This shortens `Array.findIdx?` code, by using termination_by (and
well-founded recursion) instead of a structural recursion trick, with
the intent to make it more proof friendly.
One motivation is that it makes it easier to write a proof that
`Array.findIdx?` and `List.findIdx?` are equivalent. Furthermore, this
will be useful to prove that more complex functions are equivalent.
Closes#3646
System.FilePath.parent did not return the correct parent path in the
case of absolute file paths
Example of previous behavior
```
(FilePath.mk "/foo").parent -> some (FilePath.mk "")
(System.FilePath.mk "/").parent -> some (FilePath.mk "")
```
The new behavior is based on rust's std::path::Path::parent function (as
previously described in comment in System.FilePath)
Example of updated behavior
```
(System.FilePath.mk "/foo").parent -> some (FilePath.mk "/")
(System.FilePath.mk "/").parent -> none
```
Behavior for relative file paths is unchanged
Closes#3618
Given a definition `foo`, they were previously called `foo._unfold`
until 4.7.0. We tried to rename them to `foo.def`, but it created too
many issues in the Mathlib repo. We decided to rename it again to
`foo.eq_def`. The new name is also consistent with the `eq_<idx>`
theorems generated for different "cases". That is, `foo.eq_def` is the
equality theorem for the whole definition, and `foo.eq_<idx>` is the
equality theorem for case `<idx>`.
cc @semorrison
`FileMap.lines` is an array that seems to be manually managed to have
the form `#[1, 2, ..., n-1, n-1]` with same length as
`FileMap.positions`. Remove this structure field in favour of
calculating the line number as `min(x+1, positions.size-1)` when needed.
Follow-up on #3221
Given
```lean
structure A where
x : Nat
structure B extends A where
y : Nat
```
rather than pretty printing `{ x := 1, y := 2 : B }` as `{ toA := { x :=
1 }, y := 2 }`, it now pretty prints as `{ x := 1, y := 2 }`.
The option `pp.structureInstances.flatten` controls whether to flatten
structure instances like this.
This updates the rw? tactic from Mathlib to use lazy discriminator trees
and upstreams it.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Sets the default value to `pp.fieldNotation.generalized` to `true`.
Updates tests, and fixes some minor flaws in the implementation of the
generalized field notation pretty printer.
Now generalized field notation won't be used for any function that has a
`motive` argument. This is intended to prevent recursors from pretty
printing using it as (1) recursors are more like control flow structures
than actual functions and (2) generalized field notation tends to cause
elaboration problems for recursors.
Note: be sure functions that have an `@[app_unexpander]` use
`@[pp_nodot]` if applicable. For example, `List.toArray` needs
`@[pp_nodot]` to ensure the unexpander prints it using `#[...]`
notation.
The concrete dependency that is stale isn't really actionable
information for users (ideally we'd like something like "amount of
dependencies that will be rebuilt when you restart file"). This also
makes the diagnostic an "information" diagnostic so that non-infoview
users can still see it.
Since we are moving away from using notifications for stale dependency
information, we don't need to provide an ID anymore, either.
Refactors app delaborator, merging in the projection delaborator, to
support pretty printing with generalized field notation.
Renames option `pp.structureProjections` to `pp.fieldNotation` and adds
sub-option `pp.fieldNotation.generalized` to enable/disable generalized
field notation. Adds `@[pp_nodot]` attribute to permanently disable
using field notation for a given declaration.
For now, the default value of `pp.fieldNotation.generalized` is false
since we need a stage0 update to add `@[pp_nodot]` to some core
definitions (such as `List.toArray`) before updating the tests.
[Zulip
discussion](https://leanprover.zulipchat.com/#narrow/stream/270676-lean4/topic/.60pp.2EgeneralizedFieldNotation.60/near/425856054)
This attribute, which was implemented in #3640, is applied to the
following structures: `Sigma`, `PSigma`, `PProd`, `And`, `Subtype`, and
`Fin`. These were given this attribute in Lean 3.
This coercion caused difficult-to-diagnose bugs sometimes. Because there
are some situations where converting a string to a name should be done
by parsing the string, and others where it should not, an explicit
choice seems better here.
---------
Co-authored-by: Mac Malone <tydeu@hatpress.net>
I forgot to use the sticky diagnostics in `getInteractiveDiagnostics` in
#3247, leading to them not consistently showing up in the "Messages"
panel of the InfoView.
This lead to incorrect diagnostic spans in the editor and resulted in
header errors that did not show up under "Messages" everywhere in the
file because the `fullRange?` property was missing.
Also changes the "Import out of date" warning diagnostic severity to
"Hint" so that it doesn't show up in the "Problems" view.
This is a rewrite of the `UnusedVariables` lint to inline and simplify
many of the dependent functions to try to improve the performance of
this lint, which quite often shows up in perf reports.
* The mvar assignment scanning is one of the most expensive parts of the
process, so we do two things to improve this:
* Lazily perform the scan only if we need it
* Use an object-pointer hashmap to ensure that we don't have quadratic
behavior when there are many mvar assignments with slight differences.
* The dependency on `Lean.Server` is removed, meaning we don't need to
do the LSP conversion stuff anymore. The main logic of reference finding
is inlined.
* We take `fvarAliases` into account, and union together fvars which are
aliases of a base fvar. (It would be great if we had `UnionFind` here.)
More docs will be added once we confirm an actual perf improvement.
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
Closes#3706
This derive handler's implementation is very similar to `BEq`'s, which
already ignores unused alternative so as to work correctly on indexed
inductive types. This PR simply implements the same solution as the one
present in
[`BEq.lean`](2c15cdda04/src/Lean/Elab/Deriving/BEq.lean (L94)).
After some tests, it doesn't seem like any other derive handler present
in Core suffers from the same issue (though some handlers don't work on
indexed inductives for other reasons).
Sends a diagnostic informing the user to run Restart File when a file
dependency is saved.
Based on #3014 because this feature was easier to implement with the new
architecture.
ToDo:
- [x] Adjust vscode-lean4 to display a notification when this diagnostic
appears in a non-annoying way
(https://github.com/leanprover/vscode-lean4/pull/393)
- [x] Use a file watcher to identify changes to files not tracked by VS
Code
- [x] Rebase onto master when #3014 is merged
These are used in Mathlib's `congr!` and `convert` tactics, which will
be upstreamed soon.
---------
Co-authored-by: Kyle Miller <kmill31415@gmail.com>
- Removes the public definitions `Array.eraseIdxAux` and
`Array.eraseIdxSzAux` which were implementation details.
- Motivation: `Array.eraseIdxAux` and `Array.eraseIdxSzAux` were clearly
not intended to remain public, but simply making them private would make
it inconvenient to unfold them when writing proofs in Std.
- Adds documentation comments to the public `Array.eraseIdx`-related
definitions which remain.
- Removes `Array.eraseIdx'` which was just `Array.feraseIdx` wrapped in
a subtype and adds `Array.size_feraseIdx` to prove the subtype property
as a standalone theorem.
Co-Authored-By: Daniel Windham <daniel@atlascomputing.org>
Previously, if there was a `nightly-testing-YYYY-MM-DD` tag at Std, but
not Mathlib, we were erroneously proceeding with Mathlib CI, and hence
using a probably-broken version of Mathlib.
using the `substVars` tactic on the goal can remove too much
information, as it does not take into account that the `motive` may
depend on the fixed parameters.
This is fixed by etracting `substVar` from `subst` which expects the
`x`, not the `h : x = rhs`, and then using this tactic on the local
declarations _after_ the `motive` exclusively.
a common pattern for recursive functions is
```
def countUp (n i acc : Nat) : Nat :=
if i < n then
countUp n (i+1) (acc + i)
else
acc
```
where we increase a value `i` until it hits an upper bound. This is
particularly common with array processing functions:
```
$ git grep 'termination_by.*size.*-' src/|wc -l
26
```
GuessLex now recognizes this pattern. The general approach is:
For every recursive call, check if the context contains hypotheses of
the form `e₁ < e₂` (or similar comparisions), and then consider `e₂ -
e₁` as a termination argument.
Currently, this only fires when `e₁` and `e₂` only depend on the
functions parameters, but not local let-bindings or variables bound in
local pattern matches.
Duplicates are removed.
In the table showing the termination argument failures, long termination
arguments are now given a number and abbreviated as e.g. `#4` in the
table headers.
More examples in the test file, here as some highlights:
```
def distinct (xs : Array Nat) : Bool :=
let rec loop (i j : Nat) : Bool :=
if _ : i < xs.size then
if _ : j < i then
if xs[j] = xs[i] then
false
else
loop i (j+1)
else
loop (i+1) 0
else
true
loop 0 0
```
infers
```
termination_by (Array.size xs - i, i - j)
```
and the weird functions where `i` goes up or down
```
def weird (xs : Array Nat) (i : Nat) : Bool :=
if _ : i < xs.size then
if _ : 0 < i then
if xs[i] = 42 then
weird xs.pop (i - 1)
else
weird xs (i+1)
else
weird xs (i+1)
else
true
decreasing_by all_goals simp_wf; omega
```
infers
```
termination_by (Array.size xs - i, i)
```
but unfortunately needs `decreasing_by` pending the “big
decreasing_tactic refactor” that
I expect we’ll want to do at some point.
this refactor prepares GuessLex to be able to infer more complex
termination arguments.
As a side-effect it fixes an (obscure) bug where `sizeOf` would be
applied to a term of the wrong type and thus a wrong `SizeOf` instance
could be inferred.
On Windows, we now compile all core `.o`s twice, once with and without
`dllexport`, for use in the shipped dynamic and static libraries,
respectively. On other platforms, we export always as before to avoid
the duplicate work.
---------
Co-authored-by: tydeu <tydeu@hatpress.net>
This reverts commit 4e3a8468c3 for PR
#3619. It looks like the CI in that commit didn't inform me that a test
was broken by the PR, so I managed to commit it despite the broken test.
- Add support for reserved declaration names. We use them for theorems
generated on demand.
- Equation theorems are not private declarations anymore.
- Generate equation theorems on demand when resolving symbols.
- Prevent users from creating declarations using reserved names. Users
can bypass it using meta-programming.
See next test for examples.
Before, the termination argument as inferred by `GuessLex` was passed
further
on as `Syntax`, to be elaborated later in `WF.Rel`.
This didn’t feel quite right anymore. In particular if we want to teach
`GuessLex` about guessing more complex termination arguments like
`xs.size -
i`, using `Expr` here is more natural.
So this introduces `TerminationArgument` based on an `Expr` to be used
here.
A side-effect of how the termination arguments are elaborated is that
the unused
variables linter will now look at `termination_by` variables, and that
parameters
past the colon are not even invisibly in scope, so `‹_›` will not find
them
See https://github.com/leanprover-community/mathlib4/pull/11370/files
for examples
of fixing these changes.
This PR fixes an issue where the file worker would not provide the
client with semantic tokens until the file had been elaborated
completely. The file worker now also tells the client to refresh its
semantic tokens after running "Restart File". This PR is based on #3271.
Fixes#1170.
This PR adds the module name to `RefIdent` in order to distinguish
conflicting names from different files. This also fixes related issues
in find-references or the call hierarchy feature.
It also adds some docstrings and stylistically refactors a bunch of
code.
This PR adds support for requests from the server to the client in the
language server. It is based on #3014 and was developed during an
experiment for #3247 that unfortunately did not go anywhere.
This replaces a few uses of initialize with builtin_initialize, and
removes some unneeded functionality added when it was unclear if lazy
discriminator trees would be efficient enough.
This introduces the `ArgsPacker` module and abstraction, to replace the
exising `PackDomain`/`PackMutual` code. The motivation was that we now
have more uses besides `Fix.lean` (`GuessLex` and `FunInd`), and the
code was spread in various places.
The goals are
* consistent function naming withing the the `PSigma` handling, the
`PSum` handling, and the combined interface
* avoid taking a type apart just based on the `PSigma`/`PSum` nesting,
to be robust in case the user happens to be using `PSigma`/`PSum`
somewhere. Therefore, always pass an `arity` or `numFuncs` or `varNames`
around.
* keep all the `PSigma`/`PSum` encoding logic contained within one
module (`ArgsPacker`), and keep that module independent of its users (so
no `EqnInfos` visible here).
* pick good variable names when matching on a packed argument
* the unary function now is either called `fun1._unary` or
`fun1._mutual`, never `fun1._unary._mutual`.
This file has less heavy dependencies than `PackMutual` had, so build
parallelism is improved as well.
This makes trailing whitespace visible and protectes them against
trimming by the editor, by appending the symbol ⏎ to such a line (and
also to any line that ends with such a symbol, to avoid ambiguities in
the case the message already had that symbol).
(Only the code action output / docstring parsing is affected; the error
message as sent
to the InfoView is unaffected.)
Fixes#3571
Modifies `dsimpLocation'` (which implements `dsimp?`) to take a
`simprocs : SimprocsArray` argument, like `simpLocation` and
`dsimpLocation`. This ensures that the behavior of `dsimp` matches
`dsimp?`.
---
Closes#3653
Enables the combination of `noncomputable unsafe` to be used for
definitions. Outside of pure theory, `noncomputable` is also useful to
prevent Lean from compiling a definition which will be implemented with
external code later. Such definitions may also wish to be marked
`unsafe` if they perform morally impure or memory-unsafe functions.
this makes `termination_by?` even slicker.
The heuristics is agressive in the non-mutual case (will omit `sizeOf`
if the argument is non-dependent and the `WellFoundedRelation` relation
is via `sizeOfWFRel`.
In the mutual case we'd also have to check the arguments, as they line
up in the termination argument, have the same types. I did not bother at
this point; in the mutual case we omit `sizeOf` only if the argument
type is `Nat`.
As a drive-by fix, `termination_by?` now also works on functions that
have only one plausible measure.
This is a temporary workaround for a limitation on Windows shared
libraries. We are getting errors of the form:
```
ld.lld: error: too many exported symbols (got 65572, max 65535)
```
Replaces `@[eliminator]` with two attributes `@[induction_eliminator]`
and `@[cases_eliminator]` for defining custom eliminators for the
`induction` and `cases` tactics, respectively.
Adds `Nat.recAux` and `Nat.casesAuxOn`, which are eliminators that are
defeq to `Nat.rec` and `Nat.casesOn`, but these use `0` and `n + 1`
rather than `Nat.zero` and `Nat.succ n`.
For example, using `induction` to prove that the factorial function is
positive now has the following goal states (thanks also to #3616 for the
goal state after unfolding).
```lean
example : 0 < fact x := by
induction x with
| zero => decide
| succ x ih =>
/-
x : Nat
ih : 0 < fact x
⊢ 0 < fact (x + 1)
-/
unfold fact
/-
...
⊢ 0 < (x + 1) * fact x
-/
simpa using ih
```
Thanks to @adamtopaz for initial work on splitting the `@[eliminator]`
attribute.
Floris van Doorn [reported on
Zulip](https://leanprover.zulipchat.com/#narrow/stream/270676-lean4/topic/have.20tactic.20error.20recovery/near/425283053)
that it is confusing that the `have : T := e` tactic completely fails if
the body `e` is not of type `T`. This is in contrast to `have : T := by
exact e`, which does not completely fail when `e` is not of type `T`.
This ends up being caused by `elabTermEnsuringType` throwing an error
when it fails to insert a coercion. Now, it detects this case, and it
checks the `errToSorry` flag to decide whether to throw the error or to
log the error and insert a `sorry`.
This is justified by `elabTermEnsuringType` being a frontend to
`elabTerm`, which inserts `sorry` on error.
An alternative would be to make `ensureType` respect `errToSorry`, but
there exists code that expects being able to catch when `ensureType`
fails. Making such code manipulate `errToSorry` seems error prone, and
this function is not a main entry point to the term elaborator, unlike
`elabTermEnsuringType`.
Remark: this commit removes the `jason1.lean` test. Motivation: It
breaks all the time due to changes we make, and it is not clear anymore
what it is testing.
---------
Co-authored-by: Joachim Breitner <mail@joachim-breitner.de>
This migrates lemmas about Nat `compare`, `min`, `max`, `dvd`, `gcd`,
`lcm` and `div`/`mod` from Std to Lean itself.
Std still has some additional recursors, `CoPrime` and a few additional
definitions that might merit further discussion prior to upstreaming.
If the `LEAN_GITHASH` environment variable is set, Lake will now use it
instead of the detected Lean's githash when computing traces for builds
and the elaborated Lake configuration. This override allows one to
replace the Lean version used by a library
(e.g., Mathlib) without completely rebuilding it, which is useful for
testing custom builds of Lean.
The `delabConstWithSignature` delaborator is responsible for pretty
printing constants with a declaration-like signature, with binders, a
colon, and a type. This is used by the `#check` command when it is given
just an identifier.
It used to accumulate binders from pi types indiscriminately, but this
led to unfriendly behavior. For example, `#check String.append` would
give
```
String.append (a✝ : String) (a✝¹ : String) : String
```
with inaccessible names. These appear because `String.append` is defined
using patterns, so it never names these parameters.
Now the delaborator stops accumulating binders once it reaches an
inaccessible name, and for example `#check String.append` now gives
```
String.append : String → String → String
```
We do not synthesize names for the sake of enabling binder syntax
because the binder names are part of the API of a function — one can use
`(arg := ...)` syntax to pass arguments by name. The delaborator also
now stops accumulating binders once it reaches a parameter with a name
already seen before — we then rely on the main delaborator to provide
that parameter with a fresh name when pretty printing the pi type.
As a special case, instance parameters with inaccessible names are
included as binders, pretty printing like `[LT α]`, rather than
relegating them (and all the remaining parameters) to after the colon.
It would be more accurate to pretty print this as `[inst✝ : LT α]`, but
we make the simplifying assumption that such instance parameters are
generally used via typeclass inference. Likely `inst✝` would not
directly appear in pretty printer output, and even if it appears in a
hover, users can likely figure out what is going on. (We may consider
making such `inst✝` variables pretty print as `‹LT α›` or
`infer_instance` in the future, to make this more consistent.)
Something we note here is that we do not do anything to make sure
parameters that can be used as named arguments actually appear named
after the colon (nor do we assure that the names are the correct names).
For example, one sees `foo : String → String → String` rather than `foo
: String → (baz : String) → String`. We can investigate this later if it
is wanted.
We also give `delabConstWithSignature` a `universes` flag to enable
turning off pretty printing universe levels parameters.
Closes#2846
this makes the ugly `fst`/`snd` variable names in the functional
induction principles go away.
Ironically I thought in order to fix these name, I should touch the
mutual/n-ary argument packing code used for well-founded recursion, and
embarked on a big refactor/rewrite of that code, only to find that at
least this particular instance of the issue was somewhere else. Hence
breaking this into its own PR; the refactoring will follow (and will
also improve some other variable names.)
closes#3022
With this commit, given the declaration
```
def foo : Nat → Nat
| 0 => 2
| n + 1 => foo n
```
when we unfold `foo (n+1)`, we now obtain `foo n` instead of `foo
(Nat.add n 0)`.
This fixes an issue discovered in Mathlib with the meta cache being
poisoned by using a name generator. It is difficult to reproduce due to
the name collisions being rare, but here is a minimal module with
definitions that result in an error:
```lean
prelude
universe u
inductive Unit2 : Type where
| unit : Unit2
inductive Eq2 {α : Sort u} : α → α → Prop where
| refl (a : α) : Eq2 a a
structure Subtype2 {α : Sort u} (p : α → Prop) where
val : α
def End (α) := α → α
theorem end_app_eq (α : Type u) (f : End α) (a : α) : Eq2 (f a) (f a) := Eq2.refl _
theorem Set.coe_eq_subtype {α : Type u} (s : α → Prop) : Eq2 (Subtype2 s) (Subtype2 s) := Eq2.refl _
def succAboveCases {_ : Unit2} {α : Unit2 → Sort u} (i : Unit2) (v : α i) : α i := v
theorem succAbove_cases_eq_insertNth : Eq2 @succAboveCases.{u + 1} @succAboveCases.{u + 1} := Eq2.refl _
```
Removing any of thee last 5 definitions avoids the error. Testing
against Mathlib shows this PR fixes the issue.
This bug is the real cause of leanprover/vscode-lean4#392.
At the end of a tactic state, the client calls
`getInteractiveDiagnostics` with a range `[last line of proof, last line
of proof + 1)`. The `fullRange` span of the `unresolved goals` error
however is something like `[(first line of proof, start character),
(last line of proof, nonzero end character)).
Since it operates on line numbers, `getInteractiveDiagnostics` would
then check whether `[last line of proof, last line of proof + 1)` and
`[first line of proof, last line of proof)` intersect, which is false
because of the excluded upper bound on the latter interval, despite the
fact that the end character in the last line may be nonzero.
This fix adjusts the intersection logic to use `[first line of proof,
last line of proof]` if the end character is nonzero.
Closesleanprover/vscode-lean4#392.
This PR enables import auto-completion to complete partial words in
imports.
Other inconsistencies that I've found in import completion already seem
to be fixed by #3014. Since it will be merged soon, there is no need to
invest time to fix these issues on master.
This adds the concept of **functional induction** to lean.
Derived from the definition of a (possibly mutually) recursive function,
a **functional
induction principle** is tailored to proofs about that function. For
example from:
```
def ackermann : Nat → Nat → Nat
| 0, m => m + 1
| n+1, 0 => ackermann n 1
| n+1, m+1 => ackermann n (ackermann (n + 1) m)
derive_functional_induction ackermann
```
we get
```
ackermann.induct (motive : Nat → Nat → Prop) (case1 : ∀ (m : Nat), motive 0 m)
(case2 : ∀ (n : Nat), motive n 1 → motive (Nat.succ n) 0)
(case3 : ∀ (n m : Nat), motive (n + 1) m → motive n (ackermann (n + 1) m) → motive (Nat.succ n) (Nat.succ m))
(x x : Nat) : motive x x
```
At the moment, the user has to ask for the functional induction
principle explicitly using
```
derive_functional_induction ackermann
```
The module docstring of `Lean/Meta/Tactic/FunInd.lean` contains more
details on the
design and implementation of this command.
More convenience around this (e.g. a `functional induction` tactic) will
follow eventually.
This PR includes a bunch of `PSum`/`PSigma` related functions in the
`Lean.Tactic.FunInd`
namespace. I plan to move these to `PackArgs`/`PackMutual` afterwards,
and do some cleaning
up as I do that.
---------
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
This is still WIP: the checklist for release candidates will get
finished as I do the release of `v4.7.0-rc1`.
---------
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
This adds a number of lemmas for simplification of `Bool` and `Prop`
terms. It pulls lemmas from Mathlib and adds additional lemmas where
confluence or consistency suggested they are needed.
It has been tested against Mathlib using some automated test
infrastructure.
That testing module is not yet included in this PR, but will be included
as part of this.
Note. There are currently some comments saying the origin of the simp
rule. These will be removed prior to merging, but are added to clarify
where the rule came from during review.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Proves
`Nat.mod_mul : x % (a * b) = x % a + a * (x / a % b)` and
`Nat.mod_pow_succ : x % b ^ (k + 1) = x % b ^ k + b ^ k * ((x / b ^ k) %
b)`, helpful for bitblasting.
We use `let_delayed` to elaborate `match_expr` join points, which
elaborate the body of the `let` before its value. Thus, there is a
difference between:
- `let_delayed f (x : Expr) := <val>; <body>`
- `let_delayed f := fun (x : Expr) => <val>; <body>`
In the latter, when `<body>` is elaborated, the elaborator does not know
that `f` takes an argument of type `Expr`, and that `f` is a function.
Before this commit ensures the former representation is used.
In v4.6.0, there was a significant regression in initial server startup
performance because the .ilean files got bigger in #3082 and we load the
information stored in all .ilean files synchronously when the server
starts up.
This PR makes this loading asynchronous. The trade-off is that requests
that are issued right after the initial server start when the references
are not fully loaded yet may yield incomplete results.
Benchmark for this in a separate PR soon after this one.
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
Else the `case` will now allow introducing all necessary variables.
Induction principles with `let` in the types of the cases will be more
common with #3432.
This implementation no longer reduces the type as it goes, but really
only counts
manifest foralls and lets. I find this more sensible and predictable: If
you have
```
theorem induction₂_symm {P : EReal → EReal → Prop} (symm : Symmetric P) …
```
then previously, writing
```
case symm =>
```
would actually bring a fresh `x` and `y` and variable `h : P x y` into
scope and produce a
goal of `P y x`, because `Symmetric P` happens to be
```
def Symmetric := ∀ ⦃x y⦄, x ≺ y → y ≺ x
```
After this change, after `case symm =>` will leave `Symmetric P` as the
goal.
This gives more control to the author of the induction hypothesis about
the actual
goal of the cases. This shows up in mathlib in two places; fixes in
https://github.com/leanprover-community/mathlib4/pull/11023.
I consider these improvements.
the user can now write `termination_by?` to see the termination argument
inferred by GuessLex, and turn it into `termination_by …` using the “Try
this” widget or a code action.
To be done later, maybe: Avoid writing `sizeOf` if it's not necessary.
When editing core Lean, the `pp.proofs` feature causes goal states to fail to display in the Infoview, instead showing only "error when printing message: unknown constant '«term⋯»'". This PR moves the `⋯` syntax from Init.NotationExtra to Lean.Elab.BuiltinTerm
It also makes it so that `⋯` elaborates as `_` while logging a warning, rather than throwing an error, which should be somewhat more friendly when copy/pasting from the Infoview.
