This PR records extra mod uses that previously caused wrong unnecessary
import reports from shake.
---------
Co-authored-by: Sebastian Ullrich <sebasti@nullri.ch>
This PR fixes one potential source of inlay hint flakiness.
In the old `IO.waitAny` implementation, we could rely on the fact that
if all tasks in the list were finished, `IO.waitAny` would pick the
first finished one. In the new implementation (#9732), this isn't the
case anymore for fairness reasons, but this also means that in
`IO.AsyncList.getFinishedPrefixWithTimeout`, it can happen that we don't
scan the full finished command snapshot prefix because we pick the
timeout task before the finished snapshot task. This is likely the cause
of a flaky test failure
[here](https://github.com/leanprover/lean4/actions/runs/18215430028/job/51863870111),
where the inlay hint test yielded no result (the timeout task has an
edit delay of 0ms in the first inlay hint request that is emitted,
finishes immediately and can thus immediately cause the finished prefix
to be skipped with the new `waitAny` implementation).
This PR fixes this issue by adding a `hasFinished` check before the
`waitAny` to ensure that we always scan the finished prefix and don't
need to rely on a brittle invariant that doesn't hold anymore. It also
converts some `Task.get`s to `IO.wait` for safety so that the compiler
can't re-order them.
This PR significantly improves the test coverage of the language server,
providing at least a single basic test for every request that is used by
the client. It also implements infrastructure for testing all of these
requests, e.g. the ability to run interactive tests in a project context
and refactors the interactive test runner to be more maintainable.
Finally, it also fixes a small bug with the recently implemented unknown
identifier code actions for auto-implicits (#10442) that was discovered
in testing, where the "import all unambiguous unknown identifiers" code
action didn't work correctly on auto-implicit identifiers.
This PR fixes a bug in the unknown identifier code actions where it
would yield non-sensical suggestions for nested `open` declarations like
`open Foo.Bar`.
This PR fixes deadlocking `exit` calls in the language server.
We have previously observed deadlocking calls to `exit` inside of the
language server and deemed them irrelevant. However, child processes of
these deadlocking exiting processes can continue to consume a large
amount of CPU as they try to compile a library etc. Hence, this PR
switches to the MT safe `_Exit` inside of the language server,
in order to ensure the server finishes when it is told to.
This PR implements module docstrings in Verso syntax, as well as adding
a number of improvements and fixes to Verso docstrings in general. In
particular, they now have language server support and are parsed at
parse time rather than elaboration time, so the snapshot's syntax tree
includes the parsed documentation.
This PR refines and clarifies the `meta` phase distinction in the module
system.
* `meta import A` without `public` now has the clarified meaning of
"enable compile-time evaluation of declarations in or above `A` in the
current module, but not downstream". This is now checked statically by
enforcing that public meta defs, which therefore may be referenced from
outside, can only use public meta imports, and that global evaluating
attributes such as `@[term_parser]` can only be applied to public meta
defs.
* `meta def`s may no longer reference non-meta defs even when in the
same module. This clarifies the meta distinction as well as improves
locality of (new) error messages.
* parser references in `syntax` are now also properly tracked as meta
references.
* A `meta import` of an `import` now properly loads only the `.ir` of
the nested module for the purposes of execution instead of also making
its declarations available for general elaboration.
* `initialize` is now no longer being run on import under the module
system, which is now covered by `meta initialize`.
This PR upstreams the Verso parser and adds preliminary support for
Verso in docstrings. This will allow the compiler to check examples and
cross-references in documentation.
After a `stage0` update, a follow-up PR will add the appropriate
attributes that allow the feature to be used. The parser tests from
Verso also remain to be upstreamed, and user-facing documentation will
be added once the feature has been used on more internals.
This PR speeds up auto-completion by a factor of ~3.5x through various
performance improvements in the language server. On one machine, with
`import Mathlib`, completing `i` used to take 3200ms and now instead
yields a result in 920ms.
Specifically, the following improvements are made:
- The watchdog process no longer de-serializes and re-serializes most
messages from the file worker before passing them on to the user - a
fast partial de-serialization procedure is now used to determine whether
the message needs to be de-serialized in full or not.
- `escapePart` is optimized to perform better on ASCII strings that do
not need escaping.
- `Json.compress` is optimized to allocate fewer objects.
- A faster JSON compression specifically for completion responses is
implemented that skips allocating `Json` altogether.
- The JSON compression has been moved to the task where we convert a
request response to `Json` so that converting to a string won't block
the output task of the FileWorker and so the `Json` value is not marked
as multi-threaded when we compress is, which drastically increases the
cost of reference-counting.
- The JSON representation of the `data?` field of each completion item
is optimized.
- Both the completion kind and the set of completion tags for each
imported completion item is now cached.
- The filtering of duplicate completion items is optimized.
Other adjustments:
- `LT UInt8` and `LE UInt8` are moved to Prelude so that they can be
used in `Init.Meta` for the name part escaping fast path.
- `Array.usize` is exposed since it was marked as `@[simp]`.
This PR ports more of the post-initialization C++ shell code to Lean.
All that remains is the initialization of the profiler and task manager.
