Skip to content

BDSP AutoStory: routes stop short of warps; a dim cave floor is not a tap's fade - #269

Merged
dayfine merged 4 commits into
bdsp/autostory-mine-floaroma-wipfrom
bdsp/autostory-warp-guard
Nov 9, 2026
Merged

dayfine merged 4 commits into
bdsp/autostory-mine-floaroma-wipfrom
bdsp/autostory-warp-guard

Conversation

@dayfine

@dayfine dayfine commented Nov 9, 2026 •

Copy link
Copy Markdown
Owner

Base: bdsp/autostory-mine-floaroma-wip (#226); the diff shows only this change. Sibling of #262, which moves the same sim code; expect a small conflict in StorySim_Tests.cpp when both land.

On lab-B (2026-11-08, four runs) 01.05's open walk to the tile beside the Oreburgh Gate's mouth slid onto the warp, and the tap step after it read the Gate's dim floor as its fade and waited for a banner already gone. This fixes both halves in shared code rather than in that one row.

  • Table rule. A new Story table test plans every ROUTE step as the walker does and fails on an open leg that may end on the tile before a door it walks toward. It flagged 01.05's old by_the_gate and four more, now fixed: 00.03 above_the_door_again (counted legs from the west), 01.02 by_the_east_mouth and 01.05 through_the_gate (two tiles short of the Gate's mouths; the taps go the rest), 01.03 above_the_mine (two rows above the warp row; the door run goes on). 01.02's by_the_gate was not flagged. I kept this in the table, not the planner: when the target is itself beside a warp (row 58 above the mine is all beside warps), the planner can't fix it.
  • Real fades only. tap_faded(): a look after a tap counts as the fade only when it is black (< 25), or when it drops to dark from the look before the tap. A banner showing before the first tap ends the step with no tap ("through already"). A tap step now stops two taps past the door its map has ahead, so a step that starts inside the Gate with its banner gone fails a few tiles in instead of walking across the Gate.
  • Permanent vs temporary. The table rule, the route fixes, the two-taps-past-the-door bound and the sim tests are meant to stay. Two parts are stopgaps until the event-perception work (autostory: event-driven perception design note (P1) #268) lands, marked TODO(#268 ...) in StoryStepIO.cpp: the brightness-drop fade test in tap_faded() (slice 3, a uniform-screen fade watcher) and the "through already" banner check in tap_to_banner() (slice 2, walkers report the warp they saw). Remove each once its replacement passes the same tests.
  • The story sim's old steps moved unchanged to Simulation/PokemonBDSP_AutoStory_OldStorySteps.* (StorySim_Tests.cpp was at the 1000-line limit); review commit by commit, git diff --color-moved=dimmed-zebra c19b41111^ c19b41111.

Testing

  • Red first: the table test failed on 95b2960^ (5 legs, including 01.05 by_the_gate) and on 95b2960 (4 legs), and passes now. tap_faded failed on the Gate's floor (46 → 46) under the old rule. The two story-sim cases for a tap step that starts through the Gate's mouth (banner up, banner gone) failed with the old walker: it took 24 taps across the Gate. All of these pass now.
  • New GridSim test: a 0.375-tile offset west of (5, 13) makes the old walk's three taps west take the Gate's warp.
  • SerialProgramsTests --filter PokemonBDSP::AutoStory: 103 passed. 2 failed, and both fail on the base too: Grid sim "01.03 oreburgh_north…" and "01.03 above_the_mine… across seeds" (Unknown exception). Nothing ran on the console.

🤖 Generated with Claude Code

https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss

dayfine and others added 4 commits November 8, 2026 16:19
01.05's open walk west to (2, 13), beside the Oreburgh Gate's mouth, slid
onto the warp (lab-B 2026-11-08, four runs). A table test plans every
ROUTE step as the walker does and fails on an open leg that may end on
the tile before a door it walks toward. It flagged 01.05's old
by_the_gate and four more: 00.03 above_the_door_again (now counted legs
from the west), 01.02 by_the_east_mouth and 01.05 through_the_gate (two
tiles short of the Gate's mouths, the taps go the rest), and 01.03
above_the_mine (two rows above the warp row, the door run goes on).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
Moved unchanged out of StorySim_Tests.cpp, which is at the 1000-line
limit: r75's and r81's old cave exits and the old mine steps. The next
commit adds 01.05's old tap into the Oreburgh Gate beside them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
On lab-B (2026-11-08 15:31 to 15:54) 01.05's walk slid into the Oreburgh
Gate, and its tap step read the Gate's floor (about 46) as the fade and
waited 8 s for a banner already gone. A look after a tap is the fade only
when black (< 25) or a drop to dark from the look before the tap
(tap_faded()). A banner before the first tap means the walk before took
us through: the step ends with no tap. A tap step taps at most two past
the door its map has ahead, so one begun through the warp with its banner
gone fails a few tiles in, not across the Gate.

Tests: tap_faded() on the Gate's floor and real fades; the story sim's
01.05 tap step begun through the Gate's mouth, banner up and gone (red
before: 24 taps across the Gate); GridSim's part-tile offset taking the
warp on the old open walk west.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
…s stopgaps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
@dayfine
dayfine marked this pull request as ready for review November 9, 2026 00:45
@dayfine
dayfine merged commit 013c4bd into bdsp/autostory-mine-floaroma-wip Nov 9, 2026
16 of 23 checks passed
@dayfine
dayfine deleted the bdsp/autostory-warp-guard branch November 9, 2026 00:50
dayfine added a commit that referenced this pull request Nov 9, 2026
…warps

#269 moved the route's target to (46, 57); the authored plan follows it
in the merge, and its doc now says so too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss
dayfine added a commit that referenced this pull request Nov 9, 2026
… line (#264)

* Make existing BDSP recovery tests compile across GCC and Clang-cl

* BDSP: load mine/Roark checkpoint from shared authored JSON (#263)

* Load BDSP mine checkpoint from shared JSON without changing its executor

* Pin the BDSP replay schema correction for paired CI

* Compare normalized checkpoint semantics instead of optional JSON defaults

* Share authored checkpoint envelopes and embedded resource loading

* Use shared authored checkpoint loading for BDSP

* Make the native mine plan authoritative and address data review

* Space BDSP data declarations and test mutation helpers

* Resolve simulator staging root across Python versions

* Keep existing Git normalization rules for authored story data

* Report actual BDSP story data values in test failures

* BDSP AutoStory: the mine doc's above_the_mine stops two short of the warps

#269 moved the route's target to (46, 57); the authored plan follows it
in the merge, and its doc now says so too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013VUubhHeYSMGSSDXZkKZss

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
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