Closes#3476
When using `set_option tactic.skipAssignedInstances false`, `simp` and
`rw` will synthesize instance implicit arguments even if they have
assigned by unification. If the synthesized argument does not match the
assigned one the rewrite is not performed. This option has been added
for backward compatibility.
```
example (a : Nat) :
(((a + (2 ^ 64 - 1)) % 2 ^ 64 + 1) * 8 - 1 - (a + (2 ^ 64 - 1)) % 2 ^ 64 * 8 + 1) = 8 := by
omega
```
used to time out, and now is fast.
(We will probably make separate changes later so the defeq checks would
be fast in any case here.)
During the switch to `.lake`, I overlooked updating the paths in
`LakeInstall`. This fixes that and helps prevent further mistakes by
using the same default definitions as the package configuration itself.
The `decide` tactic produces error messages that users find to be
obscure. Now:
1. If the `Decidable` instance reduces to `isFalse`, it reports that
`decide` failed because the proposition is false.
2. If the `Decidable` instance fails to reduce, it explains what
proposition it failed for, and it shows the reduced `Decidable` instance
rather than the `Decidable.decide` expression. That expression tends to
be less useful since it shows the unreduced `Decidable` argument (plus
it's a lot longer!)
Examples:
```lean
example : 1 ≠ 1 := by decide
/-
tactic 'decide' proved that the proposition
1 ≠ 1
is false
-/
opaque unknownProp : Prop
open scoped Classical in
example : unknownProp := by decide
/-
tactic 'decide' failed for proposition
unknownProp
since its 'Decidable' instance reduced to
Classical.choice ⋯
rather than to the 'isTrue' constructor.
-/
```
When reporting the error, `decide` only shows the whnf of the
`Decidable` instance. In the future we could consider having it reduce
all decidable instances present in the term, which can help with
determining the cause of failure (this was explored in
8cede580690faa5ce18683f168838b08b372bacb).
The elaboration function `Lean.Meta.coerceMonadLift?` inserts these
coercion helper functions into a term and tries to unfolded them with
`expandCoe`, but because that function only unfolds up to
reducible-and-instance transparency, these functions were not being
unfolded. The fix here is to give them the `@[reducible]` attribute.
with this, more functions will be proven terminating automatically,
namely those where after `simp_wf`, lexicographic order handling,
possibly `subst_vars` the remaining goal can be solved by `omega`.
Note that `simp_wf` already does simplification of the goal, so
this adds `omega`, not `(try simp) <;> omega` here.
There are certainly cases where `(try simp) <;> omega` will solve more
goals (e.g. due to the `subst_vars` in `decreasing_with`), and
`(try simp at *) <;> omega` even more. This PR errs on the side of
taking
smaller steps.
Just appending `<;> omega` to the existing
`simp (config := { arith := true, failIfUnchanged := false })` call
doesn’t work nicely, as that leaves forms like `Nat.sub` in the goal
that
`omega` does not seem to recognize.
This does *not* remove any of the existing ad-hoc `decreasing_trivial`
rules based on `apply` and `assumption`, to not regress over the status
quo (these rules may apply in cases where `omega` wouldn't “see”
everything, but `apply` due to defeq works).
Additionally, just extending makes bootstrapping easier; early in `Init`
where
`omega` does not work yet these other tactics can still be used.
(Using a single `omega`-based tactic was tried in #3478 but isn’t quite
possible yet, and will be postponed until we have better automation
including forward reasoning.)
with this, hopefully more obvious array accesses will be handled
automatically.
Just like #3503, this PR does not investiate which of the exitsting
tactics in `get_elem_tactic_trivial` are subsumed now and could be
dropped without (too much) breakage.
Before, app unexpanders would only be applied to entire applications.
However, some notations produce functions, and these functions can be
given additional arguments. The solution so far has been to write app
unexpanders so that they can take an arbitrary number of additional
arguments. However, as reported in [this Zulip
thread](https://leanprover.zulipchat.com/#narrow/stream/270676-lean4/topic/pretty.20printer.20bug/near/420662236),
this leads to misleading hover information in the Infoview. For example,
while `HAdd.hAdd f g 1` pretty prints as `(f + g) 1`, hovering over `f +
g` shows `f`. There is no way to fix the situation from within an app
unexpander; the expression position for `HAdd.hAdd f g` is absent, and
app unexpanders cannot register TermInfo.
This commit changes the app delaborator to try running app unexpanders
on every prefix of an application, from longest to shortest prefix. For
efficiency, it is careful to only try this when app delaborators do in
fact exist for the head constant, and it also ensures arguments are only
delaborated once. Then, in `(f + g) 1`, the `f + g` gets TermInfo
registered for that subexpression, making it properly hoverable.
The app delaborator is also refactored, and there are some bug fixes:
- app unexpanders only run when `pp.explicit` is false
- trailing parameters in under-applied applications are now only
considered up to reducible & instance transparency, which lets, for
example, optional arguments for `IO`-valued functions to be omitted.
(`IO` is a reader monad, so it's hiding a pi type)
- app unexpanders will no longer run for delaborators that use
`withOverApp`
- auto parameters now always pretty print, since we are not verifying
that the provided argument equals the result of evaluating the tactic
Furthermore, the `notation` command has been modified to generate an app
unexpander that relies on the app delaborator's new behavior.
The change to app unexpanders is reverse-compatible, but it's
recommended to update `@[app_unexpander]`s in downstream projects so
that they no longer handle overapplication themselves.
If Lake fails to download a cloud release, it will now print a warning
indicating that it is falling back to a local build. For example:
```
[0/2] Downloading cloud_test cloud release
[0/2] Building CloudTest
error: > curl -s -f -o [...] -L [...]
error: external command `curl` exited with code 22
warning: fetching cloud release failed; falling back to local build
```
This PR addresses several performance issues in the auto-completion
implementation. It also fixes a number of smaller bugs related to
auto-completion.
In a file with `import Mathlib`, the performance of various kinds of
completions has improved as follows:
- Completing `C`: 49000ms -> 1400ms
- Completing `Cat`: 14300ms -> 1000ms
- Completing `x.` for `x : Nat`: 3700ms -> 220ms
- Completing `.` for an expected type of `Nat`: 11000ms -> 180ms
The following bugs have been fixed as well:
- VS Code never used our custom completion order. Now, the server fuzzy
completion score decides the order that completions appear in.
- Dot auto-completion for private types did not work at all. It does
now.
- Completing `.<identifier>` (where the expected type is used to infer
the namespace) did not filter by the expected type and instead displayed
all matching constants in the respective namespace. Now, it uses the
expected type for filtering. Note that this is not perfect because
sub-namespaces are technically correct completions as well (e.g.
`.Foo.foobar`). Implementing this is future work.
- Completing `.` was often not possible at all. Now, as long as the `.`
is not used in a bracket (where it may be used for the anonymous lambda
feature, e.g. `(. + 1)`), it triggers the correct completion.
- Fixes#3228.
- The auto-completion in `#check` commands would always try to complete
identifiers using the full declaration name (including namespaces) if it
could be resolved. Now it simply uses the identifier itself in case
users want to complete this identifier to another identifier.
## Details
Regarding completion performance, I have more ideas on how to improve it
further in the future.
Other changes:
- The feature that completions with a matching expected type are sorted
to the top of the server-side ordering was removed. This was never
enabled in VS Code because it would use its own completion item order
and when testing it I found it to be more confusing than useful.
- In the server-side ordering, we would always display keywords at the
top of the list. They are now displayed according to their fuzzy match
score as well.
The following approaches have been used to improve performance:
- Pretty-printing the type for every single completion made up a
significant amount of the time needed to compute the completions. We now
do not pretty-print the type for every single completion that is offered
to the user anymore. Instead, the language server now supports
`completionItem/resolve` requests to compute the type lazily when the
user selects a completion item.
- Note that we need to keep the amount of properties that we compute in
a resolve request to a minimum. When the server receives the resolve
request, the document state may have changed from the state it was in
when the initial auto-completion request was received. LSP doesn't tell
us when it will stop sending resolve requests, so we cannot keep this
state around, as we would have to keep it around forever.
LSP's solution for this dilemma is to have servers send all the state
they need to compute a response to a resolve request to the client as
part of the initial auto completion response (which then sends it back
as part of the resolve request), but this is clearly infeasible for all
real language servers where the amount of state needed to resolve a
request is massive.
This means that the only practical solution is to use the current state
to compute a response to the resolve request, which may yield an
incorrect result. This scenario can especially occur when using
LiveShare where the document is edited by another person while cycling
through available completions.
- Request handlers can now specify a "header caching handler" that is
called after elaborating the header of a file. Request handlers can use
this caching handler to compute caches for information stored in the
header. The auto-completion uses this to pre-compute non-blacklisted
imported declarations, which in turn allow us to iterate only over
non-blacklisted imported declarations where we would before iterate over
all declarations in the environment. This is significant because
blacklisted declarations make up about 4/5 of all declarations.
- Dot completion now looks up names modulo private prefixes to figure
out whether a declaration is in the namespace of the type to the left of
the dot instead of first stripping the private prefix from the name and
then comparing it. This has the benefit that we do not need to scan the
full name in most cases.
This PR also adds a couple of regression tests for fixed bugs, but *no
benchmarks*. We will add these in the future when we add proper support
for benchmarking server interaction sessions to our benchmarking
architecture.
All tests that were broken by producing different completion output
(empty `detail` field, added `sortText?` and `data?` fields) have been
manually checked by me to be still correct before replacing their
expected output.
This is still a draft PR, but includes the core exact? and apply?
tactics.
Still need to convert to builtin syntax and test on Std.
---------
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
The current `ToExpr Int` instance produces `@Int.ofNat (@OfNat.ofNat Nat
i ...)` for nonnegative `i` and `@Int.negSucc (@OfNat.ofNat Nat (-i+1)
...)` for negative `i`.
However it should be producing `@OfNat.ofNat Int i ...` for nonnegative
`i`, and `@Neg.neg ... (@OfNat.ofNat Int (-i) ...)` for negative `i`.
Make `x.toNat * 2 + b.toNat` the simp normal form of `(concat x
b).toNat`.
The choice for multiplication and addition was inspired by `Nat.bit_val`
from Mathlib.
Also, because we have considerably more lemmas about multiplication and
`_ + 1` than about shifts and `_ ||| 1`.
This is very helpful when dealing with bitvectors, where a case analysis
on the bitwidth leaves one with hypotheses of the form `x<2^(Nat.succ
w)`.
Design decisions I am unsure about:
- Is creating a helper `succ?` the correct way to match on the exponent
`e+1`?
- I'm not certain why the prior call to `Int.ofNat_pow` also checked
that the exponent was a ground natural. I removed this, since we now
explicitly handle cases where the exponent is a term of the form `e+1`.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Co-authored-by: Joe Hendrix <joe@lean-fro.org>
Co-authored-by: Alex Keizer <alex@keizer.dev>
First (baby)-step to a `concat`-based `bitblast`: a characterization of
`concat` in terms of `getLsb`.
The proof might benefit slightly from a `toNat_concat` lemma, but I
wasn't sure what the normal form there should be, so I avoided it.
---------
Co-authored-by: Scott Morrison <scott@tqft.net>
Every usage of `carry` followed the pattern: `carry _ x.toNat y.toNat`,
so we've refactorod `carry` to take the `BitVec`s as arguments, and made
the `toNat` part of its definition.
PR #3432 will introduce more operations on `MatcherApp`, including somet
that have more dependencies.
This change prepares by introducing `Lean.Meta.Match.MatcherApp.Basic`
for the basic definition, and `Lean.Meta.MatcherApp.Transform` for the
transformations, currently `addArg` and `refineThrough`, but more to
come.
This PR is an effort to improve reasoning at the Nat level about
bitvectors and reduce of Fin and Nat.
It slightly tightens some proofs, but is generally aimed at reducing
inconsistencies between definitions at the Nat and Fin types in favor of
more consistently using Nat operations.
This ports leanprover/std4#664 to Lean core.
Here was the rational I provided in the discussion for
leanprover/std4#664:
It's mostly about consistency. If we use the same types and style in
definitions and proofs, there is less surprise when unfolding or
otherwise using definitions. We use some Nat based operations that
haven't been extended to Fin such as the bitwise operations, and I don't
want to pay the overhead of introducing a Fin version of every Bitvector
operation.
So this basically means Nat is preferred.
One argument potentially in favor of Fin is that we could reuse results
proven there, but that doesn't really seem to be the case so far.
A second argument is that we want to simplify expression to use more
canonical forms and we currently can pretty-print those operations
better using ofNat than ofFin. We could define the notations using ofFin
of course though, but that's additional operators that will show up in
expressions.
Adds a simple error-recovery mechanism to Lean's parser, similar to
those used in other combinator parsing libraries.
Lean itself isn't very amenable to error recovery with this mechanism,
as it requires global knowledge of the grammar in question to write
recovery rules that don't break backtracking or `<|>`. I only found a
few opportunities.
But for DSLs, this is really important. In particular, Verso parse
errors interacted very badly with Lean parse errors in a way that
required frequent "restart file" commands, but this mechanism allows me
to both recover from Verso parse errors and to have Lean skip the rest
of the file rather than repeatedly trying to parse it as Lean commands.
#3408 was somewhat large and didn't properly test the symm and label
attribute code after edits to the builtin versions.
This migrates the code for generating labeled attributes from Init back
to Lean so that the required definitions are in scope.
This also addresses a mistake in the symm elaborator that prevented symm
without location information from elaborating.
Both fixes have been tested on the Std test suite and successfully
passed.
Adds documentation to the `String.Iterator` API, mentored by
@eric-wieser and @david-christiansen
---------
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
This is a quite substantial tactic.
It also includes the infamour `NatCast` typeclass (which I've equipped
with a module-doc). I wasn't at all sure where that should live, so it
is currently randomly in `Lean/Elan/Tactic/NatCast.lean`: presumably if
we're doing this it will go somewhere in `Init`.
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
When updating Std, be careful that not every lemma has been upstreamed,
so we need to be careful to only delete things that have already been
declared.
Before the `zeta` / `zetaDelta` split, `dsimp` was performing `zeta`
by going inside of a `let`-expression, performing `zetaDelta`, and
then removing the unused `let`-expression.
Loose fvars are never supposed to be pretty printed, but having them
print with "fvar" in the name can help with debugging broken tactics and
elaborators.
Metaprogramming users often do not realize at first that `_uniq.???` in
pretty printing output refers to fvars not in the current local context.
in all uses of `CasesOnApp`, we treat `MatcherApp`s the same way,
dupliating a fair amount of relatively hairy code (and there is more to
come).
However, the `MatcherApp` abstraction is perfectly capable of
also representing `casesOn` applications, at least for the use cases
encountered so far.
So lets just (optionally) include `casesOn` applications when looking
for matchers,
and remove the `CasesOnApp` abstraction completely.
Incrementally unveil trace children for excessively large nodes to
improve infoview rendering time, adjust particularly chatty
`simp.ground` trace to make use of it.
Lake previously opened the configuration trace as read-write even if it
does not update the configuration. This meant it failed if the trace was
read-only. With this change, it now first acquires a read-only handle
and then, if and only if it determines the need for a reconfigure, does
it re-open the file with a read-write handle. Also, this change fixes a
potential deadlock (Lake will error instead) and generally clarifies the
trace locking code.
Fixes a bug with Lake cloud releases where a cloud release would produce
a different trace if the package was the root of the workspace versus a
dependency. Also, an explicit fetch of a cloud release (e.g., via `lake
build :release`) will now error out with a non-zero exit code if it
fails to find, download, and unpack a release.
Previously, `CasesOn.addArg?` would do that check inline, while
`MatcherApp.addArg?` would do it after the fact.
Now `MatcherApp.addArg?` uses the same idiom.
Also, makes both `addArg?` always fail if the argument was not refined.
The work on functional induction principles calls for more unification
between the handling of `CasesOnApp` and `MatcherApp`, so this is a step
in that direction.
This is pretty big PR that upstreams all of Std.Data.Int.Init in one go.
So far lemmas have seen minimal changes needed to adapt to Lean core
environment.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Changes the goal to `False`, retaining as much information as possible:
* If the goal is `False`, do nothing.
* If the goal is an implication or a function type, introduce the
argument and restart.
(In particular, if the goal is `x ≠ y`, introduce `x = y`.)
* Otherwise, for a propositional goal `P`, replace it with `¬ ¬ P`
(attempting to find a `Decidable` instance, but otherwise falling back
to working classically)
and introduce `¬ P`.
* For a non-propositional goal use `False.elim`.
`nat?` checks if an expression is a "natural number in normal form",
i.e. of the form `OfNat n`, where `n` matches `.lit (.natVal n)` for
some `n`.
and if so returns `n`.
This is a widely used helper function in Std/Mathlib when matching on
expressions.
I've reordered some definitions to keep things together. This
introduces:
```
/-- Return the function (name) and arguments of an application. -/
def getAppFnArgs (e : Expr) : Name × Array Expr :=
withApp e λ e a => (e.constName, a)
```
and
```
/-- If the expression is a constant, return that name. Otherwise return `Name.anonymous`. -/
def constName (e : Expr) : Name :=
e.constName?.getD Name.anonymous
```
This upstreams NatCast and IntCast alone independent of norm_cast in
#3322.
This will allow more efficiently upstreaming parts of Std.Data.Int
relevant for omega.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
By having the `pp.proofs` feature use `⋯` when omitting proofs, when
users copy/paste terms from the InfoView the elaborator can give an
error message explaining why the term cannot be elaborated.
Also adds `pp.proofs.threshold` option to allow users to pretty print
shallow proof terms. By default, only atomic proof terms are pretty
printed.
This adjustment was suggested in PR #3201, which added `⋯` and the
related `pp.deepTerms` option.
This makes it so that when `withOverApp` is handling overapplied
functions, the term produced by the supplied delaborator is hoverable in
the Infoview.
When projection functions are delaborated, intermediate parent
projections are no longer printed. For example, rather than pretty
printing as `o.toB.toA.x` with these `toB` and `toA` parent projections,
it pretty prints as `o.x`.
This feature is being upstreamed from mathlib.
This is not a complete upstreaming of that file (it also supports `∀ᵉ (x
< 2) (y < 3), p x y` as shorthand for `∀ x < 2, ∀ y < 3, p x y`, but I
don't think we need this; it is used in Mathlib).
Syntaxes still need to be made built-in.
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
This causes problems when used in conjunction with `#guard_msgs` (which
checks whitespace) and trailing whitespace removal. Discovered by
@PatrickMassot in verbose-lean4.
Again co-developed with @bollu.
Based on top of: #3225
While hunting down the performance discrepancy on qsort.lean between C
and LLVM we noticed there was a single, trivially optimizeable, alloca
(LLVM's stack memory allocation instruction) that had load/stores in the
hot code path. We then found:
https://groups.google.com/g/llvm-dev/c/e90HiFcFF7Y.
TLDR: `mem2reg`, the pass responsible for getting rid of allocas if
possible, only triggers on an alloca if it is in the first BB. The
allocas of the current implementation get put right at the location
where they are needed -> they are ignored by mem2reg.
Thus we decided to add functionality that allows us to push all allocas
up into the first BB.
We initially wanted to write `buildPrologueAlloca` in a `withReader`
style so:
1. get the current position of the builder
2. jump to first BB and do the thing
3. revert position to the original
However the LLVM C API does not expose an option to obtain the current
position of an IR builder. Thus we ended up at the current
implementation which resets the builder position to the end of the BB
that the function was called from. This is valid because we never
operate anywhere but the end of the current BB in the LLVM emitter.
The numbers on the qsort benchmark got improved by the change as
expected, however we are not fully there yet:
```
C:
Benchmark 1: ./qsort.lean.out 400
Time (mean ± σ): 2.005 s ± 0.013 s [User: 1.996 s, System: 0.003 s]
Range (min … max): 1.993 s … 2.036 s 10 runs
LLVM before aligning the types
Benchmark 1: ./qsort.lean.out 400
Time (mean ± σ): 2.151 s ± 0.007 s [User: 2.146 s, System: 0.001 s]
Range (min … max): 2.142 s … 2.161 s 10 runs
LLVM after aligning the types
Benchmark 1: ./qsort.lean.out 400
Time (mean ± σ): 2.073 s ± 0.011 s [User: 2.067 s, System: 0.002 s]
Range (min … max): 2.060 s … 2.097 s 10 runs
LLVM after this
Benchmark 1: ./qsort.lean.out 400
Time (mean ± σ): 2.038 s ± 0.009 s [User: 2.032 s, System: 0.001 s]
Range (min … max): 2.027 s … 2.052 s 10 runs
```
Note: If you wish to merge this PR independently from its predecessor,
there is no technical dependency between the two, I'm merely stacking
them so we can see the performance impacts of each more clearly.
Debugged and authored in collaboration with @bollu.
This PR fixes several performance regressions of the LLVM backend
compared to the C backend
as described in #3192. We are now at the point where some benchmarks
from `tests/bench` achieve consistently equal and sometimes ever so
slightly better performance when using LLVM instead of C. However there
are still a few testcases where we are lacking behind ever so slightly.
The PR contains two changes:
1. Using the same types for `lean.h` runtime functions in the LLVM
backend as in `lean.h` it turns out that:
a) LLVM does not throw an error if we declare a function with a
different type than it actually has. This happened on multiple occasions
here, in particular when the function used `unsigned`, as it was
wrongfully assumed to be `size_t` sized.
b) Refuses to inline a function to the call site if such a type mismatch
occurs. This means that we did not inline important functionality such
as `lean_ctor_set` and were thus slowed down compared to the C backend
which did this correctly.
2. While developing this change we noticed that LLVM does treat the
following as invalid: Having a function declared with a certain type but
called with integers of a different type. However this will manifest in
completely nonsensical errors upon optimizing the bitcode file through
`leanc` such as:
```
error: Invalid record (Producer: 'LLVM15.0.7' Reader: 'LLVM 15.0.7')
```
Presumably because the generate .bc file is invalid in the first place.
Thus we added a call to `LLVMVerifyModule` before serializing the module
into a bitcode file. This ended producing the expected type errors from
LLVM an aborting the bitcode file generation as expected.
We manually checked each function in `lean.h` that is mentioned in
`EmitLLVM.lean` to make sure that all of their types align correctly
now.
Quick overview of the fast benchmarks as measured on my machine, 2 runs
of LLVM and 2 runs of C to get a feeling for how far the averages move:
- binarytrees: basically equal performance
- binarytrees.st: basically equal performance
- const_fold: equal if not slightly better for LLVM
- deriv: LLVM has 8% more instructions than C but same wall clock time
- liasolver: basically equal performance
- qsort: LLVM is slower by 7% instructions, 4% time. We have identified
why the generated code is slower (there is a store/load in a hot loop in
LLVM that is not in C) but not figured out why that happens/how to
address it.
- rbmap: LLVM has 3% less instructions and 13% less wall-clock time than
C (woop woop)
- rbmap_1 and rbmap_10 show similar behavior
- rbmap_fbip: LLVM has 2% more instructions but 2% better wall time
- rbmap_library: equal if not slightly better for LLVM
- unionfind: LLVM has 5% more instructions but 4% better wall time
Leaving out benchmarks related to the compiler itself as I was too lazy
to keep recompiling it from scratch until we are on a level with C.
Summing things up, it appears that LLVM has now caught up or surpassed
the C backend in the microbenchmarks for the most part. Next steps from
our side are:
- trying to win the qsort benchmark
- figuring out why/how LLVM runs more instructions for less wall-clock
time. My current guesses would be measurement noise and/or better use of
micro architecture?
- measuring the larger benchmarks as well
We previously had the syntax for `change` and `change at`, but no
implementation.
This moves Kyle's implementation from Std.
This also changes the `changeLocalDecl` function to push nodes to the
infotree about FVar aliases.
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
A small fix to the `DiscrTree` documentation to reflect the fact that
implicit type arguments *are* indexed and do not become `star` or
`other`. The following is a reproduction:
```lean
import Lean
open Lean Meta Elab Tactic
elab "test_tac" t:term : tactic => do
Tactic.withMainContext do
let e ← Term.elabTerm t none
let a : DiscrTree Nat ← DiscrTree.empty.insert e 1 {}
logInfo m!"{a}"
example (α : Type) (ringAdd : Add α) : True := by
/- (Add.add => (node (Nat => (node (* => (node (0 => (node (1 => (node #[1])))))))))) -/
test_tac @Add.add Nat instAddNat 0 1
/- (Add.add => (node (_uniq.1154 => (node (* => (node (◾ => (node (◾ => (node #[1])))))))))) -/
test_tac @Add.add α ringAdd ?_ ?_
```
This moves the `rcases` and `obtain` tactics from Std, and makes them
built-in tactics.
We will separately move the test cases from Std after #3297
(`guard_expr`).
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
Moves the `@[coe]` attribute and associated elaborators/delaborators
from Std to Lean.
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
The `push_cast` tactic in Std currently uses a copy-paste version of
`mkSimpContext` that allows overriding `getSimpTheorems`. However it has
been diverging from the version in Lean.
This is one way of generalizing `mkSimpContext` in Lean to allow what is
needed downstream., but I'm not at all set on this one. As far as I can
see there are no other tactics currently using this.
`push_cast` itself just replaces `getSimpTheorems` with
`pushCastExt.getTheorems`, where `pushCastExt` is a simp extension. If
there is another approach that suits that situation it would be fine.
I've tested that the change in this PR works downstream.
This is used in the "Try this:" widget machinery powering `simp?`.
