fix(engine): transform "all <type>" as a mass effect, no target (Moonmist) - #6629
Conversation
…mist)
Moonmist's "Transform all Humans" forced a single-target prompt to pick one
Human: the parser emitted single-target Effect::Transform{target} and the
effect exposed a target slot, so a non-targeting mass effect was resolved as
a targeted one (only the source transformed).
Scope-parameterize Transform to carry the single-vs-mass axis, mirroring the
Tap/TapAll -> SetTapState{scope} refactor (no new sibling variant):
Effect::Transform { target } -> Transform { target, scope: EffectScope }
- EffectScope::Single (serde default): legacy targeted transform, unchanged.
- EffectScope::All: a non-targeting population filter enumerated at
resolution. Effect::target_filter() returns None for All (no target slot /
prompt built), mirroring SetTapState/DestroyAll (CR 115.10 — "all X" doesn't
target). resolve_all iterates the battlefield population, pre-filtering to
double-faced permanents so a matched single-faced creature is a harmless
no-op instead of an error (CR 701.27c).
Parser gains a "transform/convert all|each <filter>" mass branch (before the
single branch); the census/loop-safety classifiers scope-split Transform to
LiveBoardCensus/Census for All (mirroring DestroyAll), and every
SetTapState{Single}-gated site gets the parallel Transform gate. Serde default
keeps old Transform{target} deserializing to Single. Frontend unchanged.
Closes phase-rs#6403
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Warning Review limit reached
Next review available in: 36 minutes Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughTransform effects now carry explicit ChangesTransform scope support
Estimated code review effort: 4 (Complex) | ~45 minutes Sequence Diagram(s)sequenceDiagram
participant OracleParser
participant EffectLowering
participant TransformResolver
participant Battlefield
participant GameEvent
OracleParser->>EffectLowering: produce Transform with Single or All scope
EffectLowering->>TransformResolver: resolve scoped Transform
TransformResolver->>Battlefield: evaluate mass filter when scope is All
Battlefield-->>TransformResolver: matching permanents
TransformResolver->>GameEvent: emit EffectResolved after transformations
Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/engine/src/parser/oracle_trigger_tests.rs`:
- Line 5391: Update the Effect::Transform match in the relevant parser test to
destructure scope separately while retaining .. for other payload fields, then
assert that the parsed anaphoric “transform it” effect has EffectScope::Single.
In `@crates/mtgish-import/src/convert/action.rs`:
- Around line 3821-3822: Update the CR citation in the TransformPermanent
conversion block to reference CR 701.28/701.28a, keeping EffectScope::Single
unchanged.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 086549cc-2810-47af-9a7a-d2df8bac0bbe
📒 Files selected for processing (14)
crates/engine/src/game/ability_rw.rscrates/engine/src/game/ability_scan.rscrates/engine/src/game/ability_utils.rscrates/engine/src/game/coverage.rscrates/engine/src/game/effects/mod.rscrates/engine/src/game/effects/token.rscrates/engine/src/game/effects/transform_effect.rscrates/engine/src/parser/oracle_effect/imperative.rscrates/engine/src/parser/oracle_effect/tests.rscrates/engine/src/parser/oracle_ir/ast.rscrates/engine/src/parser/oracle_tests.rscrates/engine/src/parser/oracle_trigger_tests.rscrates/engine/src/types/ability.rscrates/mtgish-import/src/convert/action.rs
| // CR 701.27a: mtgish `TransformPermanent` is a single targeted transform. | ||
| scope: EffectScope::Single, |
There was a problem hiding this comment.
🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
rg -n -A8 -B2 '^701\.28' docs/MagicCompRules.txtRepository: phase-rs/phase
Length of output: 216
🏁 Script executed:
#!/bin/bash
set -euo pipefail
printf 'Repo root: '; pwd
# Find the rules document and the target source file.
fd -a 'MagicCompRules.txt' .
fd -a 'action.rs' crates/mtgish-import/src/convert
# Inspect the relevant area around the cited annotation.
if [ -f crates/mtgish-import/src/convert/action.rs ]; then
sed -n '3808,3830p' crates/mtgish-import/src/convert/action.rs | cat -n
fi
# Search the rules document for transform-related sections if present.
rules_doc="$(fd -a 'MagicCompRules.txt' . | head -n 1 || true)"
if [ -n "$rules_doc" ]; then
rg -n -A6 -B2 '701\.(27|28)' "$rules_doc"
fiRepository: phase-rs/phase
Length of output: 1722
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Locate the rules document anywhere in the repo.
