jpmschweitzer/settled-reach
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(client): Atlas screenshot capture matrix + goldens + set_view API (T-1120) #180

Closed
jpmschweitzer wants to merge 0 commits from atlas-screenshot-goldens into main
pull from: atlas-screenshot-goldens
merge into: :main
Conversation 1 Commits - Files Changed -
jpmschweitzer commented 2026-07-16 20:36:21 +02:00
Owner
Copy Link
Copy Source

Summary

T-1120 (T-1037 test infra): programmatic Atlas screenshot capture + the first Atlas golden set. Delivered to Jeroen as 42 captures + contact sheet; 12 curated shots promoted into the existing golden machinery.

What

  • AtlasViewer gains a public set_view(zoom, offset) / get_view_offset() API (clamped; the review-flagged alternative to poking _view_zoom/_view_offset from harnesses) + a gdUnit suite.
  • visual_capture.gd gains an atlas_matrix scenario: ONE Godot boot loops tests/atlas_shots.json (7 DB-verified bodies x zoom {fit,2.0,4.0} x overlay sets A/B/C, ~42 shots in ~90s), with signal-gated waits (atlas_layers_received Ready + city_names_received, 15s timeout fallback). 12 individually-addressable golden scenarios registered in tests/visual.json.
  • 12 golden PNGs in client/tests/golden/visual/ (+.import sidecars force-added per the existing convention), self-consistency 12/12 exact pixel match on this box.
  • Root-level settings.db* gitignored (capture-spawned servers write their settings store to cwd).

Bodies

GJ380c, GJ144e, GJ338Bd, GJ820Bc (Metropolis + Cygni yard visible), GJ251c (deliberate T-1116 water/no-roads exhibit), GJ244Ad (frozen, most cities of its class), GJ445c-m1 (inhabited moon).

Bugs found by building this

  • GDScript lambdas capture value-type locals BY VALUE — the layer-ready wait flag silently never flipped; moved to an Array carrier (team-note-worthy gotcha).
  • RegionalScreen.enter() no-ops when the screen id is unchanged — body #2+ would silently show body #1; harness calls enter() directly per body.
  • Test-infra: three new gdUnit tests initially skipped scene-tree entry (nil _canvas) — caught by the push gate, fixed.

Out of scope / follow-ups

T-1121 filed: all 19 PRE-EXISTING goldens fail 12-14% on this box (cross-machine font/AA/GPU baseline drift; fog_live_hub stalls without xvfb) — portability investigation, unrelated to the new set.

Verification

Full gate green (GDScript parse, client suite 2956 incl. the new suites, JSON). Captures verified by eye by the lead before delivery; goldens are exact-match self-consistent on the capture box.