There is a test file in Std, which I am not upstreaming at the same
time, as that relies on more code actions / #guard_msgs material. That
test file will still of course test things from Std, and later it can be
reunited with the code it is testing.
---------
Co-authored-by: Leonardo de Moura <leomoura@amazon.com>
These additional options are currently implemented in Std in a function
`Format.prettyExtra` (via `open private`), and used to implement the
`simp?` functionality.
This just adds the options to the core function.
This does not completely empty `Std.Lean.Name`, as working out how to
document the difference between `Name.isInternalDetail` and
`Name.isImplementationDetail` requires further thought.
The induction principle used by `induction` may have explicit parameters
that are
not motive, target or “real” alternatives (that have the `motive` as
conclusion), e.g. restrictions on the `motive` or other parameters.
Previously, `induction` would treat them as normal alternatives, and try
to re-introduce the automatically reverted hypotheses. But this only
works when the `motive` is actually the conclusion in the type of that
alternative.
We now pay attention to that, thread that information through, and only
revert when needed.
Fixes#3212.
This PR adds links to some folder references in the docs, making them
easier to navigate.
Please advise if these need to be made to be full URIs rather than
relative paths in order to work correctly with the doc generation
tooling that is in place.
Implements the pretty printer option `pp.numericTypes` for including a
type ascription for numeric literals. For example, `(2 : Nat)`, `(-2 :
Int)`, and `(-2 / 3 : Rat)`. This is useful for debugging how arithmetic
expressions have elaborated or have been otherwise transformed. For
example, with exponentiation is is helpful knowing whether it is `x ^ (2
: Nat)` or `x ^ (2 : Real)`. This is like the Lean 3 option
`pp.numeralTypes` but it has a wider notion of a numeric literal.
Also implements the pretty printer option `pp.natLit` for including the
`nat_lit` prefix for raw natural number literals.
Closes#3021
When we declare a `simp` set using `register_simp_attr`, we
automatically create `simproc` set. However, users may create `simp`
sets programmatically, and the associated `simproc` set may be missing
and vice-versa.
Before this commit, `Simproc`s were defined as `Expr -> SimpM (Option Step)`, where `Step` is inductively defined as follows:
```
inductive Step where
| visit : Result → Step
| done : Result → Step
```
Here, `Result` is a structure containing the resulting expression and a proof demonstrating its equality to the input. Notably, the proof is optional; in its absence, `simp` assumes reflexivity.
A simproc can:
- Fail by returning `none`, indicating its inapplicability. In this case, the next suitable simproc is attempted, along with other simp extensions.
- Succeed and invoke further simplifications using the `.visit`
constructor. This action returns control to the beginning of the
simplification loop.
- Succeed and indicate that the result should not undergo further
simplifications. However, I find the current approach unsatisfactory, as it does not align with the methodology employed in `Transform.lean`, where we have the type:
```
inductive TransformStep where
/-- Return expression without visiting any subexpressions. -/
| done (e : Expr)
/--
Visit expression (which should be different from current expression) instead.
The new expression `e` is passed to `pre` again.
-/
| visit (e : Expr)
/--
Continue transformation with the given expression (defaults to current expression).
For `pre`, this means visiting the children of the expression.
For `post`, this is equivalent to returning `done`. -/
| continue (e? : Option Expr := none)
```
This type makes it clearer what is going on. The new `Simp.Step` type is similar but use `Result` instead of `Expr` because we need a proof.
Modifies the structure instance elaborator to
1. Fill in missing fields from sources in strict left-to-right order. In
`{a, b with}`, sometimes the elaborator
would ignore `a` even if both `a` and `b` provided the same field,
depending on what subobject fields they had.
2. Use the sources, or subobjects of the sources, to fill in entire
subobjects of the target structure as much as possible.
Currently, a field cannot be filled directly by a source itself
resulting in the term being eta expanded.
This change avoids this unnecessary and surprisingly costly extra eta
expansion.
Adds two new tests to illustrate the performance benefit (one courtesy
@semorrison). These are currently failing on master and succeed on this
branch.
There is one additional test to exercise the changes to the elaboration
of structure instances.
Changes to make mathlib build are in leanprover-community/mathlib4#9843
Closes#2451
This combines a few platform-related changes:
* Add a ternary `platformIndependent` Lean configuration option to
assert whether Lake should assume Lean code is platform-independent. If
`true`, Lake will exclude platform-independent objects like external
libraries or dynlibs created through `precompileModules` from module
traces. If `false`, Lake will add the platform to module traces. If
`none` (the default), Lake will retain the current behavior (modules are
platform-dependent if and only if it depends on native objects).
* Use `System.Platform.target` from #3207 as the platform descriptor in
Lake for the configuration file trace, the cloud release archive, and as
the platform trace in Lean modules and native artifacts (e.g., object
files, and static and shared libraries).
* Do not add the platform descriptor into custom build archive names
(i.e., a user-set `buildArchive` configuration). This allows users to
create cross-platform / platform-independent archives via a name
override should they so desire.
Closes#2754.
This replaces the no-op `unusedVariablesIgnoreFnsExt` environment
extension with an actual environment extension which can be extended
using either `@[unused_variables_ignore_fn]` or
`@[builtin_unused_variables_ignore_fn]` (although for the present all
the builtin `unused_variables_ignore_fn`s are being added using direct
calls to `builtin_initialize addBuiltinUnusedVariablesIgnoreFn`, because
this also works and a stage0 update is required before the attribute can
be used).
We would like to use this attribute to disable unused variables in
syntaxes defined in std and mathlib, like
[`proof_wanted`](https://leanprover.zulipchat.com/#narrow/stream/113488-general/topic/Unused.20variables.20and.20proof_wanted/near/408554690).
This PR adds two new delaboration settings: `pp.deepTerms : Bool`
(default: `true`) and `pp.deepTerms.threshold : Nat` (default: `20`).
Setting `pp.deepTerms` to `false` will make the delaborator terminate
early after `pp.deepTerms.threshold` layers of recursion and replace the
omitted subterm with the symbol `⋯` if the subterm is deeper than
`pp.deepTerms.threshold / 4` (i.e. it is not shallow). To display the
omitted subterm in the InfoView, `⋯` can be clicked to open a popup with
the delaborated subterm.
<details>
<summary>InfoView with pp.deepTerms set to false (click to show
image)</summary>
</details>
### Implementation
- The delaborator is adjusted to use the new configuration settings and
terminate early if the threshold is exceeded and the corresponding term
to omit is shallow.
- To be able to distinguish `⋯` from regular terms, a new constructor
`Lean.Elab.Info.ofOmissionInfo` is added to `Lean.Elab.Info` that takes
a value of a new type `Lean.Elab.OmissionInfo`.
- `ofOmissionInfo` is needed in `Lean.Widget.makePopup` for the
`Lean.Widget.InteractiveDiagnostics.infoToInteractive` RPC procedure
that is used to display popups when clicking on terms in the InfoView.
It ensures that the expansion of an omitted subterm is delaborated using
`explicit := false`, which is typically set to `true` in popups for
regular terms.
- Several `Info` widget utility functions are adjusted to support
`ofOmissionInfo`.
- The list delaborator is adjusted with special support for `⋯` so that
long lists `[x₁, ..., xₖ, ..., xₙ]` are shortened to `[x₁, ..., xₖ, ⋯]`.
this way this function does not have to peek at the `altType` to see
when there are no more arguments, which makes it a bit more explicit,
and also a bit more robust should one apply this function to the type of
an alternative with the motive already instantiated.
It seems this uncovered a variable shadow bug, where the counter `i` was
accidentially reset after removing the `i`’th entry in `ys`.
Adds support for `let_fun` to the `intro` and `intros` tactics. Also
adds support to `intro` for anonymous binder names, since the default
variable name for a `letFun` with an eta reduced body is anonymous.
Encouraged by the performance gains from making `rewrite` produce
smaller proof objects
(#3121) I am here looking for low-hanging fruit in `simp`.
Consider this typical example:
```
set_option pp.explicit true
theorem test
(a : Nat)
(b : Nat)
(c : Nat)
(heq : a = b)
(h : (c.add (c.add ((c.add b).add c))).add c = c)
: (c.add (c.add ((c.add a).add c))).add c = c
```
We get a rather nice proof term when using
```
:= by rw [heq]; assumption
```
namely
```
theorem test : ∀ (a b c : Nat),
@Eq Nat a b →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c :=
fun a b c heq h =>
@Eq.mpr (@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c)
(@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c)
(@congrArg Nat Prop a b (fun _a => @Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c _a) c))) c) c) heq) h
```
(this is with #3121).
But with `by simp only [heq]; assumption`, it looks rather different:
```
theorem test : ∀ (a b c : Nat),
@Eq Nat a b →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c :=
fun a b c heq h =>
@Eq.mpr (@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c)
(@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c)
(@id
(@Eq Prop (@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c)
(@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c))
(@congrFun Nat (fun a => Prop) (@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c))
(@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c))
(@congrArg Nat (Nat → Prop) (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c)
(Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) (@Eq Nat)
(@congrFun Nat (fun a => Nat) (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))))
(Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))))
(@congrArg Nat (Nat → Nat) (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c)))
(Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) Nat.add
(@congrArg Nat Nat (Nat.add c (Nat.add (Nat.add c a) c)) (Nat.add c (Nat.add (Nat.add c b) c)) (Nat.add c)
(@congrArg Nat Nat (Nat.add (Nat.add c a) c) (Nat.add (Nat.add c b) c) (Nat.add c)
(@congrFun Nat (fun a => Nat) (Nat.add (Nat.add c a)) (Nat.add (Nat.add c b))
(@congrArg Nat (Nat → Nat) (Nat.add c a) (Nat.add c b) Nat.add
(@congrArg Nat Nat a b (Nat.add c) heq))
c))))
c))
c))
h
```
Since simp uses only single-step `congrArg`/`congrFun` congruence lemmas
here, the proof
term grows very large, likely quadratic in this case.
Can we do better? Every nesting of `congrArg` (and it's little brother
`congrFun`) can be
turned into a single `congrArg` call.
In this PR I make making the smart app builders `Meta.mkCongrArg` and
`Meta.mkCongrFun` a bit
smarter and not only fuse with `Eq.refl`, but also with
`congrArg`/`congrFun`.
Now we get, in this simple example,
```
theorem test : ∀ (a b c : Nat),
@Eq Nat a b →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c →
@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c :=
fun a b c heq h =>
@Eq.mpr (@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c a) c))) c) c)
(@Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c b) c))) c) c)
(@congrArg Nat Prop a b (fun x => @Eq Nat (Nat.add (Nat.add c (Nat.add c (Nat.add (Nat.add c x) c))) c) c) heq) h
```
Let’s see if it works and how much we gain.
right now, the `induction` tactic accepts a custom eliminator using the
`using <ident>` syntax, but is restricted to identifiers. This
limitation becomes annoying when the elminator has explicit parameters
that are not targets, and the user (naturally) wants to be able to write
```
induction a, b, c using foo (x := …)
```
This generalizes the syntax to expressions and changes the code
accordingly.
This can be used to instantiate a multi-motive induction:
```
example (a : A) : True := by
induction a using A.rec (motive_2 := fun b => True)
case mkA b IH => exact trivial
case A => exact trivial
case mkB b IH => exact trivial
```
For this to work the term elaborator learned the `heedElabAsElim` flag,
`true` by default. But in the default setting, `A.rec (motive_2 := fun b
=> True)`
would fail to elaborate, because there is no expected type. So the
induction
tactic will elaborate in a mode where that attribute is simply ignored.
As a side effect, the “failed to infer implicit target” error message
is improved and prints the name of the implicit target that could not be
instantiated.
This PR adds support for the "call hierarchy" feature of LSP that allows
quickly navigating both inbound and outbound call sites of functions. In
this PR, "call" is taken to mean "usage", so inbound and outbound
references of all kinds of identifiers (e.g. functions or types) can be
navigated. To implement the call hierarchy feature, this PR implements
the LSP requests `textDocument/prepareCallHierarchy`,
`callHierarchy/incomingCalls` and `callHierarchy/outgoingCalls`.
<details>
<summary>Showing the call hierarchy (click to show image)</summary>
</details>
<details>
<summary>Incoming calls (click to show image)</summary>
</details>
<details>
<summary>Outgoing calls (click to show image)</summary>
</details>
It is based on #3159, which should be merged before this PR.
To route the parent declaration name through to the language server, the
`.ilean` format is adjusted, breaking backwards compatibility with
version 1 of the ILean format and yielding version 2.
This PR also makes the following more minor adjustments:
- `Lean.Server.findModuleRefs` now also combines the identifiers of
constants and FVars and prefers constant over FVars for the combined
identifier. This is necessary because e.g. declarations declared using
`where` yield both a constant (for usage outside of the function) and an
FVar (for usage inside of the function) with the same range, whereas we
would typically like all references to refer to the former. This also
fixes a bug introduced in #2462 where renaming a declaration declared
using `where` would not rename usages outside of the function, as well
as a bug in the unused variable linter where `where` declarations would
be reported as unused even if they were being used outside of the
function.
- The function converting `Lean.Server.RefInfo` to `Lean.Lsp.RefInfo`
now also computes the `Lean.DeclarationRanges` for parent declaration
names via `MetaM` and must hence be in `IO` now.
- Add a utility function `Array.groupByKey` to `HashMap.lean`.
- Stylistic refactoring of `Watchdog.lean` and `LanguageFeatures.lean`.
In the new snapshot design, we have a tree of `Task`s that represents
the asynchronously processed document structure. When transforming this
tree in response to a user edit, we want to quickly run through
reusable, already computed nodes of the tree synchronously and then
spawn new tasks for the new parts. The new flag allows us to do such
mixed sync/async tree transformations uniformly. This flag exists as
e.g.
[`ExecuteSynchronously`](https://learn.microsoft.com/en-us/dotnet/api/system.threading.tasks.taskcontinuationoptions?view=net-8.0)
in other runtimes.
I deleted internal links that seemed to have the character of "TODO". I
think that the residual TODO is of little value, given that we plan a
big revamp and revision soon anyway, but I could do it some other way as
well.
This makes changes to the definitions of Associativity, Commutativity,
Idempotence and Identity classes to be more aligned with Mathlib's
versions.
The changes are:
* Move classes are moved from `Lean` to root namespace.
* Drop `Is` prefix from names.
* Rename `IsNeutral` to `LawfulIdentity` and add Left and Right
subclasses.
* Change neutral/identity element to outParam.
* Introduce `HasIdentity` for operations not intended for proofs to
implement
The identity changes are to make this compatible with
[Mathlib](718042db9d/Mathlib/Init/Algebra/Classes.lean)
and to enable nicer fold operations in Std that can use type classes to
infer the identity/initial element on binary operations.
---------
Co-authored-by: Kyle Miller <kmill31415@gmail.com>
Makes the LLVM triple of the current platform available to Lean code
towards a solution for #2754.
Defaults to the empty string if the compiler is not clang, which can
introduce some divergence between CI and local builds but should not be
noticeable in most cases and is not really possible to avoid.
Recursive predefinitions contains “rec app” markers as mdata in the
predefinitions,
but sometimes these get in the way of termination checking, when you
have
```
[mdata (fun x => f)] arg
```
Therefore, the `preprocess` pass floats them out of applications
(originally
only for structural recursion, since #2818 also for well-founded
recursion).
But the code was incomplete: Because `Meta.transform` calls `post` on `f
x y` only
once (and not also on `f x`) one has to float out of nested applications
as well.
A consequence of this can be that in a recursive proof, `rw [foo]` does
not work
although `rw [foo _ _]` does.
Also adding the testcase where @david-christiansen and I stumbled over
this
(Maybe the two preprocess modules can be combined, now that #2973 is
landed, will try that
in a follow-up).
As suggested by @kmill, removing an unnecessary `let` (possibly only
there in the first place for copy/paste reasons) seems to fix the
included test.
This makes `~q()` matching in quote4 noticeably more useful in things
like `norm_num` (as it fixes
https://github.com/leanprover-community/quote4/issues/29)
It also makes a quote4 bug slightly more visible
(https://github.com/leanprover-community/quote4/issues/30), but the bug
there already existed anyway, and isn't caused by this patch.
Fixes#3065
Give n-ary `Expr.app` constructors such as `mkApp2`, `mkApp3`, ...,
`mkApp10` the `@[match_pattern]` attribute so that it is easier to read
and write pattern matching for applications.
This PR facilitates augmenting the context of an `InfoTree` with
*partial* contexts while elaborating a command. Using partial contexts,
this PR also adds support for tracking the parent declaration name of a
term in the `InfoTree`. The parent declaration name is needed to compute
the call hierarchy in #3082.
Specifically, the `Lean.Elab.InfoTree.context` constructor is refactored
to take a value of the new type `Lean.Elab.PartialContextInfo` instead
of a `Lean.Elab.ContextInfo`, which now refers to a full `InfoTree`
context. The `PartialContextInfo` is then merged into a `ContextInfo`
while traversing the tree using
`Lean.Elab.PartialContextInfo.mergeIntoOuter?`. The partial context
after executing `liftTermElabM` is stored in values of a new type
`Lean.Elab.CommandContextInfo`.
As a result of this, `Lean.Elab.ContextInfo.save` moves to
`Lean.Elab.CommandContextInfo.save`.
For obtaining the parent declaration for a term, a new typeclass
`MonadParentDecl` is introduced to save the parent declaration in
`Lean.Elab.withSaveParentDeclInfoContext`. `Lean.Elab.Term.withDeclName
x` now calls `withSaveParentDeclInfoContext x` to save the declaration
name.
### Migration
**The changes to the `InfoTree.context` constructor break backwards
compatibility with all downstream users that traverse the `InfoTree`
manually instead of going through the functions in `InfoUtils.lean`.**
To fix this, you can merge the outer `ContextInfo` in a traversal with
the `PartialContextInfo` of an `InfoTree.context` node using
`PartialContextInfo.mergeIntoOuter?`. See e.g.
`Lean.Elab.InfoTree.foldInfo` for an example:
```lean
partial def InfoTree.foldInfo (f : ContextInfo → Info → α → α) (init : α) : InfoTree → α :=
go none init
where go ctx? a
| context ctx t => go (ctx.mergeIntoOuter? ctx?) a t
| node i ts =>
let a := match ctx? with
| none => a
| some ctx => f ctx i a
ts.foldl (init := a) (go <| i.updateContext? ctx?)
| _ => a
```
Downstream users that manually save `InfoTree`s may need to adjust calls
to `ContextInfo.save` to use `CommandContextInfo.save` instead and
potentially wrap their `CommandContextInfo` in a
`PartialContextInfo.commandCtx` constructor when storing it in an
`InfoTree` or `ContextInfo.mk` when creating a full context.
### Motivation
As of now, `ContextInfo`s are always *full* contexts, constructed as if
they were always created in `liftTermElabM` after running the
`TermElabM` action. This is not strictly true; we already create
`ContextInfo`s in several places other than `liftTermElabM` and work
around the limitation that `ContextInfo`s are always full contexts in
certain places (e.g. `Info.updateContext?` is a crux that we need
because we can't always create partial contexts at the term-level), but
it has mostly worked out so far. Note that one must be very careful when
saving a `ContextInfo` in places other than `liftTermElabM` because the
context may not be as complete as we would like (e.g. it may lack
meta-variable assignments, potentially leading to a language server
panic).
Unfortunately, the parent declaration of a term is another example of a
context that cannot be provided in `liftTermElabM`: The parent
declaration is usually set via `withDeclName`, which itself lives in
`TermElabM`. So by the time we are trying to save the full
`ContextInfo`, the declaration name is already gone. There is no easy
fix for this like in the other cases where we would really just like to
augment the context with an extra field.
The refactor that we decided on to resolve the issue is to refactor the
`InfoTree` to take a `PartialContextInfo` instead of a `ContextInfo` and
have code that traverses the `InfoTree` merge inner contexts with outer
contexts to produce a full `ContextInfo` value.
### Bumps for downstream projects
- `lean-pr-testing-3159` branch at Std, not yet opened as a PR
- `lean-pr-testing-3159` branch at Mathlib, not yet opened as a PR
- https://github.com/leanprover/LeanInk/pull/57
- https://github.com/hargoniX/LeanInk/pull/1
- https://github.com/tydeu/lean4-alloy/pull/7
- https://github.com/leanprover-community/repl/pull/29
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
Currently we create `lean-pr-testing-NNNN` branches at Mathlib
automatically for each Lean PR.
We don't automatically create one at Std; mostly simply because Std
fails less often, so it has been okay to do this manually as needed. It
is conceptually simpler, however, if this is done uniformly.
This PR:
* does not proceed with Std/Mathlib CI unless the appropriate
`nightly-testing-YYYY-MM-DD` tag exists at Std (like it already doesn't
proceed if that tag is missing at Mathlib)
* creates `lean-pr-testing-NNNN` branches at Std
* when it creates `lean-pr-testing-NNNN` branches at Mathlib, updates
the Std dependency to use the `lean-pr-testing-NNNN` branch at Std
- [x] depends on #3199
Note that because most users do not have write access at Std, in order
to make updates to `lean-pr-testing-NNNN` branches there they will need
to make PRs. These will be merged with a very low bar, and feel free to
ping me for assistance on this. If this is annoying we will automate.
Also, frequent contributors to Lean may ask @digama0 or @joehendrix for
write access in order to easily work on these branches.
This PR requires that we have a secret here with write access at Std.
I'm arranging that [on
zulip](https://leanprover.zulipchat.com/#narrow/stream/348111-std4/topic/bot.20access/near/416686090).
I will update the documentation at
https://leanprover-community.github.io/contribute/tags_and_branches.html
to reflect these changes when they are merged.
---------
Co-authored-by: Joachim Breitner <mail@joachim-breitner.de>
As discussed during the FRO meeting 2024-01-18, we are changing the
`nightly-testing-YYYY-MM-DD` branches at Std and Mathlib from branches
to tags, in:
* https://github.com/leanprover/std4/pull/545
* https://github.com/leanprover-community/mathlib4/pull/9842
This PR updates the script that creates the `lean-pr-testing-NNNN`
branches at Mathlib so it is agnostic about whether
`nightly-testing-YYYY-MM-DD` will be a branch or a tag.
---------
Co-authored-by: Joachim Breitner <mail@joachim-breitner.de>
`Array.set!` and `Array.swap!` are fairly similar operations, both
modify an array, both take an index that it out of bounds.
But they behave different; all of these return `true`
```
#eval #[1,2].set! 2 42 == #[1,2] -- with panic
#reduce #[1,2].set! 2 42 == #[1,2] -- no panic
#eval #[1,2].swap! 0 2 == #[1,2] -- with panic
#reduce #[1,2].swap! 0 2 == default -- no panic
```
The implementations are
```
@[extern "lean_array_set"]
def Array.set! (a : Array α) (i : @& Nat) (v : α) : Array α :=
Array.setD a i v
```
but
```
@[extern "lean_array_swap"]
def swap! (a : Array α) (i j : @& Nat) : Array α :=
if h₁ : i < a.size then
if h₂ : j < a.size then swap a ⟨i, h₁⟩ ⟨j, h₂⟩
else panic! "index out of bounds"
else panic! "index out of bounds"
```
It seems to be more consistent to unify the behaviors, and define
```
@[extern "lean_array_swap"]
def swap! (a : Array α) (i j : @& Nat) : Array α :=
if h₁ : i < a.size then
if h₂ : j < a.size then swap a ⟨i, h₁⟩ ⟨j, h₂⟩
else a
else a
```
Also adds docstrings.
Fixes#3196
Consider
```
import Std.Tactic.ShowTerm
opaque a : Nat
opaque b : Nat
axiom a_eq_b : a = b
opaque P : Nat → Prop
set_option pp.explicit true
-- Using rw
example (h : P b) : P a := by show_term rw [a_eq_b]; assumption
```
Before, a typical proof term for `rewrite` looked like this:
```
-- Using the proof term that rw produces
example (h : P b) : P a :=
@Eq.mpr (P a) (P b)
(@id (@Eq Prop (P a) (P b))
(@Eq.ndrec Nat a (fun _a => @Eq Prop (P a) (P _a))
(@Eq.refl Prop (P a)) b a_eq_b))
h
```
which is rather round-about, applying `ndrec` to `refl`. It would be
more direct to write
```
example (h : P b) : P a :=
@Eq.mpr (P a) (P b)
(@id (@Eq Prop (P a) (P b))
(@congrArg Nat Prop a b (fun _a => (P _a)) a_eq_b))
h
```
which this change does.
This makes proof terms smaller, causing mild general speed up throughout
the code; if the brenchmarks don’t lie the highlights are
* olean size -2.034 %
* lint wall-clock -3.401 %
* buildtactic execution s -10.462 %
H'T to @digama0 for advice and help.
NB: One might even expect the even simpler
```
-- Using the proof term that I would have expected
example (h : P b) : P a :=
@Eq.ndrec Nat b (fun _a => P _a) h a a_eq_b.symm
```
but that would require non-local changes to the source code, so one step
at a time.
The `checkTargets` function introduced in 4a0f8bf2 as
```
checkTargets (targets : Array Expr) : MetaM Unit := do
let mut foundFVars : FVarIdSet := {}
for target in targets do
unless target.isFVar do
throwError "index in target's type is not a variable (consider using the `cases` tactic instead){indentExpr target}"
if foundFVars.contains target.fvarId! then
throwError "target (or one of its indices) occurs more than once{indentExpr target}"
```
looks like it tries to check for duplicate indices, but it doesn’t
actually, as `foundFVars` is never written to.