fd -a 'MagicCompRules.txt' /home/jailuser/git || true
fd -a 'MagicCompRules.txt' . || true
# If found, show the transform sections.
rules_doc="$(fd -a 'MagicCompRules.txt' /home/jailuser/git | head -n 1 || true)"
if [ -n "$rules_doc" ]; then
echo "RULES_DOC=$rules_doc"
rg -n -A8 -B2 '701\.(27|28)' "$rules_doc"
fi
# Also inspect nearby comments in the source for the exact wording/context.
sed -n '3806,3826p' /home/jailuser/git/crates/mtgish-import/src/convert/action.rs | cat -nRepository: phase-rs/phase
Length of output: 1525
Use CR 701.28/701.28a here. TransformPermanent is transform, so both citations should point to CR 701.28 instead of 701.27/701.27a.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@crates/mtgish-import/src/convert/action.rs` around lines 3821 - 3822, Update
the CR citation in the TransformPermanent conversion block to reference CR
701.28/701.28a, keeping EffectScope::Single unchanged.
Source: Path instructions
Parse changes introduced by this PR · 4 card(s), 6 signature(s) (baseline: main
|
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/engine/tests/integration/issue_6403_moonmist_mass_transform.rs`:
- Around line 67-97: Add a single-faced Human creature without BackFaceData to
the Moonmist scenario, then assert via its identifier that it remains
untransformed after casting. Keep the existing transformed Human assertions,
non-Human assertion, and outcome.final_waiting_for() WaitingFor::Priority check
unchanged so the matching non-transformable permanent is validated through the
cast path.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: ca9d48d8-8299-4e82-8378-80403f3b9e32
📒 Files selected for processing (3)
crates/engine/src/parser/oracle_trigger_tests.rscrates/engine/tests/integration/issue_6403_moonmist_mass_transform.rscrates/engine/tests/integration/main.rs
🚧 Files skipped from review as they are similar to previous changes (1)
- crates/engine/src/parser/oracle_trigger_tests.rs
matthewevans
left a comment
matthewevans
left a comment
There was a problem hiding this comment.
[HIGH] Mass transform drops the printed on their front face restriction for That's No Moonmist. Evidence: imperative.rs:5334 discards the remainder from parse_target_with_ctx and immediately emits EffectScope::All; the current parse-diff therefore records That's No Moonmist only as artifact or creature Phyrexian. Its current Oracle text is: “Transform all artifacts and Phyrexian creatures on their front face.” resolve_all then transforms every matched DFC, including a matching back-face permanent that the card must leave unchanged.
Please either model the front-face condition as a typed population predicate and add a regression that keeps a matching transformed permanent unchanged, or strict-fail/preserve Effect::unimplemented whenever this mass-transform branch leaves that semantic suffix unparsed. The latter is acceptable if the predicate is intentionally deferred; silently marking the card supported is not.
matthewevans
left a comment
matthewevans
left a comment
There was a problem hiding this comment.
[HIGH] Mass transform still silently drops the printed on their front face restriction for That's No Moonmist. imperative.rs:5334 discards the parsed suffix as (target, _) before emitting EffectScope::All; the PR's published parse-diff consequently records only the artifact or creature Phyrexian filter. The card's Oracle text also restricts the population to objects on their front face.
resolve_all currently filters only by that target filter and double-faced status, so it can transform a matching DFC already on its back face, although that object must be left unchanged.
The prior requested change at 6400ed502b78d239d407b4006c853957da5cc522 was not remediated by the merge-only current-head change. Please carry the front-face restriction as a typed population predicate and add a production cast-path regression containing a matching back-face permanent; otherwise strict-fail this rules-bearing suffix until it is represented.
…s unparsed (phase-rs#6403 review) Address matthewevans review on phase-rs#6629: the mass "transform all|each <filter>" branch discarded the parse_target_with_ctx remainder, silently dropping the "on their front face" restriction on That's No Moonmist ("Transform all artifacts and Phyrexian creatures on their front face") — resolve_all could then transform a matching back-face permanent CR 712.2 must leave unchanged. The branch now captures the remainder: an empty remainder (Moonmist's fully consumed "all Humans") still emits EffectScope::All; a surviving rules-bearing suffix returns None so the line falls through to Effect::unimplemented (the single authority) rather than shipping a silently-lossy Transform — and does not fall through to the single "transform <target>" branch. Strict-failing the suffix-bearing cards also means resolve_all only runs on fully-parsed populations like "all Humans" (where the Human typing already implies front face), so the mass path can never transform a matching back-face permanent. Also thread the new Effect::Transform.scope field through two merge-drift sites that main added after the branch point: cr733_resolved_transform.rs (scope: Single) and the maintainer-added regression's BackFaceData (printed_loyalty: None). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Thanks — fixed at head Front-face restriction (That's No Moonmist). Took the strict-fail option you offered: there's no The
This also closes the Regressions (both pass): CI (was red). The auto-merge of |
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (3)
crates/engine/tests/integration/cr733_resolved_transform.rs (1)
111-125: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick winAssert the replayed transformation count.