As initialization tasks rather than main shell code, they were left in
C++ (where the rest of the initialization code currently is).
The `max_memory` and `timeout` Lean options used by the the `--memory`
and `--timeout` command-line options are now properly registered. The
server defaults for max memory and max heartbeats (timeout) were removed
as they were not actually used (because the `server` option that was
checked was neither set nor exists).
This PR also makes better use of the module system in `Shell.lean` and
fixes a minor bug in a previous port where the file name check was
dependent on building the `.ilean` rather than the `.c` file (as was
originally the case).
Fixes#9879.
This PR fixes a bug that caused the Lean server process tree to survive
the closing of VS Code.
The cause of this issue was that the file worker main task was blocked
on waiting for the result of `lake setup-file` because the blocking call
was lifted outside of the dedicated server task that was supposed to
contain it by the compiler.
This PR allows Lean's parser to run with a final position prior to the
end of the string, so it can be invoked on a sub-region of the input.
This has applications in Verso proper, which parses Lean syntax in
contexts such as code blocks and docstrings, and it is a prerequisite to
parsing the contents of Lean docstrings.
This PR addresses a missing check in the module system where private
names that remain in the public environment map for technical reasons
(e.g. inductive constructors generated by the kernel and relied on by
the code generator) accidentally were accessible in the public scope.
This PR re-implements `IO.waitAny` using Lean instead of C++. This is to
reduce the size and
complexity of `task_manager` in order to ease future refactorings.
There is an import behavioral change of `IO.waitAny` in this PR.
Consider a situation where we have
two promises `p1`, `p2` and call `IO.waitAny [p1.result!, p2.result!]`
and `p1` resolves instantly.
Previously this would just return the result of `p1` and require nothing
else. With the new
implementation if `p2` is released before being resolved this can cause
a panic, even if
`IO.waitAny` has already finished. I argue that this is reasonable
behavior, given that an
invocation of `result!` promises that the promise will eventually be
resolved.
This PR adjusts the import graph, primarily of `Lean`, such that the
worst case rebuild time of core (`lean` only) is below 3 minutes on the
speedcenter machine (not captured by benchmark yet).
This PR changes `lake setup-file` to use the server-provided header for
workspace modules.
This also reverts #9163 as the underlying issue is now fixed.
This PR modifies dot identifier notation so that `(.a : T)` resolves
`T.a` with respect to the root namespace, like for generalized field
notation. This lets the notation refer to private names, follow aliases,
and also use open namespaces. The LSP completions are improved to follow
how dot ident notation is resolved, but it doesn't yet take into account
aliases or open namespaces.
Closes#9629
This PR consolidates common attribute-related error messages into
reusable functions and updates the wording and formatting of relevant
error messages.
This PR fixes an issue that caused some `deriving` handlers to fail when
the name of the type being declared matched that of a declaration in an
open namespace.
Closes#9366
This PR improves the 'Go to Definition' UX, specifically:
- Using 'Go to Definition' on a type class projection will now extract
the specific instances that were involved and provide them as locations
to jump to. For example, using 'Go to Definition' on the `toString` of
`toString 0` will yield results for `ToString.toString` and `ToString
Nat`.
- Using 'Go to Definition' on a macro that produces syntax with type
class projections will now also extract the specific instances that were
involved and provide them as locations to jump to. For example, using
'Go to Definition' on the `+` of `1 + 1` will yield results for
`HAdd.hAdd`, `HAdd α α α` and `Add Nat`.
- Using 'Go to Declaration' will now provide all the results of 'Go to
Definition' in addition to the elaborator and the parser that were
involved. For example, using 'Go to Declaration' on the `+` of `1 + 1`
will yield results for `HAdd.hAdd`, `HAdd α α α`, `Add Nat`,
``macro_rules | `($x + $y) => ...`` and `infixl:65 " + " => HAdd.hAdd`.
- Using 'Go to Type Definition' on a value with a type that contains
multiple constants will now provide 'Go to Definition' results for each
constant. For example, using 'Go to Type Definition' on `x` for `x :
Array Nat` will yield results for `Array` and `Nat`.
### Details
'Go to Definition' for type class projections was first implemented by
#1767, but there were still a couple of shortcomings with the
implementation. E.g. in order to jump to the instance in `toString 0`,
one had to add another space within the application and then use 'Go to
Definition' on that, or macros would block instances from being
displayed. Then, when the .ilean format was added, most 'Go to
Definition' requests were already handled using the .ileans in the
watchdog process, and so the file worker never received them to handle
them with the semantic information that it has available.
This PR resolves most of the issues with the previous implementation and
refactors the 'Go to Definition' control flow so that 'Go to Definition'
requests are always handled by the file worker, with the watchdog merely
using its .ilean position information to update the positions in the
response to a more up-to-date state. This is necessary because the file
worker obtains its position information from the .oleans, which need to
be rebuilt in order to be up-to-date, while the watchdog always receives
.ilean update notifications from each active file worker with the
current position information in the editor.
Finally, all of the 'Go to Definition' code is refactored to be easier
to maintain.