This adds
```
foundFVars := foundFVars.insert target.fvarId!
```
and a test case.
Maybe a linter that warns about `let mut` that are never writen to would
be useful?
I keep messing things up, so time for some guard rails, so check them
using
[actionlint](https://github.com/raven-actions/actionlint).
This also runs [shellcheck](https://www.shellcheck.net/) on the files.
Shellcheck
is a bit picky about putting double quotes around variables, and will
flag many
cases where we know it’s safe, but why not simply always write the safer
variant.
Unfortunately, actionlint does not (yet) check `actions/github-script`
scripts, which is
unfortunate. Maybe they will in the future
(https://github.com/rhysd/actionlint/issues/389)
there was a check
if !Structural.recArgHasLooseBVarsAt recFnName fixedPrefixSize e then
that would avoid going through `.refineThrough`/`.addArg` for
matcher/casesOn applications. It seems it tries to detect when refining
the motive/param is pointless, but it was too eager, and cause confusion
with, for example, this reasonably reasonable function:
def foo : (n : Nat) → (i : Fin n) → Bool
| 0, _ => false
| 1, _ => false
| _+2, _ => foo 1 ⟨0, Nat.zero_lt_one⟩
decreasing_by simp_wf; simp_arith
In particular, the `GuessLex` code later expects that the (implict)
`PProd.casesOn` in the implementation of `foo._unary` will refine the
paramter, because else the (rather picky) `unpackArg` fails. But it also
prevents this from being provable.
So let's try without this shortcut.
Fixing this also revealed that `withRecApps` wasn’t looking in all
corners
of a matcherApp/casesOnApp.
Fixes#3175
this didn’t recognize the new comments with an intro, and thus the bot
would post multiple comments.
The code was also out of sync with mathlib, fixing.
The `first(…)` in the `jq` program makes it more robust in case this
went wrong once (as on #3171) and there are now multiple PRs matching.
This uses the improved termination_by syntax to give Nat.gcd a cleaner
definition. It removes the last explicit use of WellFounded.fix in Init.
This was also partly motivated by leanprover/std4#520 so that unfold
Nat.gcd gives a sensible definition.
If the current manifest is from unsupported (or has errors), a bare
`lake update` will now discard it and create a new one from scratch
rather than erroring and requiring you to manually delete the manifest.
Lake will produce warnings noting it is ignoring such invalid manifests.
so far, our reference manual did not mention these at all, this takes
the discussion of recursive definition out of the “equation compiler”
section, put it into its own section, and expands it a bit.
This is more a MVP doc change to at least mention the features briefly,
and not the most polished and thought through didactic exposition. But
it provides a start for more improvements.
---------
Co-authored-by: Arthur Adjedj <arthur.adjedj@gmail.com>
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
This change
* moves `termination_by` and `decreasing_by` next to the function they
apply to
* simplify the syntax of `termination_by`
* apply the `decreasing_by` goal to all goals at once, for better
interactive use.
See the section in `RELEASES.md` for more details and migration advise.
This is a hard breaking change, requiring developers to touch every
`termination_by` in their code base. We decided to still do it as a
hard-breaking change, because supporting both old and new syntax at the
same time would be non-trivial, and not save that much. Moreover, this
requires changes to some metaprograms that developers might have
written, and supporting both syntaxes at the same time would make
_their_ migration harder.
To handle delaborating notations that are functions that can be applied
to arguments, extracts the core function application delaborator as a
separate function that accepts the number of arguments to process and a
delaborator to apply to the "head" of the expression.
Defines `withOverApp`, which has the same interface as the combinator of
the same name from std4, but it uses this core function application
delaborator.
Uses `withOverApp` to improve a number of application delaborators,
notably projections. This means Mathlib can stop using `pp_dot` for
structure fields that have function types.
Incidentally fixes `getParamKinds` to specialize default values to use
supplied arguments, which impacts how default arguments are delaborated.
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
Allow `simproc`s to be declared without setting the `[simproc]`
attribute. A `simproc` declaration is function + pattern.
Motivation: allow them to be provided as arguments to `simp` **and** `simp only`.
TODO: track their use in `simp`.
TODO: builtin simprocs
Motivations:
- We can simplify the big mutual recursion and the implementation.
- We can implement the support for `match`-expressions in the `pre` method.
- It is easier to define and simplify `Simprocs`.
The example was looping with the new `simp` reduction strategy. Here
is the looping trace.
```
List.reverseAux (List.reverseAux as []) bs
==> rewrite using reverseAux_reverseAux
List.reverseAux [] (List.reverseAux (List.reverseAux as []) bs)
==> unfold reverseAux
List.reverseAux (List.reverseAux as []) bs
==> rewrite using reverseAux_reverseAux
List.reverseAux [] (List.reverseAux (List.reverseAux as []) bs)
==> ...
```
See new test for example that takes exponential time without new simp
theorems.
TODO: replace auxiliary theorems with simprocs as soon as we implement them.
I was about to to address the TODO
/- TODO: check arity of the given function. If it takes a PSigma as the
last argument,
this function will produce incorrect results. -/
because we now have an arity-observing variant of `decodePackedArg?` in
`unpackArg` in `PackMutual`, and it would be prudent to use it here.
But I first wanted to create a test case that would actually exhibit
this corner case, and failed.
This code was added in 096e4eb6d0 and it had a test case, but not even
that test case seems to be actually using the `decodePackedArg?`
function, neither back then nor now.
Also, mathlib works without this code.
So this seems to be dead code, possibly due to other changes to the
system, and thus can be removed. A strategically place comments points
back to this PR in case we need to resurrect that code.
The pattern
```
for h : i in [:xs.size] do
let x := xs[i]'h.2
```
is occassionally useful to iterate over an array with the index in
hand. This PR extends the `get_elem_tactic_trivial` so that one can
simply write
```
for h : i in [:xs.size] do
let x := xs[i]
```
fixes#3032.
When looking at a PR I sometimes wonder which `nightly` release is this
based on, and is used for the mathlib testing.
Right now, the action uses a label (`toolchain-available`) for this, but
a label cannot easily carry more information.
It seems a rather simple way to communicate extra information is by
setting [commit
statuses](https://docs.github.com/en/rest/commits/statuses?apiVersion=2022-11-28#create-a-commit-status);
with this change the following statuses will appear in the PR:
One could also use
[checks](https://docs.github.com/en/rest/checks/runs?apiVersion=2022-11-28#create-a-check-run)
to add more information, even with a nicely formatted markdown
description as in [this
example](https://github.com/nomeata/lean4/pull/1/checks?check_run_id=20165137082),
but it seems there you can’t set a summary that’s visible without an
extra click, and Github seems to associate these checks to “the first
workflow”, which is odd. So using statuses seems fine here.
Often one uses bots writing PR comments for this purpose, but that's a
bit noisy (extra notifications etc.), especially for stuff that happens
on every PR, but isn’t always interesting/actionable
If this works well, we can use this for more pieces of information, and
a link can be added as well.
This removes checks in `Lean.Meta.reduceNat?` that caused it to fail on
terms it could handle because they contain meta variables in arguments.
This lead to those operations being reduced using their equational
definitions and slow performance on large patterns:
```
set_option profiler true
set_option profiler.threshold 1
def testMod (x:Nat) :=
match x with
| 128 % 1024 => true
| _ => false
-- elaboration took 3.02ms
def testMul (x:Nat) :=
match x with
| 128 * 1 => true
| _ => false
-- type checking took 11.1ms
-- compilation of testMul.match_1 took 313ms
-- compilation of testMul took 65.7ms
-- elaboration took 58.9ms
```
Performance is slower on `testMul` than `testMod` because `whnf` ends up
evaluateing `128 * 1` using Peano arithmetic while `128 % 1024` is able
to avoid that treatment since `128 < 1024`.
This makes hover info, go to definition, etc work for the `h` in `cases
h : e`. The implementation is similar to that used for the `generalize h
: e = x` tactic.
#2966 was the `@[extern]` bug that prompted development of the
`test_extern` command, but then we merged the fix to #2966 without
updating the tests to use `test_extern`.
# Summary
This makes a small addition to our take on the LSP protocol
in the form of supporting snippet text edits.
It has been discussed
[here](https://github.com/microsoft/language-server-protocol/issues/592)
on the LSP issue tracker for a while,
but seems unlikely to be added anytime soon.
This feature was requested by @PatrickMassot for the purposes
of supporting Lean code templates in code actions and widgets.
---------
Co-authored-by: David Thrane Christiansen <david@davidchristiansen.dk>
This PR fixes the documentation error in "Extended Setup Notes", where
the path of builded binary is pointed to
`./build/bin/foo`, but the truly path is `./lake/build/bin/foo`.
---
Closes#3094 (`RFC` or `bug` issue number fixed by this PR, if any)
There were no `quot_precheck` instances registered for the expression
tree elaborators, which prevented them from being usable in a `notation`
expansion without turning off the quotation prechecker.
Users can evaluate whether `set_option quotPrecheck false` is still
necessary for their `notation` definitions.
given that we now use the PR description as the commit message, the PR
template should point that out. Also, a `# Summary` is relatively
strange in a commit message, so removed it.
---------
Co-authored-by: Scott Morrison <scott.morrison@gmail.com>
Switches from encoding `let_fun` using an annotated `(fun x : t => b) v`
expression to a function application `letFun v (fun x : t => b)`.
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
Getting the original PR number from a `workflow_run` cleanly and
reliably seems to be
basically impossible. See
<https://github.com/orgs/community/discussions/25220> for a discussion.
So for now let’s go back to the working state, even though it’s
deprecated and throws warnings.
This prompts users opening the workspace (on a new device) for the first
time to install the lean extension
# Summary
Link to `RFC` or `bug` issue: N/A
#3066 is causing CI failures, e.g.
[here](https://github.com/leanprover/lean4/actions/runs/7202184616/job/19619827364).
Although there are plenty of examples of using `await` in a Github
workflow script block, the error *seems* to be about this. This refactor
hopefully works around that, but I'm still uncertain of a root cause.
This fixe a surprisingly embarrassing bug introduced by me in
fa26d222cb (maybe while testing).
Enable more debug output while we are at it, to find out why sometimes
`context.payload.workflow_run.pull_requests[0]` is undefined.
If a user deleted `lakefile.olean` manually without deleting
`lakefile.olean.lock`, Lake would still attempt to load it and thus
produce an error. Now it should properly re-elaborate the configuration
file.
This adds a `test_extern` command.
Usage:
```
import Lean.Util.TestExtern
test_extern Nat.add 17 37
```
This:
* Checks that the head symbol has an `@[extern]` attribute.
* Writes down `t == t'`, where `t` is the term provided, and `t'` is the
reference implementation (specifically, `t` with the head symbol
unfolded).
* Tries to reduce this to `true`, and complains if this fails.
Note that the type of the term must have a `BEq` instance for this to
work: there's a self-explanatory error message if it isn't available.
the workflow is triggered not only by pull-request-CI-runs but also by
others. These should be skipped.
Also, no need to query the Github API to get the pull request number and
head sha, they are part of the payload, it seems.
Since the vscode-lean4 setup guide allows us to provide information on
setting up Lean 4 tailored to the user's operating system, this PR
adjusts the quickstart guide to reference the vscode-lean4 setup guide
instead.
This definition was clearly meant to be in the `List` namespace, but it
is also in a `namespace Lean` so it ended up as `Lean.List.toSMap`
instead of `List.toSMap`. It would be nice if #3031 made this
unnecessary, but for now this seems to be the convention.
I noticed this because of another side effect: it defines `Lean.List` as
a namespace, which means that
```lean
import Std
namespace Lean
open List
#check [1] <+ [2]
```
does not work as expected, it opens the `Lean.List` namespace instead of
the `List` namespace. Should there be a regression test to ensure that
the `Lean.List` namespace (and maybe others) are not accidentally
created? (Unfortunately this puts a bit of a damper on #3031.)
In particular:
* Do not use deprecated `potiuk/get-workflow-origin`.
* Use a bare checkout to push PR to `pr-releases`
* Replace `script/most-recent-nightly-tag.sh` by a one-liner inside the
workflow, so that th workflow is self-contained
Fixes an issue reported on Zulip; see the test case.
* Modifies the `MonadBacktrack` instance for `SimpM` to also backtrack
the `UsedSimps` field.
* When calling the discharger, `saveState`, and then `restoreState` if
something goes wrong.
I'm not certain that it makes sense to restore the `MetaM` state if
discharging fails. I can easily change this to more conservatively just
backtrack the `UsedSimps` after failed discharging.
Changes the implementation of `List.all` and `List.any` so they
short-circuit. The implementations are tail-recursive.
This replaces https://github.com/leanprover/std4/pull/392, which was
going to do this with `@[csimp]`.
until around 7fe6881 the way to define well-founded recursions was to
specify a `WellFoundedRelation` on the argument explicitly. This was
rather low-level, for example one had to predict the packing of multiple
arguments into `PProd`s, the packing of mutual functions into `PSum`s,
and the cliques that were calculated.
Then the current `termination_by` syntax was introduced, where you
specify the termination argument at a higher level (one clause per
functions, unpacked arguments), and the `WellFoundedRelation` is found
using type class resolution.
The old syntax was kept around as `termination_by'`. This is not used
anywhere in the lean, std, mathlib or the theorem-proving-in-lean
repositories,
and three occurrences I found in the wild can do without
In particular, it should be possible to express anything that the old
syntax
supported also with the new one, possibly requiring a helper type with a
suitable instance, or the following generic wrapper that now lives in
std
```
def wrap {α : Sort u} {r : α → α → Prop} (h : WellFounded r) (x : α) : {x : α // Acc r x}
```
Since the old syntax is unused, has an unhelpful name and relies on
internals, this removes the support. Now is a good time before the
refactoring that's planned in #2921.
The test suite was updated without particular surprises.
The parametric `terminationHint` parser is gone, which means we can
match on syntax more easily now, in `expandDecreasingBy?`.
This Github action automatically updates `stage0` on `master` if
`src/stdlib_flags.h` and `stage0/src/stdlib_flags.h`
are out of sync there.
It bypasses the merge queue to be quick, this way, an out-of-date stage0
on on
master should only exist for a few minutes.
Needs access to a _deploy SSH key_ with write permission.
The elaborator is prone to duplicate terms, including recursive calls,
even if the user only wrote a single one. This duplication is wasteful
if we run the tactics on duplicated calls, and confusing in the output
of GuessLex. So prune the list of recursive calls, and remove those
where another call exists that has the same goal and context that is no
more specific.
Implements "gaps" in string literals. These are escape sequences of the
form `"\" newline whitespace+` that have the interpretation of an empty
string. For example,
```
"this is \
a string"
```
is equivalent to `"this is a string"`. These are modeled after string
continuations in
[Rust](https://doc.rust-lang.org/beta/reference/tokens.html#string-literals).
Implements RFC #2838
before code like
def dup (a : Nat) (b : Nat := a) := a + b
def rec : Nat → Nat
| 0 => 1
| n+1 => dup (dup (dup (rec n)))
decreasing_by decreasing_tactic
would run the `decreasing_tactic` 8 tims, because the recursive call
`rec n` gets duplicate due to the default paramter. Similar effects can
be observed due to dependent types or tactics like `cases`.
This is wasteful, and is confusing to the user when they use
`decreasing_by` interactively. Therfore, we now go through the proof
obligations (MVars) and if solving one would imply solving another one,
we assign the mvars to each other accordingly.
This PR is a sibling of #3004.
I found the documentation page hard to parse, so I figured I should fix
this. It's mostly indentation (e.g. in lists), some line breaks and
making URLs clickable.
With
set_option showInferredTerminationBy true
this prints a message like
Inferred termination argument:
termination_by
ackermann n m => (sizeOf n, sizeOf m)
it tries hard to use names that
* match the names that the user used, if present
* have no daggers (so that it can be copied)
* do not shadow each other
* do not shadow anything from the environment (just to be nice)
it does so by appending sufficient `'` to the name.
Some of the emitted `sizeOf` calls are unnecessary, but they are needed
sometimes with dependent parameters. A follow-up PR will not emit them
for non-dependent arguments, so that in most cases the output is pretty.
Somewhen down the road we also want a code action, maybe triggered by
`termination_by?`. This should come after #2921, as that simplifies that
feature (no need to merge termination arguments from different cliques
for example.)
by showing the matrix of calls and measures, and what we know about that
call (=, <, ≤, ?), e.g.
guessLexFailures.lean:27:0-33:31: error: Could not find a decreasing
measure.
The arguments relate at each recursive call as follows:
(<, ≤, =: relation proved, ? all proofs failed, _: no proof attempted)
x1 x2 x3
1) 29:6-25 = = =
2) 30:6-23 = ? <
3) 31:6-23 < _ _
Please use `termination_by` to specify a decreasing measure
It’s a bit more verbose for mutual functions.
It will use the user-specified argument names for functions written
```
foo (n : Nat) := …
```
but not with pattern matching like
```
foo : Nat → …
| n => …
```
This can be refined later and separately (and maybe right away in
`expandMatchAltsWhereDecls`).
This is pure refactoring: Instead of solving each subgoal as we
encounter it while traversing the syntax tree, we leave the `MVar`
there, at the end collect them all using `getMVarsNoDelayed`, and then
solve them.
This is a refactoring preparing for two upcoming changes:
* removing unexpected duplicate goals that can arise from term
duplication
* running interactive tactics on all, not each goal (#2921)
In order to not regress with error locations, we have to associated the
`TermElabM`’s syntax refernce with the `MVar` somehow. I do this using
the existing `mkRecAppWithSyntax` expression annotation, on the `MVar`’s
type. Alternatives would be stack another `StateT` on the traversal
and accumulate `Array (MVarId, Syntax)` explicitly, but that did not
seem to be more appealing.
This sets the build directory to `build/release` for the "CMake Tools
for Visual Studio Code" extension documented at
https://vector-of-bool.github.io/docs/vscode-cmake-tools/settings.html#cmake-builddirectory.
It also sets the generator to `make`, since otherwise it tries `Ninja`
which doesn't work.
Without these settings, the extension runs configure in a bad place at
startup.
This does *not* add the cmake tools extension to the default workspace
configuration; the goal is simply to prevent bad behavior for users who
already have the extension enabled.
# Summary
Screenshot of this in action:
Link to `RFC` or `bug` issue: N/A, this is not a bug nor a user-visible
feature.
Now that we're, at least temporarily, relying more on the Nix CI,
replace some old hacks of mine with better solutions people have figured
out in the meantime.
Cachix support could probably be dropped at this point but it doesn't
really hurt.
In order to familiarize myself with this code, and so that the next
person has an easier time, I
* added docstrings explaining what I found out these things to
* rewrote the syntax expansion functions using syntax pattern matches,
to the extend possible
else I see
```
[ 69%] Building CXX object runtime/CMakeFiles/leanrt.dir/platform.cpp.o
/home/jojo/build/lean/lean4/src/runtime/io.cpp:509:75: warning: 'static_assert' with no message is a C++17 extension [-Wc++17-extensions]
static_assert(sizeof(std::chrono::milliseconds::rep) <= sizeof(uint64));
^
, ""
/home/jojo/build/lean/lean4/src/runtime/io.cpp:517:74: warning: 'static_assert' with no message is a C++17 extension [-Wc++17-extensions]
static_assert(sizeof(std::chrono::nanoseconds::rep) <= sizeof(uint64));
^
, ""
2 warnings generated.
```
when building
CI will now run on _any_ manually added label; hard to avoid.
Fun fact: Because the `toolchain-available` label is added by a github
action with the default token, it will _not_ trigger the workflow. Lucky
coincidence.
Following up on #2986, stop running the test suite in ci.yml in quick
mode; the test suite is run in the Nix job, and we do not need to run it
twice.
With a cold nix cache, when `lean` is rebuilt, not much changes, as both
jobs take ~20mins. But when `lean` is unchanged, the nix build should
be faster, and shaving off the (currently) 4mins in the CI.yaml run
should get us to a green PR sooner.
Another benefit is that we get the PR release sooner and even get it
when the test suite fails, which can be useful if you want to test
mathlib or other things before fixing the lean test suite.
which also removes an error condition at the use site.
While I am at it, I rename a parameter in `GuessLex` that I forgot to
rename earlier.
The effect will be user-visible (in obscure corner cases) with #2960, so
I’ll have the test there.
A few places would benefit from a `lambdaTelescopeBounded` that
garantees the result has the right length (eta-expanding when
necessary). I’ll look into that separately, and left TODOs here.
The goal of this change is to run a trimmed-down CI on PRs by default,
but allows opt-in the full CI as necessary.
### Specification
The CI workflow runs in “quick” mode if it was triggered from a pull
request, and that pull request does not have the `full-ci` label set.
In “quick” mode the build matrix contains fewer jobs. At the moment
only:
* Linux-release, to get the PR releases.
In non-quick mode everything should be as before.
### Implementation notes
I created a `configure` job that combines all the previous `set-` jobs,
I guess this is faster than firing up separate jobs.
The matrix is calculated in this job; this seems to be the cleanest way
to get a dynamic matrix going (experiments using `exclude` failed). The
downside is that the matrix is now in JSON rather than Yaml syntax. The
upside is that we can (later) make it’s calculation simpler, e.g. set
default `shell` values etc.
I was not able to make it so that CI runs when the `full-ci` label is
added, but don’t do anything otherwise. I think it can be done with
another workflow listening to `labeled` and then triggering this one,
but let’s do that separately. For now, add the label and then push (or
close and reopen).
The checks
```
if: matrix.build-stage2 || matrix.check-stage3
if: matrix.check-stage3
```
were dead code, we did not have these fields in the matrix anymore, so I
replaced them with
```
if: matrix.test-speedcenter
```
Now that there is a helpful message at the point of use when
`supportInterpreter` is required, we don't need to clutter every
`lakefile` with the advice.
there is a little dance with `if: success()` because otherwise a failed
`build` job would make this new job skipped, not failed, and I fear
skipped means ok when it is a required job.
So let’s make sure this job actually fails.
Also turn this into a proper check, run when a PR is opened or edited.
I took the liberty to rename the workflow file and name, so that one
doesn't have to look inside to guess what the workflow is doing.
If here is only one plausible measure, there is no point having the
`GuessLex` code see if it
is terminating, running all the tactics, only for the `MkFix` code then
run the tactics again.
So if there is only one plausible measure (non-mutual recursion with
only one varying
parameter), just use that measure.
Side benefit: If the function isn’t terminating, more detailed error
messages are shown
(failing proof goals), located at the recursive calls.
Removes the `CI` option from the `math` template. Since the template
does not currently generate a GitHub workflow, it does not do anything
out of the box except add unnecessary complexity.
The `math` template is also now tested in `tests/init` (minus the
Mathlib `require`).
This improves Lean’s capabilities to guess the termination measure for
well-founded
recursion, by also trying lexicographic orders. For example:
def ackermann (n m : Nat) := match n, m with
| 0, m => m + 1
| .succ n, 0 => ackermann n 1
| .succ n, .succ m => ackermann n (ackermann (n + 1) m)
now just works.
The module docstring of `Lean.Elab.PreDefinition.WF.GuessLex` tells the
technical story.
Fixes#2837
Closes#2548.
Later packages and libraries in the dependency tree are now preferred
over earlier ones. That is, the later ones "shadow" the earlier ones.
Such an ordering is more consistent with how declarations generally work
in programming languages.
This will break any package that relied on the previous ordering.
Also includes a related fix to `findModule?` that mistakenly treated
executable roots as importable.
Improves executable handling in `lake exe` and `lake init`:
* `lake exe <target>` now parses `target` like a build target (as the
help text states it should) rather than as a basic name.
* `lake new foo.bar [std]` now generates executables named `foo-bar`.
* `lake new foo.bar exe` now properly creates `foo/bar.lean`.
these are compagnions to `MatcherApp.addArg` and `CasesOnApp.addArg`
when one only has an
expression (which may not be a type) to transform, but not a concret
values.
This is a prerequisite for guessing lexicographic order (#2874). Keeping
this on a separate PR because it’s sizable, and has a clear independent
specification.
This is an additional safety net on top of #2749: it protects users that
circumvent the build system (e.g. with `lake env`) as well as obviates
the need for TOCTOU-like race condition checks in the build system.
The check is activated by `CHECK_OLEAN_VERSION=ON`, which now defaults
to `OFF` as the sensible default for local development. When activated,
`USE_GITHASH=ON` is also force-enabled for stage 0 in order to make sure
that stage 1 can load its own core library.
This PR adds per-package server options to resolve#2455. It is based on
the previous work in #2456, but takes a different approach: options are
loaded for the specific file in the file worker when `print-paths` is
called, instead of loading them in the watchdog with a separate Lake
command. This change addresses review comments made in #2456.
In doing so, it introduces two new Lake config fields:
- `leanOptions`: `-D` flag options that are passed to both the language
server and `lean` when building.
- `moreServerOptions`: `-D` flag options that are passed to the language
server.
Since `print-paths` must also accept a file path to compute the options
for that file, this PR is changing the API for `print-paths`. As there
have been numerous complaints about the name `print-paths`, I also
decided to change it to `setup-file` in this PR, since it would break
compatibility with the old Lake API anyways.
This PR deprecates the Lakefile field `moreServerArgs` in favor of
`moreGlobalServerArgs`, as suggested in the review for #2456.
