Skip to content

feat: Exp3 for adversarial linear bandits with the Kiefer–Wolfowitz exploration, for features that need not span the space - #459

Open
RemyDegenne wants to merge 5 commits into
LeanMachineLearning:mainfrom
RemyDegenne:roadmap/linear-exp3-kiefer-wolfowitz-worker2
Open

RemyDegenne wants to merge 5 commits into
LeanMachineLearning:mainfrom
RemyDegenne:roadmap/linear-exp3-kiefer-wolfowitz-worker2

Conversation

@RemyDegenne

@RemyDegenne RemyDegenne commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

This PR completes Bandits Layer 10.1, Exp3 with exploration: "prove [LS, Theorem 27.1] … hence $R_n\le2\sqrt{3dn\log k}$ with the Kiefer–Wolfowitz design of Layer 8.6". It also covers the case of the completion check "The Kiefer–Wolfowitz theorem applies to the finite action sets of Layers 9.3 and 10.1, including sets that do not span $E$".

On main, LinearExp3.integral_regret_le_two_mul_sqrt_three takes the Kiefer–Wolfowitz properties of the exploration as hypotheses: hπu : IsUnit V(π) and hg : ⟪φ b, V(π)⁻¹ φ b⟫ ≤ d. No caller supplies them, and they cannot hold when the features do not span E. This PR supplies the design and removes both hypotheses. The features may be arbitrary, spanning E or not. After this PR, nothing of Layer 10.1 remains. The rest of Layer 10 (item 10.2) is already on main.

The construction runs Exp3 for linear bandits on the span W of the features, so it never needs V(π) to be invertible on E. The learner only observes ⟪φ c, y⟫, which equals ⟪φ c, P_W y⟫, so a run against an adaptive adversary in E is a run against the projected adversary in W. The bound then holds with m = dim W ≤ d.

Generic Moore–Penrose material, in LMLExplore/Analysis/InnerProductSpace/MoorePenrose.lean (namespace ContinuousLinearMap, and LinearMap.IsSymmetric for the two lemmas with a symmetry hypothesis, over any RCLike field): operator-level restriction. For T : E →L[𝕜] E acting on a subspace K as an invertible T_K : K →L[𝕜] K (hT : ∀ v : K, T v = T_K v):

  • apply_coe_ringInverse_of_apply_coe_eq: T ↑(T_K⁻¹ v) = v;
  • le_range_of_apply_coe_eq: K ≤ range T;
  • LinearMap.IsSymmetric.moorePenroseInverse_apply_coe_of_apply_coe_eq (dot notation on hsym): for symmetric T on finite-dimensional E, T⁺ c = T_K⁻¹ c for c : K;
  • LinearMap.IsSymmetric.inner_moorePenroseInverse_of_apply_coe_eq: its inner-product form, ⟪a, T⁺ c⟫ = ⟪a, T_K⁻¹ c⟫ for a c : K.

The existing inner_moorePenroseInverse_designOperator_map_subtype in OptimalDesign.lean is now a one-line corollary of the last one (statement unchanged).

Design material, in LMLExplore/Analysis/InnerProductSpace/OptimalDesign.lean (namespace LMLExplore):

  • isSymmetric_sum_smul_rankOne: ∑ i ∈ s, w i • φ i ⊗ φ i is symmetric for arbitrary real weights (structural proof from Mathlib's LinearMap.isSymmetric_sum, IsSymmetric.smul, isSymmetric_rankOne_self).
  • spanFeatures φ: the features, as elements of Submodule.span ℝ (Set.range φ), with coe_spanFeatures (simp), span_range_spanFeatures (via Mathlib's Submodule.span_span_coe_preimage) and inner_spanFeatures_orthogonalProjectionOnto.
  • exists_mem_simplex_inner_ringInverse_spanFeatures_le_finrank: Kiefer–Wolfowitz for finitely many arms on the span of features that need not span E [LS, Exercise 21.6], from the existing exists_mem_simplex_inner_ringInverse_le_finrank applied on the span.
  • kieferWolfowitzDesign φ, its choice, with kieferWolfowitzDesign_mem_simplex, isUnit_sum_smul_rankOne_kieferWolfowitzDesign and inner_ringInverse_sum_smul_rankOne_kieferWolfowitzDesign_le_finrank (G-value at most m).
  • inner_ringInverse_sum_smul_rankOne_spanFeatures: when V_W(p) is invertible, ⟪φ a, V_W(p)⁻¹ φ c⟫ = ⟪φ a, V(p)⁺ φ c⟫, with the Moore–Penrose inverse on E (a corollary of LinearMap.IsSymmetric.inner_moorePenroseInverse_of_apply_coe_eq). So the algorithm on the span is Algorithm 15 with V⁺ in place of V⁻¹.