The test validates the journaled count against the live object, but replay only checks face state and timestamp.
apply_resolved_transformcould restore an incorrect count while this test still passes, despite claiming replay exactness. Add an assertion thatreplayed.transformation_countequalstransform.resulting_transformation_count.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/engine/tests/integration/cr733_resolved_transform.rs` around lines 111 - 125, Extend the replay assertions in the integration test after applying apply_resolved_transform to verify that replayed.transformation_count equals transform.resulting_transformation_count. Keep the existing face, name, and timestamp assertions unchanged and use the same descriptive assertion style.crates/engine/src/game/ability_rw.rs (2)
8181-8199: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick winExercise the ordering decision, not only the profile fields.
This test calls
rw_quantity_refandability_rw_profile, but never reachesprofiles_conflictor the production trigger-ordering path. A regression in howreads_event_liveaffects conflict classification would therefore still pass. Add a positive event-live-read × event-object-write case and controls for missing event objects or missing writes through the ordering predicate.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/engine/src/game/ability_rw.rs` around lines 8181 - 8199, Extend triggering_scry_look_count_classifies_event_live to exercise the trigger-ordering predicate, not just reads_event_live metadata: add a positive conflict case combining an event-live read with an event-object write, plus controls where the event object or write is absent. Route these cases through profiles_conflict or the production trigger-ordering path and assert the expected classifications, preserving the existing profile and twin-parity checks.Source: Path instructions
2953-2961: 📐 Maintainability & Code Quality | 🟠 Major | ⚡ Quick winAdd CR citations for both
Effect::ExileFaceDownPilearms. They encode the face-down exile pile move and theHandLibrary/membership writes, but neither branch has aCR <number>: <description>annotation.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/engine/src/game/ability_rw.rs` around lines 2953 - 2961, Add CR citations to both Effect::ExileFaceDownPile match arms, including descriptions covering the face-down exile-pile move and the HandLibrary/membership writes. Place the annotations directly on the relevant branches and preserve the existing legacy_quantity_expr, legacy_target_filter, and return behavior.Source: Path instructions
♻️ Duplicate comments (1)
crates/engine/tests/integration/issue_6403_moonmist_mass_transform.rs (1)
54-102: 🎯 Functional Correctness | 🟠 MajorCover a matching single-faced Human.
Both Humans receive
BackFaceData, so this test never exercises the CR 701.27c no-op for a matching but non-transformable permanent. Add a Human withoutBackFaceData, assert it remains untransformed, and retain the existing prompt-free resolution assertion.As per path instructions, regression tests must cover the failure path through the production pipeline.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/engine/tests/integration/issue_6403_moonmist_mass_transform.rs` around lines 54 - 102, Add a third Human permanent without BackFaceData in moonmist_cast_transforms_all_humans_without_target_selection, and assert it remains untransformed after resolution while the existing transformable Humans and non-Human assertions remain intact. Keep the test using runner.cast(moonmist).resolve() and retain the WaitingFor::Priority assertion to cover the production pipeline without target selection.Source: Path instructions
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/engine/tests/integration/issue_6403_moonmist_mass_transform.rs`:
- Around line 119-141: The test
thats_no_moonmist_front_face_restriction_stays_unimplemented must add a positive
reach guard for the exact “all artifacts and Phyrexian creatures” population
without the “on their front face” suffix, asserting it parses as
Effect::Transform with EffectScope::All before checking the suffix variant is
Unimplemented. Keep the existing Moonmist control and ensure the negative
assertion verifies the unsupported restriction rather than total population
rejection.
---
Outside diff comments:
In `@crates/engine/src/game/ability_rw.rs`:
- Around line 8181-8199: Extend triggering_scry_look_count_classifies_event_live
to exercise the trigger-ordering predicate, not just reads_event_live metadata:
add a positive conflict case combining an event-live read with an event-object
write, plus controls where the event object or write is absent. Route these
cases through profiles_conflict or the production trigger-ordering path and
assert the expected classifications, preserving the existing profile and
twin-parity checks.