Fixes#2455
---------
Co-authored-by: digama0 <mcarneir@andrew.cmu.edu>
This was a Lean 3 pretty printer option. While this pretty printer
option tends to lead to confusing situations when set, it has been
frequently requested. [It is
possible](https://github.com/leanprover-community/mathlib4/pull/7910) to
implement this pretty printer option as a user, but it comes with some
artifacts -- for instance, expressions in hovers are not beta reduced.
Adding this as a core pp option is cleanest.
(We should consider having hooks into the tactic evaluator to allow
users to transform the tactic state between tactics. This would enable
beta reducing the entire local context for real, which would be useful
for teaching.)
Closes#715
We noticed at
https://github.com/leanprover/lean4/pull/2923#discussion_r1400468371
that this instance is not used. It's arguably also incorrect (as it
doesn't backtrack the `usedTheorems` field).
Seems better to just remove to avoid confusion.
Evidence that this is dead code:
* After deleting the instance, calling `saveState` in the `SimpM` monad
raises an error `failed to synthesize instance MonadBacktrack PUnit
SimpM`.
* Understanding the `MonadBacktrack` monad leads one to believe that
would have happened, via the fact that the only instances for
`MonadBacktrack` are either concrete instances (e.g. for `MetaM`,
`TacticM`, etc), or a single lifting instance `instance [MonadBacktrack
s m] [Monad m] : MonadBacktrack s (ExceptT ε m)`. (This is good and
correct behaviour: lifting instances for `MonadBacktrack` would be hard
to model.)
* Mathlib builds after the instance is removed.
Potential evidence that I have not sought, because we don't have
sufficient tooling:
* Compiling Lean/Std/Mathlib with a debugger, breaking on entering this
code.
This PR adds basic auto-completion support for imports. Since it still
lacks Lake support for accurate completion suggestions (cc @tydeu - we
already know what needs to be done), it falls back to traversing the
`LEAN_SRC_PATH` for available imports.
Three kinds of import completion requests are supported:
- Completion of the full `import` command. Triggered when requesting
completions in an empty space within the header.
- Known issue: It is possible to trigger this completion within a
comment in the header. Fixing this would require architecture for
parsing some kind of sub-syntax between individual commands.
- Completion of the full module name after an incomplete `import`
command.
- Completion of a partial module name with a trailing dot.
Since the set of imports is potentially expensive to compute, they are
cached for 10 seconds after the last import auto-completion request.
Closes#2655.
### Changes
This PR also makes the following changes:
- To support completions on the trailing dot, the `import` syntax was
adjusted to provide partial syntax when a trailing dot is used.
- `FileWorker.lean` was refactored lightly with some larger definitions
being broken apart.
- The `WorkerState` gained two new fields:
- `currHeaderStx` tracks the current header syntax, as opposed to
tracking only the initial header syntax in `initHeaderStx`. When the
header syntax changes, a task is launched that restarts the file worker
after a certain delay to avoid constant restarts while editing the
header. During this time period, we may still want to serve import
auto-completion requests, so we need to know the up-to-date header
syntax.
- `importCachingTask?` contains a task that computes the set of
available imports.
- `determineLakePath` has moved to a new file `Lean/Util/LakePath.lean`
as it is now needed both in `ImportCompletion.lean` and
`FileWorker.lean`.
- `forEachModuleIn` from `Lake/Config/Blob.lean` has moved to
`Lean/Util/Path.lean` as it is a generally useful utility function that
was useful for traversing the `LEAN_SRC_PATH` as well.
### Tests
Unfortunately, this PR lacks tests since the set of imports available in
`tests/lean/interactive` will not be stable. In the future, I will add
support for testing LSP requests in full project setups, which is when
tests for import auto-completion will be added as well.
there were wrong italics, missing backticks, missing indentation and I
took the liberty to replace `[here]` links with link targets that better
tell the reader what to expect when clicking there.
The `packMutual` code ought to reliably replace all recursive calls to
the functions in `preDefs`, even when they are under- or over-applied.
Therefore eta-expand if need rsp. keep extra arguments around.
Needs a tweak to `Meta.transform` to avoid mistaking the `f` in
`f x1 x2` as a zero-arity application.
Includes a test case.
This fixes#2628 and #2883.
This didn't work before
```
def f (n : Nat) : Nat :=
match n with
| 0 => 0
| n + 1 => (f) n
```
because the `RecApp` metadata marker gets in the way. More practically
relevant, such code is to be produced when using `rw` or `simp` in
recursive theorems (see included test case).
We can fix this by preprocessing the definitions and floating the
`.mdata` marker out of applications.
For structural recursion, there already exists a `preprocess` function;
this now also floats out `.mdata` markers.
For well-founded recursion, this introduces an analogous `preprocess`
function.
Fixes#2810.
One test case output changes: With the `.mdata` out of the way, we get a
different error message. Seems fine.
Alternative approaches are:
* Leaving the `.mdata` marker where it is, and looking around it.
Tried in #2813, but not nice (many many places where `withApp` etc.
need to be adjusted).
* Moving the `.mdata` _inside_ the application, so that `withApp` still
works. Tried in #2814. Also not nice, the invariant that the `.mdata`
is around the `.const` is tedious to maintain.
the code stumbled over recursive functions whose type doesn’t have
enough manifest foralls, like:
```
def FunType := Nat → Nat
mutual
def foo : FunType
| .zero => 0
| .succ n => bar n
def bar : FunType
| .zero => 0
| .succ n => foo n
end
termination_by foo n => n; bar n => n
```
This can be fixed by using `whnf` in appropriate places, to expose the
`.forall` constructor.
Fixes#2925, comes with test case.
`script/most-recent-nightly-tag.sh` determines the most recent nightly
release in your current git history.
Previously it was assuming that you had a `nightly` remote, to pull tags
from. Now it just pulls directly from the repository by URL.
This only really shows up when the `decreasing_tactic` fails with
multiple goals, as in
```
macro_rules
| `(tactic|decreasing_tactic) => `(tactic| by_cases (2 > 1))
def foo (n : Nat) : Nat := foo (n - 1)
termination_by foo n => n
```
where we now get
```
unsolved goals
case inl
n: Nat
h✝: 2 > 1
⊢ (invImage (fun a => a) instWellFoundedRelation).1 (n - 1) n
case inr
n: Nat
h✝: ¬2 > 1
⊢ (invImage (fun a => a) instWellFoundedRelation).1 (n - 1) n
```
rather than
```
LeanProject.lean:3:27
unsolved goals
case inl
n: Nat
h✝: 2 > 1
⊢ (invImage (fun a => a) instWellFoundedRelation).1 (n - 1) n
LeanProject.lean:3:27
unsolved goals
case inr
n: Nat
h✝: ¬2 > 1
⊢ (invImage (fun a => a) instWellFoundedRelation).1 (n - 1) n
```
The effect is neglectible, but the code is a bit nicer, so why not,
before someone looks at it again and wonders whether the goals are
reported separately for a reason.
Some beginners have trouble finding the `if h : c then t else e`
(`dite`) version of `ite`. This augments `ite`'s docstring to mention
the dependent version.
This implements a request handler for the `textDocument/rename` LSP
request, enabling renames via F2. It handles both local renames (e.g.
`let x := 1; x` to `let y := 1; y`) as well as global renames
(definitions).
Unfortunately it does not work for "orphan" files outside a project, as
it uses ilean data for the current file and this does not seem to be
saved for orphan files. As a result, the test file does not work,
although one can manually test the implementation against a project such
as mathlib. (This issue already exists for the "references" request,
e.g. ctrl click on the first `x` in `let x := 1; x` takes you to the
second one only if you are not in an orphan file.)
* Fixesleanprover-community/mathlib4#7124
Modifies `cleanup` so that it takes (1) an array of additional fvarids
to preserve and (2) a flag to control whether to include indirect
propositions.
(This is wanted in mathlib for the `extract_goal` tactic.)
previously, it would ignore a recursive call that has extra arguments,
which can happen when the recursive functions return something of
function type. Therefore just leave them extra arguments in place.
Fixes#2883.
In the previous doc-string, the sentence
> "If any of the selected tactic applications fail, it will call
`failed` with the main goal mvar."
was false both for `Location.wildcard` (where it should have said "If
all", not "If any") or for `Location.targets` (where `failed` is never
called).
Add documentation comments with examples to `universe`, `open`,
`export`, and `variable`.
The documentation shows up when hovering over keywords, hopefully
improving the experience for beginners.
Because `Decidable` carries data,
when writing `@[simp]` lemmas which include a `Decidable` instance on the LHS,
it is best to use `{_ : Decidable p}` rather than `[Decidable p]`
so that non-canonical instances can be found via unification rather than
typeclass search.
(Previously this behaviour was often being hidden by the default `decide :=
true` in `simp`.)
We were checking out the synthetic merge commit between the PR `HEAD`
and `master`, and this was then breaking the logic to determine which
nightly-testing branches to use in Mathlib and Std.
previously, only the WellFounded code was making use of the error
location in the RecApp-metadata. We can do the same for structural
recursion. This way,
```
def f (n : Nat) : Nat :=
match n with
| 0 => 0
| n + 1 => f (n + 1)
```
will show the error with squiggly lines under `f (n + 1)`, and not at
`def f`.
`simp` was previously swallowing runtime exceptions and masking an
issue with this example.
`runT` is defined by well-founded recursion, but reducing the ground
term `runT x` takes a long time when `decide := true`.
Remark PR #2722 changes the `decide` default value to `false`.
When `decide := true`, we should probably have better diagnostics /
error messages for this kind of situation.
This is the same flag that the C test uses. Previously this was hidden
in the Lean compiler itself but now that the optimization pass is phased
out of the compiler we need to put it here.
Co-authored-by: Henrik Böving <hargonix@gmail.com>
In LLVM builds the Cmake CC is necessarily clang -> leanc will be able
to act on LLVM bitcode files if configured this way.
Co-authored-by: Siddharth <siddu.druid@gmail.com>
The notation `a ∈ as` for Arrays was previously only defined with
`DecidableEq` on the elements, for (apparently) no good reason. This
drops this requirements (by using `a ∈ as.data`), and simplifies a bunch
of proofs by simply lifting the corresponding proof from lists.
Also, `sizeOf_lt_of_mem` was defined, but not set up to be picked up by
`decreasing_trivial` in the same way that the corresponding List lemma
was set up, so this adds the tactic setup.
The definition for `a ∈ as` is intentionally not defeq to `a ∈ as.data`
so that the termination tactics for Arrays don’t spuriously apply when
recursing through lists.
This commit also removes parameter `simpleReduce` from discrimination
trees, and take WHNF configuration options.
Reason: it is more dynamic now. For example, the simplifier
will be able to use different configurations for discrimination tree insertion
and retrieval. We need this feature to address issues #2669 and #2281
This commit also removes the dead Meta.Config field `zetaNonDep`.
* feat: replay constants into an Environment
* suggestions from code review
* chore: make Environment.add private
* patch Lake to use Environment.add via extern
I initially expected `Name`s to always faithfully represent internal data, in particular that a name with macro scopes would have a form such as ```foo._@.Module._hyg.1``, and that tombstones would only appear in types that represent pretty-printed output such as as `String` or `Format`. However, that is not what happens. We have `sanitizeNames` which rewrites the `userName` field of local hypotheses to be `Name.str .anonymous "blah✝"`.
Then in the server code, we put these into `names : Array Name`e. This works fine for displaying in the infoview, but if we try to deserialize an `InteractiveHypothesisBundle` inside an RPC method for widget purposes, the `FromJson Name` instance blows up in `String.toName`.
I think my preferred solution is to, rather than 'fix' `String.toName` to accept these names with tombstones, stop pretending that they are actual `Name`s and re-type `InteractiveHypothesisBundle.names : Array String`. This should be a backwards-compatible change w.r.t. infoview code as the JSON representation is a string in either case. It is not backwards compatible w.r.t. meta code that uses this field.
Before, Lean would simply emit the message "tag not found". With this
change, it also tells the user what they could have written that would
be accepted.
* fix: make XML parser handle trailing whitespace in opening tags
* fix: make XML parser handle comments correctly
---------
Co-authored-by: György Kurucz <me@kuruczgy.com>
* fix: `withCollectingNewGoalsFrom`
do not collect old goals
* fix: update occurs check
* test: fix test `run/492.lean`
* docs: add docstring to `elabTermWithHoles`
* test: `refineFiltersOldMVars`
* test: fix `expected.out` name
* test: fix `expected.out` filename and line numbers
* docs: use long ascii dash instead of em dash
Co-authored-by: Scott Morrison <scott@tqft.net>
* docs: fix long line, mention lean4#2502
* docs: a couple more long lines
* test: fix line numbers
---------
Co-authored-by: Scott Morrison <scott@tqft.net>
* chore: add release notes for #2470 and #2480
* chore: begin development cycle for 4.2.0
* chore: add Lake-related release notes for v4.1.0
---------
Co-authored-by: Mac Malone <tydeu@hatpress.net>
* `weakLeanc/LinkArgs` for libs/mods/exes
* `weakArgs`/``extraDepTrace` for `buildO`
* include Lean trace when compiler is part of Lean toolchain
* do not include system-dependent file paths (e.g., `-I`) in dep trace
* step logging for cloud release fetching
* do not include cloud release bundle in trace
* `Package.afterReleaseSync/Async` utilities
* also cleanup `Package.recComputeDeps`
It seems that before, if `$src` isn’t a file, but a directory, that it
would contain `Bar.lean` directly, and not `Foo/Bar.lean`. This seemd
odd and would not allow dependencies to be included easily.
This obviates the need to play weak linkage games when
we build `lean.h.bc` from `lean.h`. We perform the following steps:
1. We remove the `static` modifier from all definitions in `lean.h`.
This makes all definitions have `extern` linkage. Thus, when we build
a `lean.h.bc` using `clang`, we will actually get definitions
(instead of an empty file)
2. We build `lean.h.bc` from `lean.h` using `clang`.
3. When it comes time to link, we link
`current_module.bc := LLVMLinkModules2(current_module.bc, lean.h.bc)`.
4. We loop over every symbol that arrived from `lean.h.bc`
in `current_module.bc` and we then set this symbol to have
`internal` linkage. This simulates the effect of
`#include <lean.h>` where every definition in `lean.h`
has internal linkage.
This yajna, one hopes, pleases the linker gods.
The previous type was
```
Lean.instFromJsonProd.{u, v} {α β : Type (max u v)} [FromJson α] [FromJson β] :
FromJson (α × β)
```
where universe metavariable assignment assigned the wrong universe to both types!
The test is flaky due to the presence of a fixed 'sleep()'.
The LLVM backend has introduced a performance
regression in Lake which causes this test to fail, as the
current sleep duration of 3s is insufficient.
Further investigation into the performance regression is pending.
We decided to disable the `leanlaketest_116` entirely on account
of the test being flaky by construction
(https://github.com/leanprover/lean4/pull/2358#issuecomment-1655371232).
* remove |- as an alias for ⊢
* revert false positive |->
* fix docstring
* undo previous changes
* [unchecked] use suggestion
* next attempt
* add test
* [ ] Put an X between the brackets on this line if you have done all of the following:
* Check that your issue is not already [filed](https://github.com/leanprover/lean4/issues).
* Reduce the issue to a minimal, self-contained, reproducible test case. Avoid dependencies to mathlib4 or std4.
### Description
[Clear and concise description of the issue]
### Context
[Broader context that the issue occured in. If there was any prior discussion on [the Lean Zulip](https://leanprover.zulipchat.com), link it here as well.]
### Steps to Reproduce
1.
2.
3.
**Expected behavior:** [Clear and concise description of what you expect to happen]
**Actual behavior:** [Clear and concise description of what actually happens]
### Versions
[Output of `#eval Lean.versionString` or of `lean --version` in the folder that the issue occured in]
[OS version]
### Additional Information
[Additional information, configuration or data that might be necessary to reproduce the issue]
### Impact
Add :+1: to [issues you consider important](https://github.com/leanprover/lean4/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). If others are impacted by this issue, please ask them to add :+1: to it.
Clear and detailed description of the proposal. Consider the following questions:
- **User Experience**: How does this feature improve the user experience?
- **Beneficiaries**: Which Lean users and projects benefit most from this feature/change?
- **Maintainability**: Will this change streamline code maintenance or simplify its structure?
### Community Feedback
Ideas should be discussed on [the Lean Zulip](https://leanprover.zulipchat.com) prior to submitting a proposal. Summarize all prior discussions and link them here.
### Impact
Add :+1: to [issues you consider important](https://github.com/leanprover/lean4/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). If others benefit from the changes in this proposal being added, please ask them to add :+1: to it.
* Ensure your PR follows the [External Contribution Guidelines](https://github.com/leanprover/lean4/blob/master/CONTRIBUTING.md).
* Please make sure the PR has excellent documentation and tests. If we label it `missing documentation` or `missing tests` then it needs fixing!
* Include the link to your `RFC` or `bug` issue in the description.
* If the issue does not already have approval from a developer, submit the PR as draft.
* The PR title/description will become the commit message. Keep it up-to-date as the PR evolves.
* If you rebase your PR onto `nightly-with-mathlib` then CI will test Mathlib against your PR.
* You can manage the `awaiting-review`, `awaiting-author`, and `WIP` labels yourself, by writing a comment containing one of these labels on its own line.
* Remove this section, up to and including the `---` before submitting.
---
Closes #0000 (`RFC` or `bug` issue number fixed by this PR, if any)
echo "... and Std has a 'nightly-testing-$MOST_RECENT_NIGHTLY' tag."
MESSAGE=""
if [[ -n "$MATHLIB_REMOTE_TAGS" ]]; then
echo "... and Mathlib has a 'nightly-testing-$MOST_RECENT_NIGHTLY' tag."
else
echo "... but Mathlib does not yet have a 'nightly-testing-$MOST_RECENT_NIGHTLY' tag."
MESSAGE="- ❗ Mathlib CI can not be attempted yet, as the \`nightly-testing-$MOST_RECENT_NIGHTLY\` tag does not exist there yet. We will retry when you push more commits. If you rebase your branch onto \`nightly-with-mathlib\`, Mathlib CI should run now."
fi
else
echo "... but Std does not yet have a 'nightly-testing-$MOST_RECENT_NIGHTLY' tag."
MESSAGE="- ❗ Std CI can not be attempted yet, as the \`nightly-testing-$MOST_RECENT_NIGHTLY\` tag does not exist there yet. We will retry when you push more commits. If you rebase your branch onto \`nightly-with-mathlib\`, Std CI should run now."
fi
else
echo "The most recently nightly tag on this branch has SHA: $NIGHTLY_SHA"
MESSAGE="- ❗ Std/Mathlib CI will not be attempted unless your PR branches off the \`nightly-with-mathlib\` branch. Try \`git rebase $MERGE_BASE_SHA --onto $NIGHTLY_SHA\`."
fi
if [[ -n "$MESSAGE" ]]; then
echo "Checking existing messages"
# The code for updating comments is duplicated in mathlib's
# scripts/lean-pr-testing-comments.sh
# so keep in sync
# Use GitHub API to check if a comment already exists
// check first commit only (and only once) since later commits might be intended to be squashed away
if (!/^(feat|fix|doc|style|refactor|test|chore|perf): .*[^.]($|\n\n)/.test(commits[0].commit.message)) {
await github.rest.issues.createComment({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.issue.number,
body: 'Thanks for your contribution! Please make sure to follow our [Commit Convention](https://leanprover.github.io/lean4/doc/dev/commit_convention.html).',
Thank you for your interest in contributing to Lean! There are many ways to contribute and we appreciate all of them.
In the past, we accepted most pull requests. This practice produced hard to maintain code, performance problems, and bugs. In order to improve the quality and maintainability of our codebase, we've established the following guidelines for external contributions.
## Bug reports
Helpful links
-------
Bug reports as new issues are always welcome. Please check the existing [issues](https://github.com/leanprover/lean4/issues) first.
Reduce the issue to a self-contained, reproducible test case.
If you have the chance, before reporting a bug, please search existing issues, as it's possible that
someone else has already reported your error.
If you're not sure if something is a bug or not, feel free to file a bug anyway. You may also want to discuss it with the Lean
community using the [lean4 Zulip channel](https://leanprover.zulipchat.com/#narrow/stream/270676-lean4).
Simple fixes for **typos and clear bugs** are welcome.
**Start with an Issue**: Before submitting a PR, always open an issue discussing the problem you wish to solve or the feature you'd like to add. Use the prefix `RFC:` (request for comments) if you are proposing a new feature. Ask for feedback from other users. Take the time to summarize all the feedback. This allows the maintainers to evaluate your proposal more efficiently. When creating a RFC, consider the following questions:
# **IMPORTANT**
- **User Experience**: How does this feature improve the user experience?
We are currently overwhelmed. We respectfully request that you hold off on submitting Pull Requests and creating Request for Comments (RFCs) at this time. Our team is actively seeking funding to expand the Lean development team and improve our capacity to review and integrate contributions. We appreciate your understanding and look forward to being able to accept contributions in the near future. In the meantime, the process described in the following sections is temporarily suspended.
- **Beneficiaries**: Which Lean users and projects do benefit most from this feature/change?
## Documentation
- **Community Feedback**: Have you sought feedback or insights from other Lean users?
Tutorial-like examples are very welcome.
They are useful for finding rough edges and bugs in Lean 4, for highlighting new features, and for showing how to use Lean.
If you want to store your tutorial in the Lean 4 repository to make sure future changes will not break it, we suggest the following workflow:
* Contact one of the Lean developers on Zulip, and check whether your tutorial is a good match for the Lean 4 repository.
* Send bug reports and report rough edges. We will work with you until the tutorial looks great.
* Add plenty of comments and make sure others will be able to follow it.
* Create a pull request in the Lean 4 repository. After merging, we will link it to the official documentation and make sure it becomes part of our test suite.
- **Maintainability**: Will this change streamline code maintenance or simplify its structure?
You can use `.lean` or `.md` files to create your tutorial. The `.md` files are ideal when you want to format your prose using markdown. For an example, see [this `.md` file](https://github.com/leanprover/lean4/blob/master/doc/lean3changes.md).
**Understand the Project**: Familiarize yourself with the project, existing issues, and latest commits. Ensure your contribution aligns with the project's direction and priorities.
Contributions to the reference manual are also welcome, but since Lean 4 is changing rapidly, please contact us first using Zulip
to find out which parts are stable enough to document. We will work with you to get this kind of
pull request merged. We are also happy to meet using Zoom, Skype or Google hangout to coordinate this kind of effort.
**Stay Updated**: Regularly fetch and merge changes from the main branch to ensure your branch is up-to-date and can be smoothly integrated.
As Lean 4 matures, other forms of documentation (e.g., doc-strings) will be welcome too.
**Help wanted**: We have issues tagged with ["help wanted"](https://github.com/leanprover/lean4/issues?q=is%3Aissue+is%3Aopen+label%3A%22help+wanted%22), if you want to contribute to the project, please take a look at them. If you are interested in one of them, post comments, ask questions, and engage with the core developers there.
## "Help wanted"
Quality Over Quantity:
-----
For issues marked as [`help wanted`](https://github.com/leanprover/lean4/issues?q=is%3Aissue+is%3Aopen+label%3A%22help+wanted%22), pull requests (PR) are welcome and we will work with you to get a PR merged. Some of these issues are nontrivial. If you are interested, please consider adding comments to the issue and/or messaging the Lean developers in [Zulip](https://leanprover.zulipchat.com/#).
**Focused Changes**: Each PR should address a single, clearly-defined issue or feature. Avoid making multiple unrelated changes in a single PR.
## Unexpected Pull Requests
**Write Tests**: Every new feature or bug fix should come with relevant tests. This ensures the robustness and reliability of the contribution.
We have very few core developers, and we cannot review arbitrary pull requests (PRs). Moreover, many features involve subtle tradeoffs, and it may require significant time and energy to even assess a proposed design. We suggest the following workflow:
**Documentation**: Update relevant documentation, including comments in the code, to explain the logic and reasoning behind your changes.
* First, discuss your idea with the Lean community on Zulip. Ask the community to help collect examples, document the requirements, and detect complications.
* If there is broad support, create a detailed issue for it on the Lean 4 repository at GitHub, and tag the issue with `RFC`.
* Ask the community for help documenting the requirements, and for collecting examples and concerns.
* Wait for one of the core developers to give you a "go ahead". At this point, the core developers will work with you to make sure your PR gets merged.
Coding Standards:
----
We don't want to waste your time by you implementing a feature and then us not being able to merge it.
**Follow the Code Style**: Ensure that your code follows the established coding style of the project.
## How to Contribute
**Lean on Lean**: Use Lean's built-in features and libraries effectively, avoiding reinventions.
* Always follow the [commit convention](https://leanprover.github.io/lean4/doc/dev/commit_convention.html).
* Follow the style of the surrounding code. When in doubt, look at other files using the particular syntax as well.
* Make sure your code is documented.
* New features or bug fixes should come with appropriate tests.
* Ensure all tests work before submitting a PR; see [Development Setup](https://leanprover.github.io/lean4/doc/make/index.html#development-setup) and [Fixing Tests](https://leanprover.github.io/lean4/doc/dev/fixing_tests.html).
**Performance**: Make sure that your changes do not introduce performance regressions. If possible, optimize the solution for speed and resource usage.
PR Submission:
---
**Descriptive Title and Summary**: The PR title should briefly explain the purpose of the PR. The summary should give more detailed information on what changes are made and why. Links to Zulip threads are not acceptable as a summary. You are responsible for summarizing the discussion, and getting support for it.