Bandit material, new file LMLExplore/Bandits/LinearBandit/Exp3/KieferWolfowitz.lean (namespace LMLExplore.Bandits):

  • linearExp3Estimate_spanFeatures: for spanning features, the loss estimates on the span are those on E.
  • linearExp3KieferWolfowitz φ η γ hγ, which is linearExp3 (spanFeatures φ) (kieferWolfowitzDesign φ) η γ _ hγ, with linearExp3KieferWolfowitz_def. linearExp3KieferWolfowitz_eq_linearExp3 proves that, for spanning features and γ > 0, it equals linearExp3 φ (kieferWolfowitzDesign φ) η γ _ hγ on E.
  • LinearExp3.integral_regret_linearExp3KieferWolfowitz_le_two_mul_sqrt_three: against an adaptive adversary, with η = √(log k / (3 m n)) and γ = η m, the expected regret against every arm is at most 2 √(3 m n log k). There is no hypothesis on the features or the design. When all features vanish (m = 0), the regret is 0.
  • LinearExp3.integral_regret_linearExp3KieferWolfowitz_le_two_mul_sqrt_three_finrank: the same regret is at most 2 √(3 d n log k), which is the roadmap's statement.
  • LinearExp3.integral_regret_detFeedback_linearExp3KieferWolfowitz_le_two_mul_sqrt_three: the first bound for every run against Environment.detFeedback obs (fun t x ↦ ⟪φ x.2, y t x.1⟫) _.

No public declaration of main is removed or renamed.

Reuse: I grepped Mathlib, Tau Ceti, LML, Statlib and LMLExplore/ for a restriction of a family to its span (codRestrict, rangeFactorization, maps into Submodule.span ℝ (Set.range _)) and found none. DesignSelection.kieferWolfowitz is the phased-elimination design selection on finsets with the Moore–Penrose inverse in E, a different interface. The proofs consume existing results:

  • Kiefer–Wolfowitz and its finite-arm form: exists_mem_simplex_inner_ringInverse_le_finrank.
  • Exp3 for linear bandits: integral_regret_le_two_mul_sqrt_three, isUnit_sum_smul_rankOne_linearExp3Weights.
  • Moore–Penrose: IsSymmetric.moorePenroseInverse_apply_apply_of_mem_range, moorePenroseInverse_eq_ringInverse.
  • Mathlib: Submodule.inner_orthogonalProjectionOnto_eq_of_mem_left, LinearMap.isUnit_iff_range_eq_top, LinearMap.isSymmetric_sum.

The source is cited in the files: Lattimore and Szepesvári, Bandit Algorithms, Theorem 27.1, Theorem 21.1 and Exercise 21.6.

Hypotheses:

  • [Nonempty α] is needed to choose the design.
  • 0 < γ in linearExp3KieferWolfowitz_eq_linearExp3 keeps the exploration weights invertible at every round. The tuned rates have γ > 0.
  • The rate, boundedness and measurability hypotheses of the regret theorems are those of the existing integral_regret_le_two_mul_sqrt_three.

Local checks only (CI verifies the pushed PR):

  • scope: scoped: against 20105e47ae0239a41ff62aef39ff39df15cda49c
  • local-axioms: checked 3 changed module(s), 135 declaration(s); 0 violation(s)
  • The whole-library lake build --iofail (9081 jobs), the module-system audit, the header check and the environment lint (135 declarations, 14 environment linters) passed, and the verify script exited 0.

