Lean Projects
Collection
Digesting Lean Data
•
2 items
•
Updated
Version
stringclasses 20
values | section
stringclasses 275
values | text
stringlengths 14
1.01k
| pull request
stringlengths 3
15
|
|---|---|---|---|
v4.15.0 | Language | implements `Simp.Config.implicitDefEqsProofs`. When `true` (default: `true`), `simp` will **not** create a proof term for a rewriting rule associated with an `rfl`-theorem. Rewriting rules are provided by users by annotating theorems with the attribute `@[simp]`. If the proof of the theorem is just `rfl` (reflexivity), and `implicitDefEqProofs := true`, `simp` will **not** create a proof term which is an application of the annotated theorem. | 4595 |
v4.15.0 | Language | avoid negative environment lookup | 5429 |
v4.15.0 | Language | ensure `instantiateMVarsProfiling` adds a trace node | 5501 |
v4.15.0 | Language | adds a feature to the the mutual def elaborator where the `instance` command yields theorems instead of definitions when the class is a `Prop`. | 5856 |
v4.15.0 | Language | unset trailing for `simpa?` "try this" suggestion | 5907 |
v4.15.0 | Language | changes the rule for which projections become instances. Before, all parents along with all indirect ancestors that were represented as subobject fields would have their projections become instances. Now only projections for direct parents become instances. | 5920 |
v4.15.0 | Language | make `all_goals` admit goals on failure | 5934 |
v4.15.0 | Language | introduce synthetic atoms in bv_decide | 5942 |
v4.15.0 | Language | adds a new definition `Message.kind` which returns the top-level tag of a message. This is serialized as the new field `kind` in `SerialMessaege` so that i can be used by external consumers (e.g., Lake) to identify messages via `lean --json`. | 5945 |
v4.15.0 | Language | `arg` conv tactic misreported number of arguments on error | 5968 |
v4.15.0 | Language | BitVec.twoPow in bv_decide | 5979 |
v4.15.0 | Language | simplifies the implementation of `omega`. | 5991 |
v4.15.0 | Language | fix style in bv_decide normalizer | 5992 |
v4.15.0 | Language | adds configuration options for `decide`/`decide!`/`native_decide` and refactors the tactics to be frontends to the same backend. Adds a `+revert` option that cleans up the local context and reverts all local variables the goal depends on, along with indirect propositional hypotheses. Makes `native_decide` fail at elaboration time on failure without sacrificing performance (the decision procedure is still evaluated just once). Now `native_decide` supports universe polymorphism. | 5999 |
v4.15.0 | Language | changes `bv_decide`'s configuration from lots of `set_option` to an elaborated config like `simp` or `omega`. The notable exception is `sat.solver` which is still a `set_option` such that users can configure a custom SAT solver globally for an entire project or file. Additionally it introduces the ability to set `maxSteps` for the simp preprocessing run through the new config. | 6010 |
v4.15.0 | Language | improves the validation of new syntactic tokens. Previously, the validation code had inconsistencies: some atoms would be accepted only if they had a leading space as a pretty printer hint. Additionally, atoms with internal whitespace are no longer allowed. | 6012 |
v4.15.0 | Language | removes the `decide!` tactic in favor of `decide +kernel` (breaking change). | 6016 |
v4.15.0 | Language | removes @[specilize] from `MkBinding.mkBinding`, which is a function that cannot be specialized (as none of its arguments are functions). As a result, the specializable function `Nat.foldRevM.loop` doesn't get specialized, which leads to worse performing code. | 6019 |
v4.15.0 | Language | makes the `change` tactic and conv tactic use the same elaboration strategy. It works uniformly for both the target and local hypotheses. Now `change` can assign metavariables, for example:
```lean
example (x y z : Nat) : x + y = z := by
change ?a = _
let w := ?a
-- now `w : Nat := x + y`
``` | 6022 |
v4.15.0 | Language | fixes a bug where the monad lift coercion elaborator would partially unify expressions even if they were not monads. This could be taken advantage of to propagate information that could help elaboration make progress, for example the first `change` worked because the monad lift coercion elaborator was unifying `@Eq _ _` with `@Eq (Nat × Nat) p`. As such, this is a breaking change; you may need to adjust expressions to include additional implicit arguments. | 6024 |