**Follow the commit convention**: Pull requests are squash merged, and the
commit message is taken from the pull request title and body, so make sure they adhere to the [commit convention](https://github.com/leanprover/lean4/blob/master/doc/dev/commit_convention.md). Put questions and extra information, which should not be part of the final commit message, into a first comment rather than the Pull Request description.
Because the change will be squashed, there is no need to polish the commit messages and history on the branch.
**Link to Relevant Issues**: Reference any issues that your PR addresses to provide context.
**Stay Responsive**: Once the PR is submitted, stay responsive to feedback and be prepared to make necessary revisions. We will close any PR that has been inactive (no response or updates from the submitter) for more than a month.
Reviews and Feedback:
----
**Be Patient**: Given the limited number of full-time maintainers and the volume of PRs, reviews may take some time.
**Engage Constructively**: Always approach feedback positively and constructively. Remember, reviews are about ensuring the best quality for the project, not personal criticism.
**Continuous Integration**: Ensure that all CI checks pass on your PR. Failed checks will delay the review process. The maintainers will not check PRs containing failures.
What to Expect:
----
**Not All PRs Get Merged**: While we appreciate every contribution, not all PRs will be merged. Ensure your changes align with the project's goals and quality standards.
**Feedback is a Gift**: It helps improve the project and can also help you grow as a developer or contributor.
**Community Involvement**: Engage with the Lean community on our communication channels. This can lead to better collaboration and understanding of the project's direction.
We intend to provide regular "minor version" releases of the Lean language at approximately monthly intervals.
There is not yet a strong guarantee of backwards compatibility between versions,
only an expectation that breaking changes will be documented in this file.
This file contains work-in-progress notes for the upcoming release, as well as previous stable releases.
Please check the [releases](https://github.com/leanprover/lean4/releases) page for the current status
of each version.
v4.8.0 (development in progress)
---------
* **Executables configured with `supportInterpreter := true` on Windows should now be run via `lake exe` to function properly.**
The way Lean is built on Windows has changed (see PR [#3601](https://github.com/leanprover/lean4/pull/3601)). As a result, Lake now dynamically links executables with `supportInterpreter := true` on Windows to `libleanshared.dll` and `libInit_shared.dll`. Therefore, such executables will not run unless those shared libraries are co-located with the executables or part of `PATH`. Running the executable via `lake exe` will ensure these libraries are part of `PATH`.
In a related change, the signature of the `nativeFacets` Lake configuration options has changed from a static `Array` to a function `(shouldExport : Bool) → Array`. See its docstring or Lake's [README](src/lake/README.md) for further details on the changed option.
* Lean now generates an error if the type of a theorem is **not** a proposition.
* Importing two different files containing proofs of the same theorem is no longer considered an error. This feature is particularly useful for theorems that are automatically generated on demand (e.g., equational theorems).
* New command `derive_functional_induction`:
Derived from the definition of a (possibly mutually) recursive function
defined by well-founded recursion, a **functional induction principle** is
tailored to proofs about that function. For example from:
is recognized without having to say `termination_by arr.size - i`.
* Attribute `@[pp_using_anonymous_constructor]` to make structures pretty print like `⟨x, y, z⟩`
rather than `{a := x, b := y, c := z}`.
This attribute is applied to `Sigma`, `PSigma`, `PProd`, `Subtype`, `And`, and `Fin`.
* Now structure instances pretty print with parent structures' fields inlined.
That is, if `B` extends `A`, then `{ toA := { x := 1 }, y := 2 }` now pretty prints as `{ x := 1, y := 2 }`.
Setting option `pp.structureInstances.flatten` to false turns this off.
* Option `pp.structureProjections` is renamed to `pp.fieldNotation`, and there is now a suboption `pp.fieldNotation.generalized`
to enable pretty printing function applications using generalized field notation (defaults to true).
Field notation can be disabled on a function-by-function basis using the `@[pp_nodot]` attribute.
Breaking changes:
* Automatically generated equational theorems are now named using suffix `.eq_<idx>` instead of `._eq_<idx>`, and `.def` instead of `._unfold`. Example:
* The coercion from `String` to `Name` was removed. Previously, it was `Name.mkSimple`, which does not separate strings at dots, but experience showed that this is not always the desired coercion. For the previous behavior, manually insert a call to `Name.mkSimple`.
v4.7.0
---------
* `simp` and `rw` now use instance arguments found by unification,
rather than always resynthesizing. For backwards compatibility, the original behaviour is
available via `set_option tactic.skipAssignedInstances false`.
[#3507](https://github.com/leanprover/lean4/pull/3507) and
* Improved initial language server startup performance. [#3552](https://github.com/leanprover/lean4/pull/3552)
* Changed call hierarchy to sort entries and strip private header from names displayed in the call hierarchy. [#3482](https://github.com/leanprover/lean4/pull/3482)
* There is now a low-level error recovery combinator in the parsing framework, primarily intended for DSLs. [#3413](https://github.com/leanprover/lean4/pull/3413)
* You can now write `termination_by?` after a declaration to see the automatically inferred
termination argument, and turn it into a `termination_by …` clause using the “Try this” widget or a code action. [#3514](https://github.com/leanprover/lean4/pull/3514)
* A large fraction of `Std` has been moved into the Lean repository.
This was motivated by:
1. Making universally useful tactics such as `ext`, `by_cases`, `change at`,
`norm_cast`, `rcases`, `simpa`, `simp?`, `omega`, and `exact?`
available to all users of Lean, without imports.
2. Minimizing the syntactic changes between plain Lean and Lean with `import Std`.
3. Simplifying the development process for the basic data types
`Nat`, `Int`, `Fin` (and variants such as `UInt64`), `List`, `Array`,
and `BitVec` as we begin making the APIs and simp normal forms for these types
more complete and consistent.
4. Laying the groundwork for the Std roadmap, as a library focused on
essential datatypes not provided by the core langauge (e.g. `RBMap`)
and utilities such as basic IO.
While we have achieved most of our initial aims in `v4.7.0-rc1`,
some upstreaming will continue over the coming months.
* The `/` and `%` notations in `Int` now use `Int.ediv` and `Int.emod`
(i.e. the rounding conventions have changed).
Previously `Std` overrode these notations, so this is no change for users of `Std`.
This automation will be substantially revised in the medium term,
and while `omega` does help automate some proofs, we plan to make this much more robust.
* The library search tactics `exact?` and `apply?` that were originally in
Mathlib are now available in Lean itself. These use the implementation using
lazy discrimination trees from `Std`, and thus do not require a disk cache but
have a slightly longer startup time. The order used for selection lemmas has
changed as well to favor goals purely based on how many terms in the head
pattern match the current goal.
* The `solve_by_elim` tactic has been ported from `Std` to Lean so that library
search can use it.
* New `#check_tactic` and `#check_simp` commands have been added. These are
useful for checking tactics (particularly `simp`) behave as expected in test
suites.
* Previously, app unexpanders would only be applied to entire applications. However, some notations produce
functions, and these functions can be given additional arguments. The solution so far has been to write app unexpanders so that they can take an arbitrary number of additional arguments. However this leads to misleading hover information in the Infoview. For example, while `HAdd.hAdd f g 1` pretty prints as `(f + g) 1`, hovering over `f + g` shows `f`. There is no way to fix the situation from within an app unexpander; the expression position for `HAdd.hAdd f g` is absent, and app unexpanders cannot register TermInfo.
This commit changes the app delaborator to try running app unexpanders on every prefix of an application, from longest to shortest prefix. For efficiency, it is careful to only try this when app delaborators do in fact exist for the head constant, and it also ensures arguments are only delaborated once. Then, in `(f + g) 1`, the `f + g` gets TermInfo registered for that subexpression, making it properly hoverable.
* fix `rename_i` in macros, fixes [#3553](https://github.com/leanprover/lean4/pull/3553) [#3581](https://github.com/leanprover/lean4/pull/3581)
* fix excessive resource usage in `generalize`, fixes [#3524](https://github.com/leanprover/lean4/pull/3524) [#3575](https://github.com/leanprover/lean4/pull/3575)
* an equation lemma with autoParam arguments fails to rewrite, fixing [#2243](https://github.com/leanprover/lean4/pull/2243) [#3316](https://github.com/leanprover/lean4/pull/3316)
* `add_decl_doc` should check that declarations are local [#3311](https://github.com/leanprover/lean4/pull/3311)
* instantiate the types of inductives with the right parameters, closing [#3242](https://github.com/leanprover/lean4/pull/3242) [#3246](https://github.com/leanprover/lean4/pull/3246)
* New simprocs for many basic types. [#3407](https://github.com/leanprover/lean4/pull/3407)
Lake fixes:
* Warn on fetch cloud release failure [#3401](https://github.com/leanprover/lean4/pull/3401)
* Backport of [#3552](https://github.com/leanprover/lean4/pull/3552) fixing a performance regression
in server startup.
v4.6.0
---------
* Add custom simplification procedures (aka `simproc`s) to `simp`. Simprocs can be triggered by the simplifier on a specified term-pattern. Here is an small example:
```lean
import Lean.Meta.Tactic.Simp.BuiltinSimprocs.Nat
def foo (x : Nat) : Nat :=
x + 10
/--
The `simproc` `reduceFoo` is invoked on terms that match the pattern `foo _`.
-/
simproc reduceFoo (foo _) :=
/- A term of type `Expr → SimpM Step -/
fun e => do
/-
The `Step` type has three constructors: `.done`, `.visit`, `.continue`.
* The constructor `.done` instructs `simp` that the result does
not need to be simplied further.
* The constructor `.visit` instructs `simp` to visit the resulting expression.
* The constructor `.continue` instructs `simp` to try other simplification procedures.
All three constructors take a `Result`. The `.continue` contructor may also take `none`.
`Result` has two fields `expr` (the new expression), and `proof?` (an optional proof).
If the new expression is definitionally equal to the input one, then `proof?` can be omitted or set to `none`.
-/
/- `simp` uses matching modulo reducibility. So, we ensure the term is a `foo`-application. -/
unless e.isAppOfArity ``foo 1 do
return .continue
/- `Nat.fromExpr?` tries to convert an expression into a `Nat` value -/
let some n ← Nat.fromExpr? e.appArg!
| return .continue
return .done { expr := Lean.mkNatLit (n+10) }
```
We disable simprocs support by using the command `set_option simprocs false`. This command is particularly useful when porting files to v4.6.0.
Simprocs can be scoped, manually added to `simp` commands, and suppressed using `-`. They are also supported by `simp?`. `simp only` does not execute any `simproc`. Here are some examples for the `simproc` defined above.
```lean
example : x + foo 2 = 12 + x := by
set_option simprocs false in
/- This `simp` command does not make progress since `simproc`s are disabled. -/
fail_if_success simp
simp_arith
example : x + foo 2 = 12 + x := by
/- `simp only` must not use the default simproc set. -/
fail_if_success simp only
simp_arith
example : x + foo 2 = 12 + x := by
/-
`simp only` does not use the default simproc set,
but we can provide simprocs as arguments. -/
simp only [reduceFoo]
simp_arith
example : x + foo 2 = 12 + x := by
/- We can use `-` to disable `simproc`s. -/
fail_if_success simp [-reduceFoo]
simp_arith
```
The command `register_simp_attr <id>` now creates a `simp` **and** a `simproc` set with the name `<id>`. The following command instructs Lean to insert the `reduceFoo` simplification procedure into the set `my_simp`. If no set is specified, Lean uses the default `simp` set.
```lean
simproc [my_simp] reduceFoo (foo _) := ...
```
* The syntax of the `termination_by` and `decreasing_by` termination hints is overhauled:
* They are now placed directly after the function they apply to, instead of
after the whole `mutual` block.
* Therefore, the function name no longer has to be mentioned in the hint.
* If the function has a `where` clause, the `termination_by` and
`decreasing_by` for that function come before the `where`. The
functions in the `where` clause can have their own termination hints, each
following the corresponding definition.
* The `termination_by` clause can only bind “extra parameters”, that are not
already bound by the function header, but are bound in a lambda (`:= fun x
y z =>`) or in patterns (`| x, n + 1 => …`). These extra parameters used to
be understood as a suffix of the function parameters; now it is a prefix.
Migration guide: In simple cases just remove the function name, and any
variables already bound at the header.
```diff
def foo : Nat → Nat → Nat := …
-termination_by foo a b => a - b
+termination_by a b => a - b
```
or
```diff
def foo : Nat → Nat → Nat := …
-termination_by _ a b => a - b
+termination_by a b => a - b
```
If the parameters are bound in the function header (before the `:`), remove them as well:
```diff
def foo (a b : Nat) : Nat := …
-termination_by foo a b => a - b
+termination_by a - b
```
Else, if there are multiple extra parameters, make sure to refer to the right
ones; the bound variables are interpreted from left to right, no longer from
right to left:
```diff
def foo : Nat → Nat → Nat → Nat
| a, b, c => …
-termination_by foo b c => b
+termination_by a b => b
```
In the case of a `mutual` block, place the termination arguments (without the
function name) next to the function definition:
```diff
-mutual
-def foo : Nat → Nat → Nat := …
-def bar : Nat → Nat := …
-end
-termination_by
- foo a b => a - b
- bar a => a
+mutual
+def foo : Nat → Nat → Nat := …
+termination_by a b => a - b
+def bar : Nat → Nat := …
+termination_by a => a
+end
```
Similarly, if you have (mutual) recursion through `where` or `let rec`, the
termination hints are now placed directly after the function they apply to:
```diff
-def foo (a b : Nat) : Nat := …
- where bar (x : Nat) : Nat := …
-termination_by
- foo a b => a - b
- bar x => x
+def foo (a b : Nat) : Nat := …
+termination_by a - b
+ where
+ bar (x : Nat) : Nat := …
+ termination_by x
-def foo (a b : Nat) : Nat :=
- let rec bar (x : Nat) : Nat := …
- …
-termination_by
- foo a b => a - b
- bar x => x
+def foo (a b : Nat) : Nat :=
+ let rec bar (x : Nat) : Nat := …
+ termination_by x
+ …
+termination_by a - b
```
In cases where a single `decreasing_by` clause applied to multiple mutually
recursive functions before, the tactic now has to be duplicated.
* The semantics of `decreasing_by` changed; the tactic is applied to all
termination proof goals together, not individually.
This helps when writing termination proofs interactively, as one can focus
each subgoal individually, for example using `·`. Previously, the given
tactic script had to work for _all_ goals, and one had to resort to tactic
combinators like `first`:
```diff
def foo (n : Nat) := … foo e1 … foo e2 …
-decreasing_by
-simp_wf
-first | apply something_about_e1; …
- | apply something_about_e2; …
+decreasing_by
+all_goals simp_wf
+· apply something_about_e1; …
+· apply something_about_e2; …
```
To obtain the old behaviour of applying a tactic to each goal individually,
use `all_goals`:
```diff
def foo (n : Nat) := …
-decreasing_by some_tactic
+decreasing_by all_goals some_tactic
```
In the case of mutual recursion each `decreasing_by` now applies to just its
function. If some functions in a recursive group do not have their own
`decreasing_by`, the default `decreasing_tactic` is used. If the same tactic
ought to be applied to multiple functions, the `decreasing_by` clause has to
be repeated at each of these functions.
* Modify `InfoTree.context` to facilitate augmenting it with partial contexts while elaborating a command. This breaks backwards compatibility with all downstream projects that traverse the `InfoTree` manually instead of going through the functions in `InfoUtils.lean`, as well as those manually creating and saving `InfoTree`s. See [PR #3159](https://github.com/leanprover/lean4/pull/3159) for how to migrate your code.
* Add language server support for [call hierarchy requests](https://www.youtube.com/watch?v=r5LA7ivUb2c) ([PR #3082](https://github.com/leanprover/lean4/pull/3082)). The change to the .ilean format in this PR means that projects must be fully rebuilt once in order to generate .ilean files with the new format before features like "find references" work correctly again.
* Structure instances with multiple sources (for example `{a, b, c with x := 0}`) now have their fields filled from these sources
in strict left-to-right order. Furthermore, the structure instance elaborator now aggressively use sources to fill in subobject
fields, which prevents unnecessary eta expansion of the sources,
and hence greatly reduces the reliance on costly structure eta reduction. This has a large impact on mathlib,
reducing total CPU instructions by 3% and enabling impactful refactors like leanprover-community/mathlib4#8386
which reduces the build time by almost 20%.
See [PR #2478](https://github.com/leanprover/lean4/pull/2478) and [RFC #2451](https://github.com/leanprover/lean4/issues/2451).
* Add pretty printer settings to omit deeply nested terms (`pp.deepTerms false` and `pp.deepTerms.threshold`) ([PR #3201](https://github.com/leanprover/lean4/pull/3201))
* Add pretty printer options `pp.numeralTypes` and `pp.natLit`.
When `pp.numeralTypes` is true, then natural number literals, integer literals, and rational number literals
are pretty printed with type ascriptions, such as `(2 : Rat)`, `(-2 : Rat)`, and `(-2 / 3 : Rat)`.
When `pp.natLit` is true, then raw natural number literals are pretty printed as `nat_lit 2`.
[PR #2933](https://github.com/leanprover/lean4/pull/2933) and [RFC #3021](https://github.com/leanprover/lean4/issues/3021).
Lake updates:
* improved platform information & control [#3226](https://github.com/leanprover/lean4/pull/3226)
* `lake update` from unsupported manifest versions [#3149](https://github.com/leanprover/lean4/pull/3149)
Other improvements:
* make `intro` be aware of `let_fun` [#3115](https://github.com/leanprover/lean4/pull/3115)
* produce simpler proof terms in `rw` [#3121](https://github.com/leanprover/lean4/pull/3121)
* fuse nested `mkCongrArg` calls in proofs generated by `simp` [#3203](https://github.com/leanprover/lean4/pull/3203)
* `induction using` followed by a general term [#3188](https://github.com/leanprover/lean4/pull/3188)
* allow generalization in `let` [#3060](https://github.com/leanprover/lean4/pull/3060), fixing [#3065](https://github.com/leanprover/lean4/issues/3065)
* reducing out-of-bounds `swap!` should return `a`, not `default`` [#3197](https://github.com/leanprover/lean4/pull/3197), fixing [#3196](https://github.com/leanprover/lean4/issues/3196)
* derive `BEq` on structure with `Prop`-fields [#3191](https://github.com/leanprover/lean4/pull/3191), fixing [#3140](https://github.com/leanprover/lean4/issues/3140)
* refine through more `casesOnApp`/`matcherApp` [#3176](https://github.com/leanprover/lean4/pull/3176), fixing [#3175](https://github.com/leanprover/lean4/pull/3175)
* do not strip dotted components from lean module names [#2994](https://github.com/leanprover/lean4/pull/2994), fixing [#2999](https://github.com/leanprover/lean4/issues/2999)
* fix `deriving` only deriving the first declaration for some handlers [#3058](https://github.com/leanprover/lean4/pull/3058), fixing [#3057](https://github.com/leanprover/lean4/issues/3057)
* do not instantiate metavariables in kabstract/rw for disallowed occurrences [#2539](https://github.com/leanprover/lean4/pull/2539), fixing [#2538](https://github.com/leanprover/lean4/issues/2538)
* hover info for `cases h : ...` [#3084](https://github.com/leanprover/lean4/pull/3084)
v4.5.0
---------
* Modify the lexical syntax of string literals to have string gaps, which are escape sequences of the form `"\" newline whitespace*`.
These have the interpetation of an empty string and allow a string to flow across multiple lines without introducing additional whitespace.
The following is equivalent to `"this is a string"`.
```lean
"this is \
a string"
```
[PR #2821](https://github.com/leanprover/lean4/pull/2821) and [RFC #2838](https://github.com/leanprover/lean4/issues/2838).
* Add raw string literal syntax. For example, `r"\n"` is equivalent to `"\\n"`, with no escape processing.
To include double quote characters in a raw string one can add sufficiently many `#` characters before and after
the bounding `"`s, as in `r#"the "the" is in quotes"#` for `"the \"the\" is in quotes"`.
[PR #2929](https://github.com/leanprover/lean4/pull/2929) and [issue #1422](https://github.com/leanprover/lean4/issues/1422).
* The low-level `termination_by'` clause is no longer supported.
Migration guide: Use `termination_by` instead, e.g.:
If the well-founded relation you want to use is not the one that the
`WellFoundedRelation` type class would infer for your termination argument,
you can use `WellFounded.wrap` from the std libarary to explicitly give one:
```diff
-termination_by' ⟨r, hwf⟩
+termination_by x => hwf.wrap x
```
* Support snippet edits in LSP `TextEdit`s. See `Lean.Lsp.SnippetString` for more details.
* Deprecations and changes in the widget API.
- `Widget.UserWidgetDefinition` is deprecated in favour of `Widget.Module`. The annotation `@[widget]` is deprecated in favour of `@[widget_module]`. To migrate a definition of type `UserWidgetDefinition`, remove the `name` field and replace the type with `Widget.Module`. Removing the `name` results in a title bar no longer being drawn above your panel widget. To add it back, draw it as part of the component using `<details open=true><summary class='mv2 pointer'>{name}</summary>{rest_of_widget}</details>`. See an example migration [here](https://github.com/leanprover/std4/pull/475/files#diff-857376079661a0c28a53b7ff84701afabbdf529836a6944d106c5294f0e68109R43-R83).
- The new command `show_panel_widgets` allows displaying always-on and locally-on panel widgets.
- `RpcEncodable` widget props can now be stored in the infotree.
- See [RFC 2963](https://github.com/leanprover/lean4/issues/2963) for more details and motivation.
* If no usable lexicographic order can be found automatically for a termination proof, explain why.
See [feat: GuessLex: if no measure is found, explain why](https://github.com/leanprover/lean4/pull/2960).
* Option to print [inferred termination argument](https://github.com/leanprover/lean4/pull/3012).
With `set_option showInferredTerminationBy true` you will get messages like
```
Inferred termination argument:
termination_by
ackermann n m => (sizeOf n, sizeOf m)
```
for automatically generated `termination_by` clauses.
* More detailed error messages for [invalid mutual blocks](https://github.com/leanprover/lean4/pull/2949).
* [Multiple](https://github.com/leanprover/lean4/pull/2923) [improvements](https://github.com/leanprover/lean4/pull/2969) to the output of `simp?` and `simp_all?`.
* Tactics with `withLocation *` [no longer fail](https://github.com/leanprover/lean4/pull/2917) if they close the main goal.
* Implementation of a `test_extern` command for writing tests for `@[extern]` and `@[implemented_by]` functions.
Usage is
```
import Lean.Util.TestExtern
test_extern Nat.add 17 37
```
The head symbol must be the constant with the `@[extern]` or `@[implemented_by]` attribute. The return type must have a `DecidableEq` instance.
Bug fix for [eager evaluation of default value](https://github.com/leanprover/lean4/pull/3043) in `Option.getD`.
Avoid [panic in `leanPosToLspPos`](https://github.com/leanprover/lean4/pull/3071) when file source is unavailable.
Improve [short-circuiting behavior](https://github.com/leanprover/lean4/pull/2972) for `List.all` and `List.any`.
Several Lake bug fixes: [#3036](https://github.com/leanprover/lean4/issues/3036), [#3064](https://github.com/leanprover/lean4/issues/3064), [#3069](https://github.com/leanprover/lean4/issues/3069).
v4.4.0
---------
* Lake and the language server now support per-package server options using the `moreServerOptions` config field, as well as options that apply to both the language server and `lean` using the `leanOptions` config field. Setting either of these fields instead of `moreServerArgs` ensures that viewing files from a dependency uses the options for that dependency. Additionally, `moreServerArgs` is being deprecated in favor of the `moreGlobalServerArgs` field. See PR [#2858](https://github.com/leanprover/lean4/pull/2858).
A Lakefile with the following deprecated package declaration:
```lean
def moreServerArgs := #[
"-Dpp.unicode.fun=true"
]
def moreLeanArgs := moreServerArgs
package SomePackage where
moreServerArgs := moreServerArgs
moreLeanArgs := moreLeanArgs
```
... can be updated to the following package declaration to use per-package options:
* [`pp.beta`` to apply beta reduction when pretty printing](https://github.com/leanprover/lean4/pull/2864).
* [Embed and check githash in .olean](https://github.com/leanprover/lean4/pull/2766).
* [Guess lexicographic order for well-founded recursion](https://github.com/leanprover/lean4/pull/2874).
* [Allow trailing comma in tuples, lists, and tactics](https://github.com/leanprover/lean4/pull/2643).
Bug fixes for [#2628](https://github.com/leanprover/lean4/issues/2628), [#2883](https://github.com/leanprover/lean4/issues/2883),
[#2810](https://github.com/leanprover/lean4/issues/2810), [#2925](https://github.com/leanprover/lean4/issues/2925), and [#2914](https://github.com/leanprover/lean4/issues/2914).
**Lake:**
* `lake init .` and a bare `lake init` and will now use the current directory as the package name. [#2890](https://github.com/leanprover/lean4/pull/2890)
* `lake new` and `lake init` will now produce errors on invalid package names such as `..`, `foo/bar`, `Init`, `Lean`, `Lake`, and `Main`. See issue [#2637](https://github.com/leanprover/lean4/issues/2637) and PR [#2890](https://github.com/leanprover/lean4/pull/2890).
* `lean_lib` no longer converts its name to upper camel case (e.g., `lean_lib bar` will include modules named `bar.*` rather than `Bar.*`). See issue [#2567](https://github.com/leanprover/lean4/issues/2567) and PR [#2889](https://github.com/leanprover/lean4/pull/2889).