Roadmap: Bandits

🤖 Generated with Claude Code

…xploration, for features that need not span the space

Co-Authored-By: Claude Code <noreply@github.com>
@lmlexplore-automation lmlexplore-automation Bot added awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts and removed awaiting-CI CI has not yet reported on the latest commit labels Nov 15, 2026

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 reuse — request_changes claude/claude-opus-5-5

The Moore–Penrose bridge inner_ringInverse_sum_smul_rankOne_spanFeatures and its helpers span_le_range_sum_smul_rankOne and coe_sum_smul_rankOne_spanFeatures_apply re-derive the existing inner_moorePenroseInverse_designOperator_map_subtype and designOperator_map_subtype_apply in OptimalDesign.lean with the same proof steps. The rest of the file correctly reuses the existing Kiefer–Wolfowitz and Exp3 results.

  • LMLExplore/Bandits/LinearBandit/Exp3/KieferWolfowitz.lean:152 — inner_ringInverse_sum_smul_rankOne_spanFeatures proves the same fact as inner_moorePenroseInverse_designOperator_map_subtype (OptimalDesign.lean:611): ⟪a, V⁺ c⟫ on E equals ⟪a, V_K⁻¹ c⟫ on a subspace K where V_K is invertible. The proof follows the same steps: write c as V (V_K⁻¹ c), show K lies in the range of V, then apply IsSymmetric.moorePenroseInverse_apply_apply_of_mem_range. The new lemma differs only in allowing two different vectors and arbitrary weights p. Fix: Generalise inner_moorePenroseInverse_designOperator_map_subtype in OptimalDesign.lean to two vectors a c : K, at operator level (a symmetric T on E that maps K into K, with an invertible restriction T_K, gives ⟪a, T⁺ c⟫ = ⟪a, T_K⁻¹ c⟫). Derive both the existing lemma and the Exp3 statement from it.
  • LMLExplore/Bandits/LinearBandit/Exp3/KieferWolfowitz.lean:129 — coe_sum_smul_rankOne_spanFeatures_apply restates designOperator_map_subtype_apply (OptimalDesign.lean:325) for the sum form, and span_le_range_sum_smul_rankOne (line 137) restates the hmem step inside inner_moorePenroseInverse_designOperator_map_subtype. They are the same duplicate as the previous finding. Fix: Put the generalised subspace-restriction lemma, with its range-membership step, in OptimalDesign.lean and use it here instead of these two public helpers. Remove them if nothing else uses them.

Reply in this thread to contest a finding; that re-runs only this rubric and posts an answer here. (To fix it, just push a commit — that re-reviews on its own. To contest again after an answer, post a NEW reply rather than editing an old one.)

claude/claude-opus-5-5 · 50s · 182.7k in (140.1k cache read, 42.6k cache write) / 4.4k out tokens · reviewing this diff · rubric · ⚠️ rubrics checkout differs from published main

@RemyDegenne RemyDegenne Nov 15, 2026 •

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 naming — request_changes claude/claude-opus-5-5

Two new Moore–Penrose restriction theorems take hsym : LinearMap.IsSymmetric as their first explicit argument but are placed in ContinuousLinearMap, which breaks the IsSymmetric. dot-notation precedent in the same file. All other names and namespaces are fine.

  • LMLExplore/Analysis/InnerProductSpace/MoorePenrose.lean:179 — ContinuousLinearMap.moorePenroseInverse_apply_coe_of_apply_coe_eq takes hsym : (T : E →ₗ[𝕜] E).IsSymmetric as its first explicit argument but is not in the LinearMap.IsSymmetric namespace, so hsym.… dot notation does not work. Fix: Rename it to LinearMap.IsSymmetric.moorePenroseInverse_apply_coe_of_apply_coe_eq, next to the existing IsSymmetric lemmas, and update its callers.
  • LMLExplore/Analysis/InnerProductSpace/MoorePenrose.lean:190 — Same problem for ContinuousLinearMap.inner_moorePenroseInverse_of_apply_coe_eq: its first explicit argument is hsym : LinearMap.IsSymmetric. Fix: Rename it to LinearMap.IsSymmetric.inner_moorePenroseInverse_of_apply_coe_eq and update its uses in OptimalDesign.lean (inner_moorePenroseInverse_designOperator_map_subtype, inner_ringInverse_sum_smul_rankOne_spanFeatures) to hsym.inner_moorePenroseInverse_of_apply_coe_eq ….