- Around line 2953-2961: Add CR citations to both Effect::ExileFaceDownPile
match arms, including descriptions covering the face-down exile-pile move and
the HandLibrary/membership writes. Place the annotations directly on the
relevant branches and preserve the existing legacy_quantity_expr,
legacy_target_filter, and return behavior.
In `@crates/engine/tests/integration/cr733_resolved_transform.rs`:
- Around line 111-125: Extend the replay assertions in the integration test
after applying apply_resolved_transform to verify that
replayed.transformation_count equals transform.resulting_transformation_count.
Keep the existing face, name, and timestamp assertions unchanged and use the
same descriptive assertion style.
---
Duplicate comments:
In `@crates/engine/tests/integration/issue_6403_moonmist_mass_transform.rs`:
- Around line 54-102: Add a third Human permanent without BackFaceData in
moonmist_cast_transforms_all_humans_without_target_selection, and assert it
remains untransformed after resolution while the existing transformable Humans
and non-Human assertions remain intact. Keep the test using
runner.cast(moonmist).resolve() and retain the WaitingFor::Priority assertion to
cover the production pipeline without target selection.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 186bcb2c-639d-4490-9110-4b1c605885e9
📒 Files selected for processing (13)
crates/engine/src/game/ability_rw.rscrates/engine/src/game/ability_scan.rscrates/engine/src/game/ability_utils.rscrates/engine/src/game/coverage.rscrates/engine/src/game/effects/mod.rscrates/engine/src/game/effects/token.rscrates/engine/src/game/effects/transform_effect.rscrates/engine/src/parser/oracle_effect/imperative.rscrates/engine/src/parser/oracle_effect/tests.rscrates/engine/src/types/ability.rscrates/engine/tests/integration/cr733_resolved_transform.rscrates/engine/tests/integration/issue_6403_moonmist_mass_transform.rscrates/engine/tests/integration/main.rs
🚧 Files skipped from review as they are similar to previous changes (7)
- crates/engine/tests/integration/main.rs
- crates/engine/src/parser/oracle_effect/tests.rs
- crates/engine/src/game/effects/mod.rs
- crates/engine/src/game/coverage.rs
- crates/engine/src/game/effects/token.rs
- crates/engine/src/parser/oracle_effect/imperative.rs
- crates/engine/src/game/effects/transform_effect.rs
|
Current head |
matthewevans
left a comment
matthewevans
left a comment
There was a problem hiding this comment.
Approved: current-head regression coverage now exercises the matching non-transformable Human and the strict-failed face-state suffix; all required checks are green.
2 participants
Tier: Frontier
Model: claude-opus-4-8
Closes #6403.
Summary
Moonmist — "Transform all Humans. Prevent all combat damage that would be dealt this turn by creatures other than Werewolves and Wolves." — forced a single-target prompt to pick one Human. The parser emitted a single-target
Effect::Transform { target: Typed{Subtype Human} }, and the effect exposed a target slot, so a non-targeting mass effect was resolved as a targeted one (only the source transformed). "Transform all X" doesn't target (CR 115.10) — it should transform every Human on the battlefield with no prompt.Rather than add a
TransformAllsibling, this scope-parameterizesTransformto carry the single-vs-mass axis — the exact refactor the codebase already applied toTap/TapAll→SetTapState { scope }(theEffectScopedoc-comment names it), so/add-engine-variantapproves it as parameterization, not proliferation:EffectScope::Single(serde default): the legacy targeted/anaphoric transform ("transform target creature","transform ~"), unchanged.EffectScope::All: a non-targeting population filter enumerated at resolution.Effect::target_filter()returnsNoneforAll(so no target slot / prompt is built — mirroringSetTapState/DestroyAll), and a newresolve_alliterates the battlefield population applying the transform to each match, mirroringtap_untap::resolve_all.Two rules-correctness details that make the mass path actually work:
target_filter()is the single target-exposure authority. Both production slot-builder paths route through it, soNoneforAllfully removes the prompt (CR 115.10).resolve_allpre-filters the matched population to double-faced permanents (is_double_faced_permanent) so a matched single-faced creature is skipped rather than erroring —transform_permanentreturnsErr("no back face")for a non-DFC, so the mass loop must not propagate it (a naive mirror would abort the whole transform on the first single-faced Human).The parser gains a
"transform/convert all|each <filter>"mass branch (before the single branch), the loop-safety census classifiers scope-splitTransformto a live-board census forAll(mirroringDestroyAll— a wrong classification would defeat the LoopFirewall), and everySetTapState{Single}-gated site gets the parallelTransformgate. The serde default keeps old serializedTransform{target}deserializing toSingle(byte-compatible).Implementation method (required)