* Lean and Lake now properly support non-identifier library names (e.g., `lake new 123-hello` and `import «123Hello»` now work correctly). See issue [#2865](https://github.com/leanprover/lean4/issues/2865) and PR [#2889](https://github.com/leanprover/lean4/pull/2888).
* Lake now filters the environment extensions loaded from a compiled configuration (`lakefile.olean`) to include only those relevant to Lake's workspace loading process. This resolves segmentation faults caused by environment extension type mismatches (e.g., when defining custom elaborators via `elab` in configurations). See issue [#2632](https://github.com/leanprover/lean4/issues/2632) and PR [#2896](https://github.com/leanprover/lean4/pull/2896).
* Cloud releases will now properly be re-unpacked if the build directory is removed. See PR [#2928](https://github.com/leanprover/lean4/pull/2928).
* Lake's `math` template has been simplified. See PR [#2930](https://github.com/leanprover/lean4/pull/2930).
* `lake exe <target>` now parses `target` like a build target (as the help text states it should) rather than as a basic name. For example, `lake exe @mathlib/runLinter` should now work. See PR [#2932](https://github.com/leanprover/lean4/pull/2932).
* `lake new foo.bar [std]` now generates executables named `foo-bar` and `lake new foo.bar exe` properly creates `foo/bar.lean`. See PR [#2932](https://github.com/leanprover/lean4/pull/2932).
* Later packages and libraries in the dependency tree are now preferred over earlier ones. That is, the later ones "shadow" the earlier ones. Such an ordering is more consistent with how declarations generally work in programming languages. This will break any package that relied on the previous ordering. See issue [#2548](https://github.com/leanprover/lean4/issues/2548) and PR [#2937](https://github.com/leanprover/lean4/pull/2937).
* Executable roots are no longer mistakenly treated as importable. They will no longer be picked up by `findModule?`. See PR [#2937](https://github.com/leanprover/lean4/pull/2937).
v4.3.0
---------
* `simp [f]` does not unfold partial applications of `f` anymore. See issue [#2042](https://github.com/leanprover/lean4/issues/2042).
To fix proofs affected by this change, use `unfold f` or `simp (config := { unfoldPartialApp := true }) [f]`.
* By default, `simp` will no longer try to use Decidable instances to rewrite terms. In particular, not all decidable goals will be closed by `simp`, and the `decide` tactic may be useful in such cases. The `decide` simp configuration option can be used to locally restore the old `simp` behavior, as in `simp (config := {decide := true})`; this includes using Decidable instances to verify side goals such as numeric inequalities.
* Many bug fixes:
* [Add left/right actions to term tree coercion elaborator and make `^`` a right action](https://github.com/leanprover/lean4/pull/2778)
* [Fix for #2775, don't catch max recursion depth errors](https://github.com/leanprover/lean4/pull/2790)
* [Reduction of `Decidable` instances very slow when using `cases` tactic](https://github.com/leanprover/lean4/issues/2552)
* [`simp` not rewriting in binder](https://github.com/leanprover/lean4/issues/1926)
* [`simp` unfolding `let` even with `zeta := false` option](https://github.com/leanprover/lean4/issues/2669)
* [`simp` (with beta/zeta disabled) and discrimination trees](https://github.com/leanprover/lean4/issues/2281)
* [unknown free variable introduced by `rw ... at h`](https://github.com/leanprover/lean4/issues/2711)
* [`dsimp` doesn't use `rfl` theorems which consist of an unapplied constant](https://github.com/leanprover/lean4/issues/2685)
* [`dsimp` does not close reflexive equality goals if they are wrapped in metadata](https://github.com/leanprover/lean4/issues/2514)
* [`rw [h]` uses `h` from the environment in preference to `h` from the local context](https://github.com/leanprover/lean4/issues/2729)
* [missing `withAssignableSyntheticOpaque` for `assumption` tactic](https://github.com/leanprover/lean4/issues/2361)
* [ignoring default value for field warning](https://github.com/leanprover/lean4/issues/2178)
* [Cancel outstanding tasks on document edit in the language server](https://github.com/leanprover/lean4/pull/2648).
* [Remove unnecessary `%` operations in `Fin.mod` and `Fin.div`](https://github.com/leanprover/lean4/pull/2688)
* [Avoid `DecidableEq` in `Array.mem`](https://github.com/leanprover/lean4/pull/2774)
* [Ensure `USize.size` unifies with `?m + 1`](https://github.com/leanprover/lean4/issues/1926)
* [Improve compatibility with emacs eglot client](https://github.com/leanprover/lean4/pull/2721)
**Lake:**
* [Sensible defaults for `lake new MyProject math`](https://github.com/leanprover/lean4/pull/2770)
* Changed `postUpdate?` configuration option to a `post_update` declaration. See the `post_update` syntax docstring for more information on the new syntax.
* [A manifest is automatically created on workspace load if one does not exists.](https://github.com/leanprover/lean4/pull/2680).
* The `:=` syntax for configuration declarations (i.e., `package`, `lean_lib`, and `lean_exe`) has been deprecated. For example, `package foo := {...}` is deprecated.
* [support for overriding package URLs via `LAKE_PKG_URL_MAP`](https://github.com/leanprover/lean4/pull/2709)
* Moved the default build directory (e.g., `build`), default packages directory (e.g., `lake-packages`), and the compiled configuration (e.g., `lakefile.olean`) into a new dedicated directory for Lake outputs, `.lake`. The cloud release build archives are also stored here, fixing [#2713](https://github.com/leanprover/lean4/issues/2713).
* Update manifest format to version 7 (see [lean4#2801](https://github.com/leanprover/lean4/pull/2801) for details on the changes).
* Deprecate the `manifestFile` field of a package configuration.
* There is now a more rigorous check on `lakefile.olean` compatibility (see [#2842](https://github.com/leanprover/lean4/pull/2842) for more details).
v4.2.0
---------
* [isDefEq cache for terms not containing metavariables.](https://github.com/leanprover/lean4/pull/2644).
* Make [`Environment.mk`](https://github.com/leanprover/lean4/pull/2604) and [`Environment.add`](https://github.com/leanprover/lean4/pull/2642) private, and add [`replay`](https://github.com/leanprover/lean4/pull/2617) as a safer alternative.
* `IO.Process.output` no longer inherits the standard input of the caller.
* [Do not inhibit caching](https://github.com/leanprover/lean4/pull/2612) of default-level `match` reduction.
* [List the valid case tags](https://github.com/leanprover/lean4/pull/2629) when the user writes an invalid one.
* The derive handler for `DecidableEq` [now handles](https://github.com/leanprover/lean4/pull/2591) mutual inductive types.
* [Show path of failed import in Lake](https://github.com/leanprover/lean4/pull/2616).
* [Fix linker warnings on macOS](https://github.com/leanprover/lean4/pull/2598).
* **Lake:** Add `postUpdate?` package configuration option. Used by a package to specify some code which should be run after a successful `lake update` of the package or one of its downstream dependencies. ([lake#185](https://github.com/leanprover/lake/issues/185))
* Improvements to Lake startup time ([#2572](https://github.com/leanprover/lean4/pull/2572), [#2573](https://github.com/leanprover/lean4/pull/2573))
* `refine e` now replaces the main goal with metavariables which were created during elaboration of `e` and no longer captures pre-existing metavariables that occur in `e` ([#2502](https://github.com/leanprover/lean4/pull/2502)).
* This is accomplished via changes to `withCollectingNewGoalsFrom`, which also affects `elabTermWithHoles`, `refine'`, `calc` (tactic), and `specialize`. Likewise, all of these now only include newly-created metavariables in their output.
* Previously, both newly-created and pre-existing metavariables occurring in `e` were returned inconsistently in different edge cases, causing duplicated goals in the infoview (issue [#2495](https://github.com/leanprover/lean4/issues/2495)), erroneously closed goals (issue [#2434](https://github.com/leanprover/lean4/issues/2434)), and unintuitive behavior due to `refine e` capturing previously-created goals appearing unexpectedly in `e` (no issue; see PR).
v4.1.0
---------
* The error positioning on missing tokens has been [improved](https://github.com/leanprover/lean4/pull/2393). In particular, this should make it easier to spot errors in incomplete tactic proofs.
* After elaborating a configuration file, Lake will now cache the configuration to a `lakefile.olean`. Subsequent runs of Lake will import this OLean instead of elaborating the configuration file. This provides a significant performance improvement (benchmarks indicate that using the OLean cuts Lake's startup time in half), but there are some important details to keep in mind:
+ Lake will regenerate this OLean after each modification to the `lakefile.lean` or `lean-toolchain`. You can also force a reconfigure by passing the new `--reconfigure` / `-R` option to `lake`.
+ Lake configuration options (i.e., `-K`) will be fixed at the moment of elaboration. Setting these options when `lake` is using the cached configuration will have no effect. To change options, run `lake` with `-R` / `--reconfigure`.
+ **The `lakefile.olean` is a local configuration and should not be committed to Git. Therefore, existing Lake packages need to add it to their `.gitignore`.**
* The signature of `Lake.buildO` has changed, `args` has been split into `weakArgs` and `traceArgs`. `traceArgs` are included in the input trace and `weakArgs` are not. See Lake's [FFI example](src/lake/examples/ffi/lib/lakefile.lean) for a demonstration of how to adapt to this change.
* The signatures of `Lean.importModules`, `Lean.Elab.headerToImports`, and `Lean.Elab.parseImports`
have [changed](https://github.com/leanprover/lean4/pull/2480) from taking `List Import` to `Array Import`.
* There is now [an `occs` field](https://github.com/leanprover/lean4/pull/2470)
in the configuration object for the `rewrite` tactic,
allowing control of which occurrences of a pattern should be rewritten.
This was previously a separate argument for `Lean.MVarId.rewrite`,
and this has been removed in favour of an additional field of `Rewrite.Config`.
It was not previously accessible from user tactics.
v4.0.0
---------
* [`Lean.Meta.getConst?` has been renamed](https://github.com/leanprover/lean4/pull/2454).
We have renamed `getConst?` to `getUnfoldableConst?` (and `getConstNoEx?` to `getUnfoldableConstNoEx?`).
These were not intended to be part of the public API, but downstream projects had been using them
(sometimes expecting different behaviour) incorrectly instead of `Lean.getConstInfo`.
* [`dsimp` / `simp` / `simp_all` now fail by default if they make no progress](https://github.com/leanprover/lean4/pull/2336).
This can be overridden with the `(config := { failIfUnchanged := false })` option.
This change was made to ease manual use of `simp` (with complicated goals it can be hard to tell if it was effective)
and to allow easier flow control in tactics internally using `simp`.
See the [summary discussion](https://leanprover.zulipchat.com/#narrow/stream/270676-lean4/topic/simp.20fails.20if.20no.20progress/near/380153295)
on zulip for more details.
* [`simp_all` now preserves order of hypotheses](https://github.com/leanprover/lean4/pull/2334).
In order to support the `failIfUnchanged` configuration option for `dsimp` / `simp` / `simp_all`
the way `simp_all` replaces hypotheses has changed.
In particular it is now more likely to preserve the order of hypotheses.
See [`simp_all` reorders hypotheses unnecessarily](https://github.com/leanprover/lean4/pull/2334).
(Previously all non-dependent propositional hypotheses were reverted and reintroduced.
Now only such hypotheses which were changed, or which come after a changed hypothesis,
are reverted and reintroduced.
This has the effect of preserving the ordering amongst the non-dependent propositional hypotheses,
but now any dependent or non-propositional hypotheses retain their position amongst the unchanged
non-dependent propositional hypotheses.)
This may affect proofs that use `rename_i`, `case ... =>`, or `next ... =>`.
`this` is now a regular identifier again that is implicitly introduced by anonymous `have :=` for the remainder of the tactic block. It used to be a keyword that was visible in all scopes and led to unexpected behavior when explicitly used as a binder name.
@@ -80,7 +827,7 @@ Unreleased
* New [code generator](https://github.com/leanprover/lean4/tree/master/src/Lean/Compiler/LCNF) project has started.
* Fix syntax hightlighting for recursive declarations. Example
* Fix syntax highlighting for recursive declarations. Example
```lean
inductive List (α : Type u) where
| nil : List α -- `List` is not highlighted as a variable anymore
@@ -909,7 +1656,7 @@ For example, given `f : Nat → Nat` and `g : Nat → Nat`, `f.comp g` is now no
* Various improvements to go-to-definition & find-all-references accuracy.
* Auto generated congruence lemmas with support for casts on proofs and `Decidable` instances (see [whishlist](https://github.com/leanprover/lean4/issues/988)).
* Auto generated congruence lemmas with support for casts on proofs and `Decidable` instances (see [wishlist](https://github.com/leanprover/lean4/issues/988)).
If any of the terms ``tᵢ`` in the template above contain a recursive call to ``foo``, the equation compiler tries to interpret the definition as a structural recursion. In order for that to succeed, the recursive arguments must be subterms of the corresponding arguments on the left-hand side. The function is then defined using a *course of values* recursion, using automatically generated functions ``below`` and ``brec`` in the namespace corresponding to the inductive type of the recursive argument. In this case the defining equations hold definitionally, possibly with additional case splits.
The case where patterns are matched against an argument whose type is an inductive family is known as *dependent pattern matching*. This is more complicated, because the type of the function being defined can impose constraints on the patterns that are matched. In this case, the equation compiler will detect inconsistent cases and rule them out.
```lean
universe u
inductive Vector (α : Type u) : Nat → Type u
| nil : Vector α 0
| cons : α → Vector α n → Vector α (n+1)
namespace Vector
def head : Vector α (n+1) → α
| cons h t => h
def tail : Vector α (n+1) → Vector α n
| cons h t => t
def map (f : α → β → γ) : Vector α n → Vector β n → Vector γ n
| nil, nil => nil
| cons a va, cons b vb => cons (f a b) (map f va vb)
end Vector
```
.. _recursive_functions:
Recursive functions
===================
Lean must ensure that a recursive function terminates, for which there are two strategies: _structural recursion_, in which all recursive calls are made on smaller parts of the input data, and _well-founded recursion_, in which recursive calls are justified by showing that arguments to recursive calls are smaller according to some other measure.
Structural recursion
--------------------
If the definition of a function contains recursive calls, Lean first tries to interpret the definition as a structural recursion. In order for that to succeed, the recursive arguments must be subterms of the corresponding arguments on the left-hand side.
The function is then defined using a *course of values* recursion, using automatically generated functions ``below`` and ``brec`` in the namespace corresponding to the inductive type of the recursive argument. In this case the defining equations hold definitionally, possibly with additional case splits.
If structural recursion fails, the equation compiler falls back on well-founded recursion. It tries to infer an instance of ``SizeOf`` for the type of each argument, and then show that each recursive call is decreasing under the lexicographic order of the arguments with respect to ``sizeOf`` measure. If it fails, the error message provides information as to the goal that Lean tried to prove. Lean uses information in the local context, so you can often provide the relevant proof manually using ``have`` in the body of the definition. In this case of well-founded recursion, the defining equations hold only propositionally, and can be accessed using ``simp`` and ``rewrite`` with the name ``foo``.
Well-founded recursion
---------------------
If structural recursion fails, the equation compiler falls back on well-founded recursion. It tries to infer an instance of ``SizeOf`` for the type of each argument, and then tries to find a permutation of the arguments such that each recursive call is decreasing under the lexicographic order with respect to ``sizeOf`` measures. Lean uses information in the local context, so you can often provide the relevant proof manually using ``have`` in the body of the definition.
In the case of well-founded recursion, the equation used to declare the function holds only propositionally, but not definitionally, and can be accessed using ``unfold``, ``simp`` and ``rewrite`` with the function name (for example ``unfold foo`` or ``simp [foo]``, where ``foo`` is the function defined with well-founded recursion).
```lean
namespace Hide
@@ -528,9 +569,53 @@ by rw [div]; rfl
end Hide
```
If Lean cannot find a permutation of the arguments for which all recursive calls are decreasing, it will print a table that contains, for every recursive call, which arguments Lean could prove to be decreasing. For example, a function with three recursive calls and four parameters might cause the following message to be printed
```
example.lean:37:0-43:31: error: Could not find a decreasing measure.
The arguments relate at each recursive call as follows:
(<, ≤, =: relation proved, ? all proofs failed, _: no proof attempted)
x1 x2 x3 x4
1) 39:6-27 = = _ =
2) 40:6-25 = ? _ <
3) 41:6-25 < _ _ _
Please use `termination_by` to specify a decreasing measure.
```
This table should be read as follows:
* In the first recursive call, in line 39, arguments 1, 2 and 4 are equal to the function's parameters.
* The second recursive call, in line 40, has an equal first argument, a smaller fourth argument, and nothing could be inferred for the second argument.
* The third recursive call, in line 41, has a decreasing first argument.
* No other proofs were attempted, either because the parameter has a type without a non-trivial ``WellFounded`` instance (parameter 3), or because it is already clear that no decreasing measure can be found.
Lean will print the termination argument it found if ``set_option showInferredTerminationBy true`` is set.
If Lean does not find the termination argument, or if you want to be explicit, you can append a `termination_by` clause to the function definition, after the function's body, but before the `where` clause if present. It is of the form
```
termination_by e
```
where ``e`` is an expression that depends on the parameters of the function and should be decreasing at each recursive call. The type of `e` should be an instance of the class ``WellFoundedRelation``, which determines how to compare two values of that type.
If ``f`` has parameters “after the ``:``” (for example when defining functions via patterns using `|`), then these can be brought into scope using the syntax
```
termination_by a₁ … aₙ => e
```
By default, Lean uses the tactic ``decreasing_tactic`` when proving that an argument is decreasing; see its documentation for how to globally extend it. You can also choose to use a different tactic for a given function definition with the clause
```
decreasing_by <tac>
```
which should come after ``termination_by`, if present.
Note that recursive definitions can in general require nested recursions, that is, recursion on different arguments of ``foo`` in the template above. The equation compiler handles this by abstracting later arguments, and recursively defining higher-order functions to meet the specification.
The equation compiler also allows mutual recursive definitions, with a syntax similar to that of [Mutual and Nested Inductive Definitions](#mutual-and-nested-inductive-definitions). They are compiled using well-founded recursion, and so once again the defining equations hold only propositionally.
Mutual recursion
----------------
The equation compiler also allows mutual recursive definitions, with a syntax similar to that of [Mutual and Nested Inductive Definitions](#mutual-and-nested-inductive-definitions). Mutual definitions are always compiled using well-founded recursion, and so once again the defining equations hold only propositionally.
```lean
mutual
@@ -587,29 +672,31 @@ def num_consts_lst : List Term → Nat
end
```
The case where patterns are matched against an argument whose type is an inductive family is known as *dependent pattern matching*. This is more complicated, because the type of the function being defined can impose constraints on the patterns that are matched. In this case, the equation compiler will detect inconsistent cases and rule them out.
In a set of mutually recursive function, either all or no functions must have an explicit termination argument (``termination_by``). A change of the default termination tactic (``decreasing_by``) only affects the proofs about the recursive calls of that function, not the other functions in the group.
```lean
universe u
```
mutual
theorem even_of_odd_succ : ∀ n, Odd (n + 1) → Even n
| _, odd_succ n h => h
termination_by n h => h
decreasing_by decreasing_tactic
inductive Vector (α : Type u) : Nat → Type u
| nil : Vector α 0
| cons : α → Vector α n → Vector α (n+1)
theorem odd_of_even_succ : ∀ n, Even (n + 1) → Odd n
| _, even_succ n h => h
termination_by n h => h
end
```
namespace Vector
Another way to express mutual recursion is using local function definitions in ``where`` or ``let rec`` clauses: these can be mutually recursive with each other and their containing function:
def head {α : Type} : Vector α (n+1) → α
| cons h t => h
def tail {α : Type} : Vector α (n+1) → Vector α n
| cons h t => t
def map {α β γ : Type} (f : α → β → γ) :
∀ {n}, Vector α n → Vector β n → Vector γ n
| 0, nil, nil => nil
| n+1, cons a va, cons b vb => cons (f a b) (map f va vb)
end Vector
```
theorem even_of_odd_succ : ∀ n, Odd (n + 1) → Even n
| _, odd_succ n h => h
termination_by n h => h
where
theorem odd_of_even_succ : ∀ n, Even (n + 1) → Odd n
@@ -11,6 +11,8 @@ There are two primary attributes for interoperating with other languages:
It can also be used with `def` to provide an internal definition, but ensuring consistency of both definitions is up to the user.
*`@[export sym] def leanSym : ...` exports `leanSym` under the unmangled symbol name `sym`.
For simple examples of how to call foreign code from Lean and vice versa, see <https://github.com/leanprover/lean4/blob/master/src/lake/examples/ffi> and <https://github.com/leanprover/lean4/blob/master/src/lake/examples/reverse-ffi>, respectively.
## The Lean ABI
The Lean Application Binary Interface (ABI) describes how the signature of a Lean declaration is encoded as a native calling convention.
@@ -109,6 +111,15 @@ if (lean_io_result_is_ok(res)) {
lean_io_mark_end_initialization();
```
In addition, any other thread not spawned by the Lean runtime itself must be initialized for Lean use by calling
```c
void lean_initialize_thread();
```
and should be finalized in order to free all thread-local resources by calling
```c
void lean_finalize_thread();
```
## `@[extern]` in the Interpreter
The interpreter can run Lean declarations for which symbols are available in loaded shared libraries, which includes `@[extern]` declarations.
@@ -119,4 +130,4 @@ Thus to e.g. run `#eval` on such a declaration, you need to
Note that it is not sufficient to load the foreign library containing the external symbol because the interpreter depends on code that is emitted for each `@[extern]` declaration.
Thus it is not possible to interpret an `@[extern]` declaration in the same file.
See `tests/compiler/foreign` for an example.
See [`tests/compiler/foreign`](https://github.com/leanprover/lean4/tree/master/tests/compiler/foreign/) for an example.
If you want to make changes to Lean itself, start by [building Lean](../make/index.html) from a clean checkout to make sure that everything is set up correctly.
If you want to make changes to Lean itself, start by [building Lean](../make/index.md) from a clean checkout to make sure that everything is set up correctly.
After that, read on below to find out how to set up your editor for changing the Lean source code, followed by further sections of the development manual where applicable such as on the [test suite](testing.md) and [commit convention](commit_convention.md).
If you are planning to make any changes that may affect the compilation of Lean itself, e.g. changes to the parser, elaborator, or compiler, you should first read about the [bootstrapping pipeline](bootstrap.md).
You can use `elantoolchain link` to give a specific stage build
directory a reference name, then use `elan override set` to associate
such a name to the current directory. We usually want to use `stage0`
for editing files in `src` and `stage1`for everything else (e.g.
tests).
The `lean-toolchain` files in the Lean 4 repository are set up to use the `lean4-stage0`
toolchain for editing files in `src` and the `lean4` toolchain for editing files in `tests`.
Run the following commands to make `lean4` point at `stage1` and `lean4-stage0`point at `stage0`:
```bash
# in the Lean rootdir
elan toolchain link lean4 build/release/stage1
elan toolchain link lean4-stage0 build/release/stage0
# make `lean` etc. point to stage1 in the rootdir and subdirs
elan override set lean4
cd src
# make `lean` etc. point to stage0 anywhere inside `src`
elan override set lean4-stage0
```
You can also use the `+toolchain` shorthand (e.g. `lean +lean4-debug`) to switch
@@ -57,3 +51,32 @@ You might find that debugging through elan, e.g. via `gdb lean`, disables some
things like symbol autocompletion because at first only the elan proxy binary
is loaded. You can instead pass the explicit path to `bin/lean` in your build
folder to gdb, or use `gdb $(elan which lean)`.
It is also possible to generate releases that others can use,
simply by pushing a tag to your fork of the Lean 4 github repository
(and waiting about an hour; check the `Actions` tab for completion).
If you push `my-tag` to a fork in your github account `my_name`,
you can then put `my_name/lean4:my-tag` in your `lean-toolchain` file in a project using `lake`.
(You must use a tag name that does not start with a numeral, or contain `_`).
### VS Code
There is a `lean.code-workspace` file that correctly sets up VS Code with workspace roots for the stage0/stage1 setup described above as well as with other settings.
You should always load it when working on Lean, such as by invoking
```
code lean.code-workspace
```
on the command line.
### `ccache`
Lean's build process uses [`ccache`](https://ccache.dev/) if it is
installed to speed up recompilation of the generated C code. Without
`ccache`, you'll likely spend more time than necessary waiting on
rebuilds - it's a good idea to make sure it's installed.
### `prelude`
Unlike most Lean projects, all submodules of the `Lean` module begin with the
`prelude` keyword. This disables the automated import of `Init`, meaning that
developers need to figure out their own subset of `Init` to import. This is done
such that changing files in `Init` doesn't force a full rebuild of `Lean`.
- In `RELEASES.md` remove `(development in progress)` from the `v4.7.0` section header.
- Our current goal is to have written release notes only about major language features or breaking changes,
and to rely on automatically generated release notes for bugfixes and minor changes.
- Do not wait on `RELEASES.md` being perfect before creating the `release/v4.7.0` branch. It is essential to choose the nightly which will become the release candidate as early as possible, to avoid confusion.
- If there are major changes not reflected in `RELEASES.md` already, you may need to solicit help from the authors.
- Minor changes and bug fixes do not need to be documented in `RELEASES.md`: they will be added automatically on the Github release page.
- Commit your changes to `RELEASES.md`, and push.
- Remember that changes to `RELEASES.md` after you have branched `releases/v4.7.0` should also be cherry-picked back to `master`.
- In `src/CMakeLists.txt`,
- verify that you see `set(LEAN_VERSION_MINOR 7)` (for whichever `7` is appropriate); this should already have been updated when the development cycle began.
- `set(LEAN_VERSION_IS_RELEASE 1)` (this should be a change; on `master` and nightly releases it is always `0`).
- Commit your changes to `src/CMakeLists.txt`, and push.
- `git tag v4.7.0-rc1`
- `git push origin v4.7.0-rc1`
- Now wait, while CI runs.
- You can monitor this at `https://github.com/leanprover/lean4/actions/workflows/ci.yml`, looking for the `v4.7.0-rc1` tag.
- This step can take up to an hour.
- Once the release appears at https://github.com/leanprover/lean4/releases/
- Edit the release notes on Github to select the "Set as a pre-release box".
- Copy the section of `RELEASES.md` for this version into the Github release notes.
- Use the title "Changes since v4.6.0 (from RELEASES.md)"
- Then in the "previous tag" dropdown, select `v4.6.0`, and click "Generate release notes".
- This will add a list of all the commits since the last stable version.
- Delete anything already mentioned in the hand-written release notes above.
- Delete "update stage0" commits, and anything with a completely inscrutable commit message.
- Briefly rearrange the remaining items by category (e.g. `simp`, `lake`, `bug fixes`),
but for minor items don't put any work in expanding on commit messages.
- (How we want to release notes to look is evolving: please update this section if it looks wrong!)
- Next, we will move a curated list of downstream repos to the release candidate.
- This assumes that there is already a *reviewed* branch `bump/v4.7.0` on each repository
containing the required adaptations (or no adaptations are required).
The preparation of this branch is beyond the scope of this document.
- For each of the target repositories:
- Checkout the `bump/v4.7.0` branch.
- Verify that the `lean-toolchain` is set to the nightly from which the release candidate was created.
- `git merge origin/master`
- Change the `lean-toolchain` to `leanprover/lean4:v4.7.0-rc1`
- In `lakefile.lean`, change any dependencies which were using `nightly-testing` or `bump/v4.7.0` branches
back to `master` or `main`, and run `lake update` for those dependencies.
- Run `lake build` to ensure that dependencies are found (but it's okay to stop it after a moment).
- `git commit`
- `git push`
- Open a PR from `bump/v4.7.0` to `master`, and either merge it yourself after CI, if appropriate,
or notify the maintainers that it is ready to go.
- Once this PR has been merged, tag `master` with `v4.7.0-rc1` and push this tag.
- We do this for the same list of repositories as for stable releases, see above.
As above, there are dependencies between these, and so the process above is iterative.
It greatly helps if you can merge the `bump/v4.7.0` PRs yourself!
- For Std/Aesop/Mathlib, which maintain a `nightly-testing` branch, make sure there is a tag
`nightly-testing-2024-02-29` with date corresponding to the nightly used for the release
(create it if not), and then on the `nightly-testing` branch `git reset --hard master`, and force push.
- Make an announcement!
This should go in https://leanprover.zulipchat.com/#narrow/stream/113486-announce, with topic `v4.7.0-rc1`.
Please see previous announcements for suggested language.
You will want a few bullet points for main topics from the release notes.
Please also make sure that whoever is handling social media knows the release is out.
- Begin the next development cycle (i.e. for `v4.8.0`) on the Lean repository, by making a PR that:
- Updates `src/CMakeLists.txt` to say `set(LEAN_VERSION_MINOR 8)`
- Removes `(in development)` from the section heading in `RELEASES.md` for `v4.7.0`,
and creates a new `v4.8.0 (in development)` section heading.
## Time estimates:
Slightly longer than the corresponding steps for a stable release.
Similar process, but more things go wrong.
In particular, updating the downstream repositories is significantly more work
(because we need to merge existing `bump/v4.7.0` branches, not just update a toolchain).
# Preparing `bump/v4.7.0` branches
While not part of the release process per se,
this is a brief summary of the work that goes into updating Std/Aesop/Mathlib to new versions.
@@ -5,7 +5,6 @@ After [building Lean](../make/index.md) you can run all the tests using
cd build/release
make test ARGS=-j4
```
Change the 4 to the maximum number of parallel tests you want to
allow. The best choice is the number of CPU cores on your machine as
the tests are mostly CPU bound. You can find the number of processors
@@ -17,6 +16,12 @@ adding the `-C stageN` argument. The default when run as above is stage 1. The
Lean tests will automatically use that stage's corresponding Lean
executables
Running `make test` will not pick up new test files; run
```bash
cmake build/release/stage1
```
to update the list of tests.
You can also use `ctest` directly if you are in the right folder. So
to run stage1 tests with a 300 second timeout run this:
@@ -24,6 +29,9 @@ to run stage1 tests with a 300 second timeout run this:
cd build/release/stage1
ctest -j 4 --output-on-failure --timeout 300
```
Useful `ctest` flags are `-R <name of test>` to run a single test, and
`--rerun-failed` to run all tests that failed during the last run.
You can also pass `ctest` flags via `make test ARGS="--rerun-failed"`.
To get verbose output from ctest pass the `--verbose` command line
option. Test output is normally suppressed and only summary
@@ -33,17 +41,17 @@ information is displayed. This option will show all test output.
All these tests are included by [src/shell/CMakeLists.txt](https://github.com/leanprover/lean4/blob/master/src/shell/CMakeLists.txt):
-`tests/lean`: contains tests that come equipped with a
.lean.expected.out file. The driver script `test_single.sh` runs
-[`tests/lean`](https://github.com/leanprover/lean4/tree/master/tests/lean/): contains tests that come equipped with a
.lean.expected.out file. The driver script [`test_single.sh`](https://github.com/leanprover/lean4/tree/master/tests/lean/test_single.sh) runs
each test and checks the actual output (*.produced.out) with the
checked in expected output.
-`tests/lean/run`: contains tests that are run through the lean
-[`tests/lean/run`](https://github.com/leanprover/lean4/tree/master/tests/lean/run/): contains tests that are run through the lean
command line one file at a time. These tests only look for error
codes and do not check the expected output even though output is
produced, it is ignored.
-`tests/lean/interactive`: are designed to test server requests at a
-[`tests/lean/interactive`](https://github.com/leanprover/lean4/tree/master/tests/lean/interactive/): are designed to test server requests at a
given position in the input file. Each .lean file contains comments
that indicate how to simulate a client request at that position.
using a `--^` point to the line position. Example:
@@ -53,7 +61,7 @@ All these tests are included by [src/shell/CMakeLists.txt](https://github.com/le
Bla.
--^ textDocument/completion
```
In this example, the test driver `test_single.sh` will simulate an
In this example, the test driver [`test_single.sh`](https://github.com/leanprover/lean4/tree/master/tests/lean/interactive/test_single.sh) will simulate an
auto-completion request at `Bla.`. The expected output is stored in
a .lean.expected.out in the json format that is part of the
[Language Server
@@ -70,23 +78,33 @@ All these tests are included by [src/shell/CMakeLists.txt](https://github.com/le
--^ collectDiagnostics
```
- `tests/lean/server`: Tests more of the Lean `--server` protocol.
- [`tests/lean/server`](https://github.com/leanprover/lean4/tree/master/tests/lean/server/): Tests more of the Lean `--server` protocol.
There are just a few of them, and it uses .log files containing
JSON.
- `tests/compiler`: contains tests that will run the Lean compiler and
- [`tests/compiler`](https://github.com/leanprover/lean4/tree/master/tests/compiler/): contains tests that will run the Lean compiler and
build an executable that is executed and the output is compared to
the .lean.expected.out file. This test also contains a subfolder
`foreign` which shows how to extend Lean using C++.
[`foreign`](https://github.com/leanprover/lean4/tree/master/tests/compiler/foreign/) which shows how to extend Lean using C++.
- `tests/lean/trust0`: tests that run Lean in a mode that Lean doesn't
- [`tests/lean/trust0`](https://github.com/leanprover/lean4/tree/master/tests/lean/trust0): tests that run Lean in a mode that Lean doesn't
- `tests/plugin`: tests that compiled Lean code can be loaded into
- [`tests/plugin`](https://github.com/leanprover/lean4/tree/master/tests/plugin/): tests that compiled Lean code can be loaded into
`lean` via the `--plugin` command line option.
## Writing Good Tests
Every test file should contain:
* an initial `/-! -/` module docstring summarizing the test's purpose
* a module docstring for each test section that describes what is tested
and, if not 100% clear, why that is the desirable behavior
At the time of writing, most tests do not follow these new guidelines yet.
For an example of a conforming test, see [`tests/lean/1971.lean`](https://github.com/leanprover/lean4/tree/master/tests/lean/1971.lean).
## Fixing Tests
When the Lean source code or the standard library are modified, some of the
@@ -101,7 +119,7 @@ First, we must install [meld](http://meldmerge.org/). On Ubuntu, we can do it by
sudo apt-get install meld
```
Now, suppose `bad_class.lean` test is broken. We can see the problem by going to `test/lean` directory and
Now, suppose `bad_class.lean` test is broken. We can see the problem by going to [`tests/lean`](https://github.com/leanprover/lean4/tree/master/tests/lean) directory and
executing
```
@@ -114,8 +132,3 @@ outputs. `meld` can also be used to repair the problems.
In Emacs, we can also execute `M-x lean4-diff-test-file` to check/diff the file of the current buffer.
To mass-copy all `.produced.out` files to the respective `.expected.out` file, use `tests/lean/copy-produced`.
When using the Nix setup, add `--keep-failed` to the `nix build` call and then call
@@ -5,7 +5,7 @@ If the type of keys can be totally ordered -- that is, it supports a well-behave
then maps can be implemented with binary search trees (BSTs). Insert and lookup operations on BSTs take time
proportional to the height of the tree. If the tree is balanced, the operations therefore take logarithmic time.
This example is based on a similar example found in the ["Sofware Foundations"](https://softwarefoundations.cis.upenn.edu/vfa-current/SearchTree.html)
This example is based on a similar example found in the ["Software Foundations"](https://softwarefoundations.cis.upenn.edu/vfa-current/SearchTree.html)
book (volume 3).
-/
@@ -81,9 +81,9 @@ def Tree.toList (t : Tree β) : List (Nat × β) :=
|>.toList
/-!
The implemention of `Tree.toList` is inefficient because of how it uses the `++` operator.
The implementation of `Tree.toList` is inefficient because of how it uses the `++` operator.
On a balanced tree its running time is linearithmic, because it does a linear number of
concatentations at each level of the tree. On an unbalanced tree it's quadratic time.
concatenations at each level of the tree. On an unbalanced tree it's quadratic time.
Here's a tail-recursive implementation than runs in linear time, regardless of whether the tree is balanced:
-/
defTree.toListTR(t:Treeβ):List(Nat×β):=
@@ -114,9 +114,9 @@ concatenating all goals produced by `tac'`. In this theorem, we use it to apply
The `simp` parameters `toListTR.go` and `toList` instruct the simplifier to try to reduce
and/or apply auto generated equation theorems for these two functions.
The parameter `*` intructs the simplifier to use any equation in a goal as rewriting rules.
The parameter `*` instructs the simplifier to use any equation in a goal as rewriting rules.
In this particular case, `simp` uses the induction hypotheses as rewriting rules.
Finally, the parameter `List.append_assoc` intructs the simplifier to use the
Finally, the parameter `List.append_assoc` instructs the simplifier to use the
@@ -7,15 +7,6 @@ Lean is a new open source theorem prover being developed at Microsoft Research.
It is a research project that aims to bridge the gap between interactive and automated theorem proving.
Lean can be also used as a programming language. Actually, some Lean features are implemented in Lean itself.
### Are pull requests welcome?
In the past, we accepted most pull requests. This practice produced hard to maintain code, performance problems, and bugs.
It takes time to review a pull request and make sure it is correct, useful and is not in conflict with our plans.
Small bug fixes (few lines of code) are always welcome. Any other kind of unrequested pull request is not.
Thus, before implementing a feature or modifying the system, please ask whether the change is welcome or not.
We have issues tagged with ["help wanted"](https://github.com/leanprover/lean4/issues?q=is%3Aissue+is%3Aopen+label%3A%22help+wanted%22), if you want to contribute to the project, please take a look at them.
If you are interested in one of them, post comments, ask questions, and engage with the core developers there.
### Should I use Lean?
Lean is under heavy development, and we are constantly trying new
@@ -36,7 +27,7 @@ It is a good place to interact with other Lean users.
### Should I use Lean to teach a course?
Lean has been used to teach courses on logic, type theory and programming languages at CMU and the University of Washington.
The lecture notes for the CMU course [Logic and Proof](https://leanprover.github.io/logic_and_proof) are available online,
The lecture notes for the CMU course [Logic and Proof](https://lean-lang.org/logic_and_proof) are available online,
but they are for Lean 3.
If you decide to teach a course using Lean, we suggest you prepare all material before the beginning of the course, and
make sure that Lean attends all your needs. You should not expect we will fix bugs and/or add features needed for your course.
@@ -56,7 +47,7 @@ We expect similar independent checkers will be built for Lean 4.
We use [GitHub](https://github.com/leanprover/lean4/issues) to track bugs and new features.
Bug reports are always welcome, but nitpicking issues are not (e.g., the error message is confusing).
See also our [contribution guidelines](../CONTRIBUTING.md).
See also our [contribution guidelines](https://github.com/leanprover/lean4/blob/master/CONTRIBUTING.md).
The goal of [this book](https://leanprover.github.io/functional_programming_in_lean/) is to be an accessible introduction to using Lean 4 as a programming language.
The goal of [this book](https://lean-lang.org/functional_programming_in_lean/) is to be an accessible introduction to using Lean 4 as a programming language.
It should be useful both to people who want to use Lean as a general-purpose programming language and to mathematicians who want to develop larger-scale proof automation but do not have a background in functional programming.
It does not assume any background with functional programming, though it's probably not a good first book on programming in general.
New content will be added once per month until it's done.
By default, Lean only accepts total functions. The `partial` keyword should be used when Lean cannot
establish that a function always terminates.
By default, Lean only accepts total functions.
The `partial` keyword may be used to define a recursive function without a termination proof; `partial` functions compute in compiled programs, but are opaque in proofs and during type checking.
While [Nix](https://nixos.org/nix/) can be used to quickly open a shell with all dependencies for the [standard setup](index.md) installed, the user-facing [Nix Setup](../setup.md#nix-setup) can also be used to work *on* Lean.
## Setup
Follow the setup in the link above; to open the Lean shell inside a Lean checkout, you can also use
```bash
# in the Lean root directory
$ nix-shell -A nix
```
On top of the local and remote Nix cache, we do still rely on CCache as well to make C/C++ build steps incremental, which are atomic steps from Nix's point of view.
To enable CCache, add the following line to the config file mentioned in the setup:
```bash
extra-sandbox-paths = /nix/var/cache/ccache
```
Then set up that directory as follows:
```bash
sudo mkdir -m0770 -p /nix/var/cache/ccache
# macOS standard chown doesn't support --reference
From the Lean root directory inside the Lean shell:
```bash
nix build .#stage1 # build this stage's stdlib & executable
nix build .#stage1.test # run all tests
nix run .#stage1.update-stage0 # update ./stage0 from this stage
nix run .#stage1.update-stage0-commit # ...and commit the results
```
The `stage1.` part in each command is optional:
```bash
nix build .#test # run tests for stage 1
nix build . # build stage 1
nix build # ditto
```
## Build Process Description
The Nix build process conceptually works the same as described in [Lean Build Pipeline](index.md#lean-build-pipeline).
However, there are two important differences in practice apart from the standard Nix properties (hermeneutic, reproducible builds stored in a global hash-indexed store etc.):
* Only files tracked by git (using `git add` or at least `git add --intent-to-add`) are compiled.
This is actually a general property of Nix flakes, and has the benefit of making it basically impossible to forget to commit a file (at least in `src/`).
* Only files reachable from `src/Lean.lean` are compiled.
This is because modules are discovered not from a directory listing anymore but by recursively compiling all dependencies of that top module.
## Editor Integration
As in the standard Nix setup.
After adding `src/` as an LSP workspace, it should automatically fall back to using stage 0 in there.
Note that the UX of `{emacs,vscode}-dev` is quite different from the Make-based setup regarding the compilation of dependencies:
there is no mutable directory incrementally filled by the build that we could point the editor at for .olean files.
Instead, `emacs-dev` will gather the individual dependency outputs from the Nix store when checking a file -- and build them on the fly when necessary.
However, it will only ever load changes saved to disk, not ones opened in other buffers.
The absence of a mutable output directory also means that the Lean server will not automatically pick up `.ilean` metadata from newly compiled files.
Instead, you can run `nix run .#link-ilean` to symlink the `.ilean` tree of the stdlib state at that point in time to `src/build/lib`, where the server should automatically find them.
## Other Fun Stuff to Do with Nix
Open Emacs with Lean set up from an arbitrary commit (without even cloning Lean beforehand... if your Nix is new enough):
```bash
nix run github:leanprover/lean4/7e4edeb#emacs-package
```
Open a shell with `lean` and `LEAN_PATH` set up for compiling a specific module (this is exactly what `emacs-dev` is doing internally):
```bash
nix develop .#mods.\"Lean.Parser.Basic\"
# alternatively, directly pass a command to execute:
1. Launch VS Code and install the `lean4` extension.
1. Launch VS Code and install the `lean4` extension by clicking on the "Extensions" sidebar entry and searching for "lean4".
1.Create a new file using "File > New Text File" (`Ctrl+N`). Click the `Select a language` prompt, type in `lean4`, and hit ENTER. You should see the following popup:
1.Open the Lean 4 setup guide by creating a new text file using "File > New Text File" (`Ctrl+N`), clicking on the ∀-symbol in the top right and selecting "Documentation… > Setup: Show Setup Guide".
Click the "Install Lean using Elan" button. You should see some progress output like this:
```
info: syncing channel updates for 'nightly'
info: latest update on nightly, lean version nightly-2023-06-27
info: downloading component 'lean'
```
If there is no popup, you probably have Elan installed already.
You may want to make sure that your default toolchain is Lean 4 in this case by running `elan default leanprover/lean4:nightly` and reopen the file, as the next step will fail otherwise.
1. Follow the Lean 4 setup guide. It will walk you through learning resources for Lean 4, teach you how to set up Lean's dependencies on your platform, install Lean 4 for you at the click of a button and help you set up your first project.
1. While it is installing, you can paste the following Lean program into the new file:
```lean
#eval Lean.versionString
```
When the installation has finished, the Lean Language Server should start automatically and you should get syntax-highlighting and a "Lean Infoview" popping up on the right. You will see the output of the `#eval` statement when
you place your cursor at the end of the statement.
You are set up!
## Create a Lean Project
*If your goal is to contribute to [mathlib4](https://github.com/leanprover-community/mathlib4) or use it as a depdency, please see its readme for specific instructions on how to do that.*
You can now create a Lean project in a new folder. Run `lake init foo` from "View > Terminal" to create a package, followed by `lake build` to get an executable version of your Lean program.
On Linux/macOS, you first have to follow the instructions printed by the Lean installation or log out and in again for the Lean executables to be available in you terminal.
Note: Packages **have** to be opened using "File > Open Folder..." for imports to work.
Saved changes are visible in other files after running "Lean 4: Refresh File Dependencies" (`Ctrl+Shift+X`).
## Troubleshooting
**The InfoView says "Waiting for Lean server to start..." forever.**
Check that the VS Code Terminal is not showing some installation errors from `elan`.
If that doesn't work, try also running the VS Code command `Developer: Reload Window`.
The Lean language server provides semantic highlighting information to editors. In order to benefit from this in VSCode, you may need to activate the "Editor > Semantic Highlighting" option in the preferences (this is translates to `"editor.semanticHighlighting.enabled": true,`
in `settings.json`). The default option here is to let your color theme decides whether it activates semantic highlighting (the default themes Dark+ and Light+ do activate it for instance).
However this may be insufficient if your color theme does not distinguish enough syntax categories or distinguishes them very subtly. For instance the default Light+ theme uses color `#001080` for variables. This is awfully close to `#000000` that is used as the default text color. This makes it very easy to miss an accidental use of [auto bound implicit arguments](https://leanprover.github.io/lean4/doc/autobound.html). For instance in
However this may be insufficient if your color theme does not distinguish enough syntax categories or distinguishes them very subtly. For instance the default Light+ theme uses color `#001080` for variables. This is awfully close to `#000000` that is used as the default text color. This makes it very easy to miss an accidental use of [auto bound implicit arguments](https://lean-lang.org/lean4/doc/autobound.html). For instance in
Platforms built & tested by our CI, available as nightly releases via elan (see below)
Platforms built & tested by our CI, available as binary releases via elan (see below)
* x86-64 Linux with glibc 2.27+
* x86-64 macOS 10.15+
@@ -10,13 +10,15 @@ Platforms built & tested by our CI, available as nightly releases via elan (see
### Tier 2
Platforms cross-compiled but not tested by our CI, available as nightly releases
Platforms cross-compiled but not tested by our CI, available as binary releases
Releases may be silently broken due to the lack of automated testing.
Issue reports and fixes are welcome.
* aarch64 Linux with glibc 2.27+
* aarch64 (Apple Silicon) macOS
* x86 (32-bit) Linux
* Emscripten Web Assembly
<!--
### Tier 3
@@ -32,8 +34,8 @@ Release builds for all supported platforms are available at <https://github.com/
Instead of downloading these and setting up the paths manually, however, it is recommended to use the Lean version manager [`elan`](https://github.com/leanprover/elan) instead:
```sh
$ elan self update # in case you haven't updated elan in a while
An alternative setup based on Nix provides a perfectly reproducible development environment for your project from the Lean version down to the editor and Lean extension.
However, it is still experimental and subject to change; in particular, it is heavily based on an unreleased version of Nix enabling [Nix Flakes](https://www.tweag.io/blog/2020-05-25-flakes/). The setup has been tested on NixOS, other Linux distributions, and macOS.
After installing (any version of) Nix (<https://nixos.org/download.html>), you can easily open a shell with the particular pre-release version of Nix needed by and tested with our setup (called the "Lean shell" from here on):
```bash
$ nix-shell https://github.com/leanprover/lean4/archive/master.tar.gz -A nix
```
While this shell is sufficient for executing the steps below, it is recommended to also set the following options in `/etc/nix/nix.conf` (`nix.extraOptions` in NixOS):
```
max-jobs = auto # Allow building multiple derivations in parallel
keep-outputs = true # Do not garbage-collect build time-only dependencies (e.g. clang)
# Allow fetching build results from the Lean Cachix cache
On a multi-user installation of Nix (the default), you need to restart the Nix daemon afterwards:
```bash
sudo pkill nix-daemon
```
The [Cachix](https://cachix.org/) integration will magically beam any build steps already executed by the CI right onto your machine when calling Nix commands in the shell opened above.
It can be set up analogously as a cache for your own project.
Note: Your system Nix might print warnings about not knowing some of the settings used by the Lean shell Nix, which can be ignored.
## Basic Commands
From a Lean shell, run
```bash
$ nix flake new mypkg -t github:leanprover/lean4
```
to create a new Lean package in directory `mypkg` using the latest commit of Lean 4.
Such packages follow the same directory layout as described in the standard setup, except for a `lakefile.lean` replaced by a `flake.nix` file set up so you can run Nix commands on it, for example:
$ nix run .#emacs-dev # open a pinned version of Emacs with lean4-mode fully set up
$ nix run .#emacs-dev MyPackage.lean # arguments can be passed as well, e.g. the file to open
$ nix run .#vscode-dev MyPackage.lean # ditto, using VS Code
```
Note that if you rename `MyPackage.lean`, you also have to adjust the `name` attribute in `flake.nix` accordingly.
Also note that if you turn the package into a Git repository, only tracked files will be visible to Nix.
As in the standard setup, changes need to be saved to be visible in other files, which have then to be invalidated via an editor command.
If you don't want to or cannot start the pinned editor from Nix, e.g. because you're running Lean inside WSL/a container/on a different machine, you can manually point your editor at the `lean` wrapper script the commands above use internally:
```bash
$ nix build .#lean-dev -o result-lean-dev
```
The resulting `./result-lean-dev/bin/lean` script essentially runs `nix run .#lean` in the current project's root directory when you open a Lean file or use the "refresh dependencies" command such that the correct Lean version for that project is executed.
This includes selecting the correct stage of Lean (which it will compile on the fly, though without progress output) if you are [working on Lean itself](./make/nix.md#editor-integration).
Package dependencies can be added as further input flakes and passed to the `deps` list of `buildLeanPackage`. Example: <https://github.com/Kha/testpkg2/blob/master/flake.nix#L5>
For hacking, it can be useful to temporarily override an input with a local checkout/different version of a dependency:
On a build error, Nix will show the last 10 lines of the output by default. You can pass `-L` to `nix build` to show all lines, or pass the shown `*.drv` path to `nix log` to show the full log after the fact.
Keeping all outputs ever built on a machine alive can accumulate to quite impressive amounts of disk space, so you might want to trigger the Nix GC when `/nix/store/` has grown too large:
```bash
nix-collect-garbage
```
This will remove everything not reachable from "GC roots" such as the `./result` symlink created by `nix build`.
Note that the package information in `flake.nix` is currently completely independent from `lakefile.lean` used in the standard setup.
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