Reply in this thread to contest a finding; that re-runs only this rubric and posts an answer here. (To fix it, just push a commit — that re-reviews on its own. To contest again after an answer, post a NEW reply rather than editing an old one.)

claude/claude-opus-5-5 · 28s · 83.5k in (66.5k cache read, 17.1k cache write) / 2.9k out tokens · reviewing this diff · rubric · ⚠️ rubrics checkout differs from published main

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed in 1efdf60: both theorems now live in namespace LinearMap.IsSymmetric, right after the ContinuousLinearMap section of MoorePenrose.lean. They are LinearMap.IsSymmetric.moorePenroseInverse_apply_coe_of_apply_coe_eq and LinearMap.IsSymmetric.inner_moorePenroseInverse_of_apply_coe_eq, and their statements are unchanged. The callers use dot notation on hsym: the inner-product form calls hsym.moorePenroseInverse_apply_coe_of_apply_coe_eq, and OptimalDesign.lean calls (isPositive_designOperator _).toLinearMap.isSymmetric.inner_moorePenroseInverse_of_apply_coe_eq … and (isSymmetric_sum_smul_rankOne _ φ p).inner_moorePenroseInverse_of_apply_coe_eq …. apply_coe_ringInverse_of_apply_coe_eq and le_range_of_apply_coe_eq stay in ContinuousLinearMap, because they take no symmetry hypothesis. I updated the module docstring and the PR description. Local checks pass: whole-library lake build --iofail (9081 jobs), and local-axioms: checked 3 changed module(s), 135 declaration(s); 0 violation(s).

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 proof-quality — request_changes claude/claude-opus-5-5

The proofs are short and readable, but one inline symmetry proof re-derives a structural fact by unfolding inner products instead of using existing Mathlib lemmas, and one rewrite chain is duplicated across two proofs instead of being factored into a lemma.

  • LMLExplore/Bandits/LinearBandit/Exp3/KieferWolfowitz.lean:158 — hsym proves symmetry of ∑ b, p b • rankOne ℝ (φ b) (φ b) by unfolding into a long simp only [coe_coe, sum_apply, smul_apply, rankOne_apply, sum_inner, inner_sum, real_inner_smul_left, real_inner_smul_right] plus sum_congr/real_inner_comm/ring. This chain depends on many lemma names, and it hides a reusable fact: the design operator is symmetric for arbitrary weights. Fix: Factor out a lemma isSymmetric_sum_smul_rankOne (next to isPositive_sum_smul_rankOne in OptimalDesign.lean). Prove it structurally with coe_sum + LinearMap.isSymmetric_sum + IsSymmetric.smul + isSymmetric_rankOne_self, and use it here.
  • LMLExplore/Bandits/LinearBandit/Exp3/KieferWolfowitz.lean:167 — The rewrite chain rw [coe_coe, ← coe_sum_smul_rankOne_spanFeatures_apply, apply_ringInverse_apply hu, …] appears twice: in span_le_range_sum_smul_rankOne (lines 143–144) and in hc (lines 167–168). Both prove the same fact, that V(p) ↑(V_W(p)⁻¹ w) = ↑w. Fix: Add a lemma sum_smul_rankOne_apply_coe_ringInverse (hu) (w) : V(p) ↑(Ring.inverse V_W(p) w) = ↑w. Use it in both places: span_le_range_sum_smul_rankOne becomes fun v hv ↦ ⟨_, lemma hu ⟨v, hv⟩⟩, and hc becomes (lemma hu _).symm.