Method: /engine-implementer — plan (/engine-planner + /add-engine-effect, with a run-then-delete reproduction probe confirming the single-target prompt) + the /add-engine-variant gate (verdict: scope-parameterize per the
SetTapStateprecedent; corrected the CR citations — transform is 701.27, not 701.28) → /review-engine-plan (found 5 blocking gaps: the plan compiled but did not fix the bug — thetarget_filter()linchpin was untouched,transform_permanenterrors rather than no-ops on non-DFCs, the census classifiers were left scope-blind, ≥5 compile sites and theSetTapState-gate matrix were incomplete — all corrected) → implement → /review-impl (Maintainer-Simulation Gate PASS; independently confirmed all 5 blockers correct —target_filter()is the only target-exposure surface, census classification correct, Single-gate parity complete; one unreachable LOW noted below). Each step in a fresh agent context.Gate A
./scripts/check-parser-combinators.sh→ Gate A PASS head=1613e601a (Gate G PASS). The mass parser branch is composed combinators (preceded/alt/tag/peek).Anchored on
Effect::SetTapState { target, scope, state }(types/ability.rs, theTap/TapAllunification) — the end-to-end template: theEffectScopeaxis, thetarget_filter()scope-split (Single => Some(target),All => None), the census-classifier scope arms, and theSingle-gate sites are all mirrored forTransform.game/effects/tap_untap.rs::resolve_all— the mass-resolution shape (resolved_object_filter+matches_target_filteroverstate.battlefield, per-object, no targets) thattransform_effect::resolve_allmirrors, with the DFC pre-filter added for CR 701.27c.game/ability_scan.rsDestroyAll/Suspect{All}census arms — the correct loop-safety classification (LiveBoardCensus/Census) for a mass effect that iterates the growing battlefield;Transform{All}joins that group (explicitly not theSetTapState-onlySnapshotOrEventexception, which stays pinned byobligation_ii_census_exception_is_exactly_settapstate).Verification
cargo fmt --allclean;cargo build -p engine -p mtgish-import -p phase-aiclean (every scope site threaded);cargo clippy -p engine --lib --tests -- -D warnings0 warnings;cargo test -p engine --lib17,706 passed / 0 failed (no drift from the sharedEffectchange, incl. the census/LoopFirewall tests); Gate A/G PASS.parse_oracle_text+ real resolver,back_faceDFC modeling):moonmist_transforms_all_humans_without_a_prompt(primary): parse Moonmist verbatim, resolve over 2 DFC Humans + a Goblin + a Werewolf → both Humans transform, Goblin/Werewolf untouched, noWaitingForprompt,target_filter().is_none(), PreventDamage sibling preserved.mass_transform_exposes_no_target_slot(B1):Transform{All}.target_filter() == None,Single→Some(target).mass_transform_skips_single_faced_human_without_error(B2): a single-faced Human (back_face:None) in the population → the DFC transforms, the single-faced Human is untouched, resolution does not error (removing the DFC pre-filter makes it error — revert-sensitive).parse_transform_all_is_mass_scope/parse_transform_target_is_single_scope: the mass branch emitsscope:All;"transform target creature"staysSingle(control — mass branch didn't regress single).docs/MagicCompRules.txt: 701.27 / 701.27a (transform), 701.27c (a permanent that can't transform does nothing), 115.10 / 115.10a ("all X" is non-targeting).Claimed parse impact
Serialized
Effect::Transformgains ascopefield: existing Transform cards mechanically gain"scope":"Single"(serde default → byte-compatible deserialization of old states), and Moonmist becomes"scope":"All". The parser mass branch fires only on literal"transform/convert all|each <filter>", so no other card's Transform parse changes shape.mtgish-importthreadsscope:Singleon its existingTransformPermanentaction (a forced compile-consequence of the shared reshape). Frontend unchanged (no clientEffect::Transformhandling). The exact card-data parse-diff (which cards carry a "transform all" line beyond Moonmist) is deferred to CI'scoverage-parse-diffartifact — the scope parameter is the general building block for the whole mass-transform class.Scope Expansion
None. Honest latent parity note (out of class, not a regression): the tracked-set-recording arm that special-cases
SetTapState{All}for chained "those cards" sub-abilities has noTransform{All}sibling — unreachable today (Moonmist is aTypedpopulation filter and no tracked-set mass-transform card exists;Transform{Single}had no such handling either). I left it rather than add a speculative guard; aTransform{All}arm can be added if such a card ever appears.🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Bug Fixes
Tests