## Summary T-1120 (T-1037 test infra): programmatic Atlas screenshot capture + the first Atlas golden set. Delivered to Jeroen as 42 captures + contact sheet; 12 curated shots promoted into the existing golden machinery. ### What - AtlasViewer gains a public set_view(zoom, offset) / get_view_offset() API (clamped; the review-flagged alternative to poking _view_zoom/_view_offset from harnesses) + a gdUnit suite. - visual_capture.gd gains an atlas_matrix scenario: ONE Godot boot loops tests/atlas_shots.json (7 DB-verified bodies x zoom {fit,2.0,4.0} x overlay sets A/B/C, ~42 shots in ~90s), with signal-gated waits (atlas_layers_received Ready + city_names_received, 15s timeout fallback). 12 individually-addressable golden scenarios registered in tests/visual.json. - 12 golden PNGs in client/tests/golden/visual/ (+.import sidecars force-added per the existing convention), self-consistency 12/12 exact pixel match on this box. - Root-level settings.db* gitignored (capture-spawned servers write their settings store to cwd). ### Bodies GJ380c, GJ144e, GJ338Bd, GJ820Bc (Metropolis + Cygni yard visible), GJ251c (deliberate T-1116 water/no-roads exhibit), GJ244Ad (frozen, most cities of its class), GJ445c-m1 (inhabited moon). ### Bugs found by building this - GDScript lambdas capture value-type locals BY VALUE — the layer-ready wait flag silently never flipped; moved to an Array carrier (team-note-worthy gotcha). - RegionalScreen.enter() no-ops when the screen id is unchanged — body #2+ would silently show body #1; harness calls enter() directly per body. - Test-infra: three new gdUnit tests initially skipped scene-tree entry (nil _canvas) — caught by the push gate, fixed. ### Out of scope / follow-ups T-1121 filed: all 19 PRE-EXISTING goldens fail 12-14% on this box (cross-machine font/AA/GPU baseline drift; fog_live_hub stalls without xvfb) — portability investigation, unrelated to the new set. ### Verification Full gate green (GDScript parse, client suite 2956 incl. the new suites, JSON). Captures verified by eye by the lead before delivery; goldens are exact-match self-consistent on the capture box.
jpmschweitzer added 4 commits 2026-07-16 20:36:22 +02:00
chore(meta): activate T-1120 — Atlas screenshot capture matrix + goldens (stig) 17cad2bbf9
feat(client): Atlas screenshot capture matrix + set_view API (T-1120) 799156b140 
AtlasViewer gains public set_view(zoom, offset)/get_view_offset (clamped, review-flagged API gap) + gdUnit suite. visual_capture.gd gains an atlas_matrix scenario (one boot, tests/atlas_shots.json-driven: 7 verified bodies x zoom x overlay sets, signal-gated waits on atlas_layers_received/city_names_received with timeout fallback). 12 individually-addressable golden scenarios registered in tests/visual.json. Root-level settings.db* gitignored (capture-spawned servers write their settings store to cwd). Fixes found live: GDScript lambdas capture value-type locals BY VALUE (wait-flag moved to Array carrier); RegionalScreen.enter() no-ops when the screen id is unchanged (direct enter() per body after the first).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test(client): 12 Atlas golden screenshots — 7 bodies x 3 zooms x 3 overlay sets (T-1120) 2b2f4f8abb 
Curated from the 42-shot matrix, one per body spanning all zoom levels and overlay sets; self-consistency verified 12/12 at exact pixel match on this box. .import sidecars force-added per the existing golden convention. Bodies: GJ380c, GJ144e, GJ338Bd, GJ820Bc (Metropolis + Cygni yard), GJ251c (T-1116 water/no-roads exhibit), GJ244Ad (frozen), GJ445c-m1 (moon).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test(client): add_child AtlasViewer in view-api tests — bare .new() has nil _canvas 56b9f135d0 
Three of the four new tests skipped scene-tree entry, so _ready() never resolved _canvas and set_view crashed _apply_transform. Mirrors the fourth (passing) test. Targeted suite green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer added 3 commits 2026-07-16 21:32:49 +02:00
docs(client): golden provenance notes — atlas_* goldens are machine-local until T-1121 (PR #180 review, Tyre T1) 9427b165f0 
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test(client): wall-clock deadline + 10s ceiling for input-roundtrip response waits 8d1121d13a 
Second distinct flake in this suite today (gate run: MoveEast observed mid-move at the 5s window). The old accounting accrued elapsed only on empty polls — load-dependent effective window. Wall-clock deadline via Time.get_ticks_msec; ceiling 5s->10s (exit-on-arrival, costs nothing when healthy). Targeted suite 6/6.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test(client): pace input-roundtrip moves past the D-053 stance cooldown (T-1068 fix ported) 3819e65759 
Root-caused with server trace logs: Walk stance throttles 1 move per 2 ticks and SILENTLY discards early moves (movement.rs apply_move -> try_move, TRACE-only) — under load the test's MoveEast landed exactly 1 tick after MoveNorth and was consumed. Server behavior is correct, deliberate D-053, pinned by Rust tests; production clients re-send while keys are held. Fix: COOLDOWN_TICKS=3 post-ack spacing, same as test_sprint2_proof.gd (T-1068) whose fix never reached this file due to helper copy-paste. Pre-fix repro at run 17/30 under load; post-fix 30/30 stress green + full suite 2956/2956.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer added 1 commit 2026-07-16 21:42:03 +02:00
fix(client): move golden-provenance key out of scenarios{} — crashed visual_capture --list (PR #180 review, Hoshe) fcf913a413 
The _comment_atlas_goldens key inside scenarios{} hit _print_list's .get() on a bare String (reproduced: Invalid call on String at visual_capture.gd:699). Moved to a top-level key alongside resolution/golden_dir; _print_list now also skips _-prefixed and non-Dictionary entries so future comment keys cannot crash it. Verified: --list runs clean, JSON valid.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer commented 2026-07-16 21:42:38 +02:00
Author
Owner
Copy Link
Copy Source

Re-review verdict: APPROVED — Tyre: T1 provenance resolved on both surfaces, commits 2-3 clean (the wall-clock fix repairs a real latent unbounded-timeout bug; the cooldown pacing makes the test respect a genuine D-053 gameplay contract). Hoshe: APPROVE with one condition — the provenance key crashed visual_capture --list (reproduced live) — fixed exactly as prescribed on fcf913a41 (key moved top-level + iterator hardened, repro verified gone, gate green). Root-cause note for the record: the movement flake was NOT a bridge bug — the D-053 Walk stance cooldown silently discards a move landing 1 tick after the previous one; test now paces past it (T-1068 fix ported; T-1122 filed for the shared-harness extraction that would have prevented the drift). Merging.

## Re-review verdict: APPROVED — Tyre: T1 provenance resolved on both surfaces, commits 2-3 clean (the wall-clock fix repairs a real latent unbounded-timeout bug; the cooldown pacing makes the test respect a genuine D-053 gameplay contract). Hoshe: APPROVE with one condition — the provenance key crashed visual_capture --list (reproduced live) — fixed exactly as prescribed on fcf913a41 (key moved top-level + iterator hardened, repro verified gone, gate green). Root-cause note for the record: the movement flake was NOT a bridge bug — the D-053 Walk stance cooldown silently discards a move landing 1 tick after the previous one; test now paces past it (T-1068 fix ported; T-1122 filed for the shared-harness extraction that would have prevented the drift). Merging.
jpmschweitzer closed this pull request 2026-07-16 21:46:42 +02:00

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
jpmschweitzer (Jeroen Schweitzer)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: jpmschweitzer/settled-reach#180