Reply in this thread to contest a finding; that re-runs only this rubric and posts an answer here. (To fix it, just push a commit — that re-reviews on its own. To contest again after an answer, post a NEW reply rather than editing an old one.)

claude/claude-opus-5-5 · 42s · 143.3k in (129.2k cache read, 14.1k cache write) / 3.3k out tokens · reviewing this diff · rubric · ⚠️ rubrics checkout differs from published main

@RemyDegenne

RemyDegenne commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor Author

AI review — changes requested

Each rubric is judged independently by multiple review agents; the PR merges only once every rubric is green — any rubric that is not green (changes requested, blocked, errored, stale, or not yet run) blocks the merge. See the rubrics.

⚠️ This review ran from a rubrics checkout that differs from the published rubrics (out of date, pinned, or locally edited), so its findings may reflect different rules.

rubric state judge summary
✅ correctness approved claude/claude-opus-5-5 The new regret bounds faithfully state LS Theorem 27.1 with a Kiefer–Wolfowitz design on the feature span: the design hypotheses are proved, not assumed, and the regret is stated for the original adversary in E. The Moore–Penrose restriction lemmas, the spanning-case equivalence and the m = 0 case are correct; I found no vacuity or mis-formalization.
✅ reuse approved claude/claude-opus-5-5 The new declarations (operator restriction for the Moore–Penrose inverse, spanFeatures, the Kiefer–Wolfowitz design on the span, and the Exp3 regret bound with no spanning hypothesis) have no existing counterpart: DesignSelection.kieferWolfowitz in PhasedElimination/Basic.lean is a finset design-selection interface requiring Nontrivial E. The PR rebuilds inner_moorePenroseInverse_designOperator_map_subtype on the new generic lemma and reuses exists_mem_simplex_inner_ringInverse_le_finrank, Submodule.span_span_coe_preimage, LinearMap.isSymmetric_sum, Submodule.inner_orthogonalProjectionOnto_eq_of_mem_left and integral_regret_le_two_mul_sqrt_three.
✅ scope approved claude/claude-opus-5-5 The PR proves Bandits Layer 10.1's own statement (LS Thm 27.1 with the Kiefer–Wolfowitz design, R_n ≤ 2√(3dn log k)) and the roadmap's completion check for finite action sets that do not span E (README.md:738-741, 766-767). The Moore–Penrose restriction lemmas and span-feature design are direct prerequisites, and the re-proof of the existing subtype lemma comes from the same generalisation, so this is a single topic.
✅ attribution approved claude/claude-opus-5-5 The new bandit file and the new Kiefer–Wolfowitz design declarations cite Lattimore–Szepesvári (Theorem 27.1, Theorem 21.1, Exercise 21.6) in the code, both files have References sections, and the citations match the roadmap's. The Moore–Penrose restriction lemmas are routine and need no credit.
✅ api-design approved claude/claude-opus-5-5 The general operator-restriction lemmas now sit in MoorePenrose.lean under ContinuousLinearMap; spanFeatures, kieferWolfowitzDesign and linearExp3KieferWolfowitz each have the lemmas that characterize them (simp-tagged coercion, _def by (rfl), simplex/invertibility/G-value lemmas from the choice). No bodies are exposed and there is no compatibility layer.
✅ generality approved claude/claude-opus-5-5 The prior finding is resolved: moorePenroseInverse_apply_coe_of_apply_coe_eq now states the vector identity T⁺ c = T_K⁻¹ c, and the inner-product form and the design corollaries are derived from it. The remaining statements sit at the natural level (general RCLike field for the restriction lemmas, regret bound in the span dimension m with the d bound derived from it).
✅ placement approved claude/claude-opus-5-5 The generic restriction lemmas live in MoorePenrose.lean, the span-of-features and Kiefer–Wolfowitz design material sits next to the existing finite-arm Kiefer–Wolfowitz results in OptimalDesign.lean, and the bandit results are in a new file in the existing LinearBandit/Exp3/ directory. Both imports of the new file are specific and used.
🟡 naming changes requested claude/claude-opus-5-5 Two new Moore–Penrose restriction theorems take hsym : LinearMap.IsSymmetric as their first explicit argument but are placed in ContinuousLinearMap, which breaks the IsSymmetric. dot-notation precedent in the same file. All other names and namespaces are fine.
✅ documentation approved claude/claude-opus-5-5 The new module, its definitions and its main theorems are documented accurately with [LS] citations, and no docstring claims more than its statement proves. The headers of the touched files list the new material, and the related Exp3 docstrings are still accurate.
✅ proof-quality approved claude/claude-opus-5-5 Proofs are short, built from named lemmas, and use no change/show or hidden definitional equality across wrappers. The subspace argument for the Moore–Penrose inverse is factored into reusable lemmas, and the regret theorems are brief reductions to the existing bound through the projection lemma.