### Breaking changes
`InfoTree.hoverableInfoAt?` has been generalized to
`InfoTree.hoverableInfoAtM?` and now takes a general `filter` argument
instead of several boolean flags, as was the case before.
This PR removes uses of `Lean.RBMap` in Lean itself.
Furthermore some massaging of the import graph is done in order to avoid
having `Std.Data.TreeMap.AdditionalOperations` (which is quite
expensive) be the critical path for a large chunk of Lean. In particular
we can build `Lean.Meta.Simp` and `Lean.Meta.Grind` without it thanks to
these changes.
We did previously not conduct this change as `Std.TreeMap` was not
outperforming `Lean.RBMap` yet, however this has changed with the new
code generator.
This PR migrates usages of `Std.Range` to the new polymorphic ranges.
This PR unfortunately increases the transitive imports for
frequently-used parts of `Init` because the ranges now rely on iterators
in order to provide their functionality for types other than `Nat`.
However, iteration over ranges in compiled code is as efficient as
before in the examples I checked. This is because of a special
`IteratorLoop` implementation provided in the PR for this purpose.
There were two issues that were uncovered during migration:
* In `IndPredBelow.lean`, migrating the last remaining range causes
`compilerTest1.lean` to break. I have minimized the issue and came to
the conclusion it's a compiler bug. Therefore, I have not replaced said
old range usage yet (see #9186).
* In `BRecOn.lean`, we are publicly importing the ranges. Making this
import private should theoretically work, but there seems to be a
problem with the module system, causing the build to panic later in
`Init.Data.Grind.Poly` (see #9185).
* In `FuzzyMatching.lean`, inlining fails with the new ranges, which
would have led to significant slowdown. Therefore, I have not migrated
this file either.
This PR disables the use of the header produced by `lake setup-file` in
the server for now. It will be re-enabled once Lake takes into account
the header given by the server when processing workspace modules.
Without that, `setup-file` header can produce odd behavior when the file
on disk and in an editor disagree on whether the file participates in
the module system.
This PR adjusts the experimental module system to make `private` the
default visibility modifier in `module`s, introducing `public` as a new
modifier instead. `public section` can be used to revert the default for
an entire section, though this is more intended to ease gradual adoption
of the new semantics such as in `Init` (and soon `Std`) where they
should be replaced by a future decl-by-decl re-review of visibilities.
This PR replaces all usages of `[:]` slice notation in `src` with the
new `[...]` notation in production code, tests and comments. The
underlying implementation of the `Subarray` functions stays the same.
Notation cheat sheet:
* `*...*` is the doubly-unbounded range.
* `*...a` or `*...<a` contains all elements that are less than `a`.
* `*...=a` contains all elements that are less than or equal to `a`.
* `a...*` contains all elements that are greater than or equal to `a`.
* `a...b` or `a...<b` contains all elements that are greater than or
equal to `a` and less than `b`.
* `a...=b` contains all elements that are greater than or equal to `a`
and less than or equal to `b`.
* `a<...*` contains all elements that are greater than `a`.
* `a<...b` or `a<...<b` contains all elements that are greater than `a`
and less than `b`.
* `a<...=b` contains all elements that are greater than `a` and less
than or equal to `b`.
Benchmarks have shown that importing the iterator-backed parts of the
polymorphic slice library in `Init` impacts build performance. This PR
avoids this problem by separating those parts of the library that do not
rely on iterators from those those that do. Whereever the new slice
notation is used, only the iterator-independent files are imported.
This PR fixes a bug where semantic highlighting would only highlight
keywords that started with an alphanumeric character. Now, it uses
`Lean.isIdFirst`.
This PR adds support to the server for the new module setup process by
changing how `lake setup-file` is used.
In the new server setup, `lake setup-file` is invoked with the file name
of the edited module passed as a CLI argument and with the parsed header
passed to standard input in JSON form. Standard input is used to avoid
potentially exceeding the CLI length limits on Windows. Lake will build
the module's imports along with any other dependencies and then return
the module's workspace configuration via JSON (now in the form of
`ModuleSetup`). The server then post-processes this configuration a bit
and returns it back to the Lean language processor.
The server's header is currently only fully respected by Lake for
external modules (files that are not part of any workspace library). For
workspace modules, the saved module header is currently used to build
imports (as has been done since #7909). A follow-up Lake PR will align
both cases to follow the server's header.
Lean search paths (e.g., `LEAN_PATH`, `LEAN_SRC_PATH`) are no longer
negotiated between the server and Lake. These environment variables are
already configured during sever setup by `lake serve` and do not change
on a per-file basis. Lake can also pre-resolve the `.olean` files of
imports via the `importArts` field of `ModuleSetup`, limiting the
potential utility of communicating `LEAN_PATH`.
This PR adds support for throwing named errors with associated error
explanations. In particular, it adds elaborators for the syntax defined
in #8649, which use the error-explanation infrastructure added in #8651.
This includes completions, hovers, and jump-to-definition for error
names.
Note that another stage0 rebuild will be required to define explanations
using `register_error_explanation`.
---------
Co-authored-by: Joachim Breitner <mail@joachim-breitner.de>
Co-authored-by: Marc Huisinga <mhuisi@protonmail.com>