v4.15.0 | Language | adds a normalization rule to `bv_normalize` (which is used by `bv_decide`) that converts `x / 2^k` into `x >>> k` under suitable conditions, allowing cheaper shifting circuits instead of expensive division circuits. | 6029 |
v4.15.0 | Language | fixes `simp only [· ∈ ·]` after #5020. | 6030 |
v4.15.0 | Language | introduces the and flattening pre processing pass from Bitwuzla to `bv_decide`. It splits hypotheses of the form `(a && b) = true` into `a = true` and `b = true` which synergizes with the existing embedded constraint substitution pass. | 6035 |
v4.15.0 | Language | fixes `bv_decide`'s embedded constraint substitution to generate correct counter examples in the corner case where duplicate theorems are in the local context. | 6037 |
v4.15.0 | Language | add `LEAN_ALWAYS_INLINE` to some functions | 6045 |
v4.15.0 | Language | fixes `simp?` suggesting output with invalid indentation | 6048 |
v4.15.0 | Language | mark `Meta.Context.config` as private | 6051 |
v4.15.0 | Language | fixes the caching infrastructure for `whnf` and `isDefEq`, ensuring the cache accounts for all relevant configuration flags. Cleans up `WHNF.lean` and improves `whnf` configuration. | 6053 |
v4.15.0 | Language | adds a simp_arith benchmark. | 6061 |
v4.15.0 | Language | optimize Nat.Linear.Expr.toPoly | 6062 |
v4.15.0 | Language | optimize Nat.Linear.Poly.norm | 6064 |
v4.15.0 | Language | improves the asymptotic performance of `simp_arith` when there are many variables to consider. | 6068 |
v4.15.0 | Language | adds options to `bv_decide`'s configuration so that all non mandatory preprocessing passes can be disabled. | 6077 |
v4.15.0 | Language | changes how the canonicalizer handles `forall` and `lambda`, replacing bvars with temporary fvars. Fixes a bug reported by @hrmacbeth on zulip about quantifiers in CanonM. | 6082 |
v4.15.0 | Language | use mkFreshUserName in ArgsPacker | 6093 |
v4.15.0 | Language | improves the `#print` command for structures to show all fields and which parents the fields were inherited from, hiding internal details such as which parents are represented as subobjects. The pretty printer for private constants is also improved. Private names from the current module are handled like other names; private names from other modules are hygienic. | 6096 |
v4.15.0 | Language | modifies the signature of `Nat.fold`, `Nat.foldRev`, `Nat.any`, `Nat.all`, so that the function is passed the upper bound, enabling compile time checks instead of runtime checks. | 6139 |
v4.15.0 | Language | adds a primitive for creating temporary directories, akin to the existing functionality for creating temporary files. | 6148 |
v4.15.0 | Language | completes the elementwise accessors for `ofNatLt`, `allOnes`, and `not` by adding their implementations of `getMsbD`. | 6149 |
v4.15.0 | Language | completes the `toInt` interface for `BitVec` bitwise operations. | 6151 |
v4.15.0 | Language | implements `BitVec.toInt_abs`. | 6154 |
v4.15.0 | Language | adds `toNat` theorems for `BitVec.signExtend.` | 6155 |
v4.15.0 | Language | adds toInt theorems for BitVec.signExtend. | 6157 |
v4.15.0 | Language | adds theorem `mod_eq_sub`, makes theorem `sub_mul_eq_mod_of_lt_of_le` not private anymore and moves it within the `rotate*` section to use it in other proofs. | 6160 |
v4.15.0 | Language | uses `Array.findFinIdx?` in preference to `Array.findIdx?` where possible, converting runtime checks to compile time checks. | 6184 |
v4.15.0 | Language | completes the `toNat` theorems for bitwise operations (`and`, `or`, `xor`, `shiftLeft`, `shiftRight`) of the UInt types and adds `toBitVec` theorems as well; renames `and_toNat` to `toNat_and`. | 6188 |
v4.15.0 | Language | adds the builtin simproc `USize.reduceToNat` which reduces `USize.toNat` on literals less than `UInt32.size`. | 6190 |
v4.15.0 | Language | adds `Array.zipWithAll` and lemmas relating it to `List.zipWithAll`. | 6191 |
v4.15.0 | Language | adds deprecations for `Lean.HashMap` functions which did not receive deprecation attributes initially. | 6192 |
v4.15.0 | Language | completes the TODO in `Init.Data.Array.BinSearch`, removing `partial` and converting runtime checks to compile time checks. | 6193 |
v4.15.0 | Language | changes the signature of `Array.swap`, so it takes `Nat` with tactic-provided proofs of bounds. Renames `Array.swap!` to `Array.swapIfInBounds`. | 6194 |
v4.15.0 | Language | renames `Array.setD` to `Array.setIfInBounds`. | 6195 |
v4.15.0 | Language | upstreams the definition of `Vector` from Batteries, along with basic functions. | 6197 |
v4.15.0 | Language | upstreams `Nat.lt_pow_self` and `Nat.lt_two_pow` from Mathlib and uses them to prove the simp theorem `Nat.mod_two_pow`. | 6200 |
v4.15.0 | Language | makes `USize.toUInt64` a regular non-opaque definition. | 6202 |
v4.15.0 | Language | adds theorems `le_usize_size` and `usize_size_le`, simplifying inequalities about `USize.size`. | 6203 |
v4.15.0 | Language | upstreams some UInt theorems from Batteries and adds more `toNat`-related theorems. Also adds missing `UInt8` and `UInt16` to/from `USize` conversions. | 6205 |
v4.15.0 | Language | ensures `Fin.foldl` and `Fin.foldr` are semireducible so that certain defeqs (like `example (f : Fin 3 → ℕ) : List.ofFn f = [f 0, f 1, f 2] := rfl`) succeed. | 6207 |
v4.15.0 | Language | fix Vector.indexOf? | 6208 |
v4.15.0 | Language | adds `simp` lemmas about `List`'s `==` operation. | 6217 |
v4.15.0 | Language | fixes problems in non-Ubuntu Linux distributions for `tzdata` directory by providing a fallback, and trims unnecessary characters from local time identifier. | 6221 |
v4.15.0 | Language | changes the definition of `HashSet.insertMany` so it is equivalent to repeated `HashSet.insert`. Also clarifies docstrings of `insert` and `insertMany`. | 6222 |
v4.15.0 | Language | copies some lemmas about `List.foldX` to `Array`. | 6230 |
v4.15.0 | Language | upstreams lemmas about `Vector` from Batteries. | 6233 |
v4.15.0 | Language | upstreams the definition and lemmas about `List.finRange` from Batteries. | 6234 |
v4.15.0 | Language | relates `Nat.fold`/`foldRev`/`any`/`all` to corresponding list operations over `List.finRange`. | 6235 |
v4.15.0 | Language | refactors `Array.qsort` to remove runtime array checks and avoid `partial`. Uses `Vector` API with auto_params. Benchmarks show no regression. | 6241 |
v4.15.0 | Language | deprecates `Fin.ofNat` in favor of `Fin.ofNat'` which takes `[NeZero]` instead of returning `Fin (n+1)`. | 6242 |
v4.15.0 | Language | adds `numBits_pos`, `le_numBits`, `numBits_le` to simplify inequalities about `System.Platform.numBits`. | 6247 |
v4.15.0 | Compiler | changes `lean_sharecommon_{eq,hash}` to only consider the salient bytes of an object, not uninitialized capacity bytes. | 5840 |
v4.15.0 | Compiler | fixes a bug in the constant folding of `Nat.ble` and `Nat.blt` in the old code generator, preventing miscompilation. | 6087 |
v4.15.0 | Compiler | should make Lean better-behaved around sanitizers, addressing https://github.com/google/sanitizers/issues/1688. Taking the address of a local does not produce a monotonic measure of stack usage. | 6143 |
v4.15.0 | Compiler | fixes an issue where edits to a command containing a nested docstring fail to reparse the entire command. | 6236 |
v4.15.0 | Compiler | fixes `trace.profiler.pp` not using the term pretty printer. | 6138 |
v4.15.0 | Compiler | adds core metaprogramming functions for forking off background tasks from elaboration so that their results show up in reporting and the language server. | 6170 |
v4.15.0 | Compiler | fixes a bug with the `structure`/`class` command where parents not represented as subobjects but used other parents as instances caused a kernel error. Closes #2611. | 6175 |
v4.15.0 | Compiler | fixes a non-termination bug when generating match-expression equation theorems. The proof automation repeatedly applied `injection(` to the same local decl. See issue #6067. | 6180 |
v4.15.0 | Compiler | adds the builtin simproc `USize.reduceToNat` for `USize.toNat` on small literals. (Duplicate mention from above, but relevant to the code generator.) | 6190 |
v4.15.0 | Compiler | adds a primitive for accessing the current thread ID. | 6049 |
v4.15.0 | Compiler | fixes a bug in the proof assembly procedure used by the grind tactic (placeholder text). | 6056 |
v4.15.0 | Compiler | adds `Lean.loadPlugin` exposing functionality similar to `--plugin` to Lean code. | 6130 |
v4.15.0 | Pretty Printing | adds the option `pp.parens` (default: false) to eagerly insert parentheses, helpful for teaching and clarifying expression structure. | 2934 |
v4.15.0 | Pretty Printing | prevents `Nat.succ ?_` from pretty printing as `?_.succ`, improving `apply?` usability. | 6014 |
v4.15.0 | Pretty Printing | improves term info for coercions like `DFunLike.coe` so "go to definition" on the function name works. Also shows the coercee rather than the coercion expression on hover. | 6085 |
v4.15.0 | Pretty Printing | improves the `#print` command for structures to show all fields and their inheritance, hiding subobject parent details. Private constants pretty print better, handling current-module private names normally and others hygienically. | 6096 |
v4.15.0 | Pretty Printing | adds a new delab option `pp.coercions.types` which shows all coercions with an explicit type ascription when enabled. | 6119 |
v4.15.0 | Pretty Printing | ensures whitespace is printed before `+opt` and `-opt` config options, improving tactics like `simp?`. | 6161 |
v4.15.0 | Pretty Printing | fixes a bug where `#check ident` would ignore `pp.raw`. Closes #6090. | 6181 |
v4.15.0 | Pretty Printing | exposes the difference in "synthesized type class instance is not definitionally equal" errors. | 6213 |
v4.15.0 | Documentation | fixes a typo in the docstring for prec and improves the text slightly. | 6009 |
v4.15.0 | Documentation | join → flatten in docstring | 6040 |
v4.15.0 | Documentation | does some mild refactoring of the `Lean.Elab.StructInst` module while adding documentation. | 6110 |
v4.15.0 | Documentation | converts 3 doc-strings to module docs, presumably their intended usage. | 6144 |
v4.15.0 | Documentation | refine kernel code comments | 6150 |
v4.15.0 | Documentation | adjust file reference in Data.Sum | 6158 |
v4.15.0 | Documentation | explains the order in which `Expr.abstract` introduces de Bruijn indices. | 6239 |
v4.15.0 | Server | adds auto-completion for fields in structure instance notation. Querying completions in whitespace now shows the full list of fields. Custom syntax can enable this by wrapping its fields parser in `structInstFields`. | 5835 |
v4.15.0 | Server | fixes an old auto-completion bug where `x.` gave nonsensical completions if `x.` could not be elaborated as a dot completion. | 5837 |
v4.15.0 | Server | avoid max heartbeat error in completion | 5996 |
v4.15.0 | Server | fixes a regression with go-to-definition and doc highlight misbehaving on tactic blocks. | 6031 |
End of preview. Expand
in Dataset Viewer.
Lean 4 Changelog Dataset
Dataset Description
The Lean 4 Changelog Dataset provides structured, machine-readable entries for changes in Lean 4, including language and library updates, fixes, deprecations, and other modifications. This dataset focuses on release notes from Lean 4’s official changelogs, capturing the key updates that are most likely to impact users, developers, or researchers working with Lean 4.
By offering structured data for these changes, this dataset enables tooling and machine learning workflows that analyze Lean 4’s evolution, assist in migration between versions, or build question-answering systems for Lean-specific features.
Dataset Structure
The dataset includes the following fields:
- version: The Lean 4 version associated with the change (e.g., "v4.16.0").
- section: The section of the changelog where the change appears (e.g., "## Language," "## Libraries").
- text: The content of the change
- pull request: The associated pull request.
Example Row
- version: "v4.16.0"
- section: "## Language"
- title: "#6204"
- description: "Lets _ be used in numeric literals as a separator, e.g., 1_000_000."
Source Code
The dataset was created by processing the official Lean 4 changelogs to extract and structure their content into individual, machine-readable entries. While the changelogs cover Lean 4's development comprehensively, this dataset focuses on capturing actionable updates for users and researchers. References to GitHub pull requests (e.g., "#6204") are included in the title field, while the description field provides a clean, detailed summary of the update.
Usage
This dataset is designed for:
- Formal Methods Research: Understanding the evolution of Lean 4 features and their impact on theorem proving.
- Machine Learning Applications: Training models for Lean 4 version prediction, changelog summarization, or migration tool development.
- Educational Purposes: Teaching Lean 4 users about recent changes or deprecated features in new releases.
License
This dataset is distributed under the Apache 2.0 license, aligning with the licensing of the Lean 4 project.
Acknowledgments
- Official Lean 4 Repository: https://github.com/leanprover/lean4
- Dataset Maintainer: phanerozoic (Charles Norton)
- Inspired by: Andreas Florath's Facts, Propositions, and Proofs Dataset
- Downloads last month
- 26
Size of downloaded dataset files:
355 kB
Size of the auto-converted Parquet files:
113 kB
Number of rows:
1,997