♻️ = approved on an earlier commit, re-run before merge.

Reviewing this diff at head c469622. rubrics @ 7d59664. CLI @ 7d59664. Review spend: $2.98.

@lmlexplore-automation lmlexplore-automation Bot added awaiting-author A review requested changes; author action needed and removed awaiting-review CI is green; waiting for review verdicts labels Nov 15, 2026
…Design

Move `spanFeatures` and its lemmas, and the Kiefer–Wolfowitz design on the
span of the features (now `kieferWolfowitzDesign`, with the existential
`exists_mem_simplex_inner_ringInverse_spanFeatures_le_finrank`), to
`Analysis/InnerProductSpace/OptimalDesign.lean`. Factor the Moore–Penrose
restriction argument into the operator-level
`inner_moorePenroseInverse_of_apply_coe_eq` (with
`apply_coe_ringInverse_of_apply_coe_eq` and `le_range_of_apply_coe_eq`),
derive `inner_moorePenroseInverse_designOperator_map_subtype` and
`inner_ringInverse_sum_smul_rankOne_spanFeatures` from it, and add
`isSymmetric_sum_smul_rankOne`. Rename the regret theorems after their bound.

Co-Authored-By: Claude Code <noreply@github.com>
@lmlexplore-automation lmlexplore-automation Bot added awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts awaiting-author A review requested changes; author action needed and removed awaiting-author A review requested changes; author action needed awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts labels Nov 15, 2026
`apply_coe_ringInverse_of_apply_coe_eq`, `le_range_of_apply_coe_eq` and
`inner_moorePenroseInverse_of_apply_coe_eq` are generic facts about an operator
acting on a subspace and its Moore–Penrose inverse, with no design content. Move
them from OptimalDesign.lean to MoorePenrose.lean, in the `ContinuousLinearMap`
namespace next to `moorePenroseInverse_eq_ringInverse`, over a general `RCLike`
field.

Co-Authored-By: Claude Code <noreply@github.com>
@lmlexplore-automation lmlexplore-automation Bot added awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts and removed awaiting-author A review requested changes; author action needed awaiting-CI CI has not yet reported on the latest commit labels Nov 15, 2026
@lmlexplore-automation lmlexplore-automation Bot added awaiting-author A review requested changes; author action needed and removed awaiting-review CI is green; waiting for review verdicts labels Nov 15, 2026
Add ContinuousLinearMap.moorePenroseInverse_apply_coe_of_apply_coe_eq,
T⁺ c = T_K⁻¹ c for c ∈ K, and derive the inner-product form from it.

Co-Authored-By: Claude Code <noreply@github.com>
@lmlexplore-automation lmlexplore-automation Bot added awaiting-CI CI has not yet reported on the latest commit and removed awaiting-author A review requested changes; author action needed labels Nov 15, 2026
@lmlexplore-automation lmlexplore-automation Bot added awaiting-review CI is green; waiting for review verdicts awaiting-author A review requested changes; author action needed and removed awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts labels Nov 15, 2026
…rMap.IsSymmetric

Co-Authored-By: Claude Code <noreply@github.com>
@lmlexplore-automation lmlexplore-automation Bot added awaiting-CI CI has not yet reported on the latest commit awaiting-review CI is green; waiting for review verdicts and removed awaiting-author A review requested changes; author action needed awaiting-CI CI has not yet reported on the latest commit labels Nov 15, 2026

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

awaiting-review CI is green; waiting for review verdicts

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant