade can end before the hold onto the warp ends, so a watch started after the hold misses it. Move the run details to the PR description or commit message.
  • [style] SerialPrograms/Source/Pokemon/Navigation/Pokemon_LegWalker_Tests.cpp:161 The test comment names a console, date and run number ("Console B, 2026-11-01, run 16"), which is lab-run provenance and not explanation of the scenario. Fix: Describe the scenario only: held south into a sign, the leg to a wall is interrupted by its message about 600ms into every hold, repeatedly. Drop the console, date and run number.
  • [nit] SerialPrograms/Source/Pokemon/Navigation/Pokemon_LegTypes.h:5 "Split out of Pokemon_LegWalker.h, unchanged" is change history. The same note appears in Pokemon_MapFeatures.h, Pokemon_MapTiles.h, Pokemon_RouteLogLegs.h and Pokemon_RouteLogTypes.h. Pokemon_RouteLog.h also says "lifted from PokemonBDSP_AutoStory_RouteWalk.cpp". The header will outlive the refactor and the note says nothing about the code. Fix: Replace it with one line saying what the header holds, or remove it, and keep the history in the commit message.
  • [nit] SerialPrograms/Source/Pokemon/Navigation/Pokemon_LegWalker_Tests.cpp:64 LegSpec is built positionally with a bare bool and a bare duration (LegSpec{Direction::EAST, 4, false, TILE}). The same pattern appears for HoldOutcome{0, 2200ms} and in most tests in this file, so a reader cannot tell which field false or true sets. Fix: Use designated initializers, e.g. LegSpec{.direction = Direction::EAST, .tiles = 4, .to_wall = false, .tile_time = TILE}.
  • [nit] SerialPrograms/Source/Pokemon/Navigation/Pokemon_RuntimeMap_Tests.cpp:6 is included in the application build, but std::cout and std::cerr are used only under PA_RUNTIME_MAP_STANDALONE. Fix: Move the include inside the standalone-only section.
  • [nit] devtools/runtime-map-tests/README.md:32 The README is stale against the code. It says "all eight existing registered MapData cases", but MAP_DATA_CASES has 9 after the tiles-to-door case. It says "The 26 new cases", but RUNTIME_MAP_CASES has about 39 entries. It also names the 40717c7c pin while CI uses 906956e9. Fix: Update the counts, or drop them so they cannot drift. State that CI passes --publisher-revision with its own pin.

dayfine and others added 2 commits November 12, 2026 14:59
…ment, README counts

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
The README's prepare_parity.py command had no --publisher-revision, and the
script's default pin was older than the workflow's, so following the README
on the pinned checkout failed. The default now matches the workflow pin.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
@dayfine
dayfine merged commit 2569cc5 into autostory/engine Nov 13, 2026
16 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant