- Fixes#3030
The `ComputeTypeInfo` transformation was incorrectly assuming that the
type of the body couldn't refer to the let item. When inferring the type
of a let, we now inline the let item(s) into the type of the body.
```
NLet Let {..} ->
let bodyTy = Info.getNodeType _letBody
in subst (_letItem ^. letItemValue) bodyTy
```
* Closes#2968
* Implements detection of function-like definitions, which either:
- have some arguments on the left of `:`, or
- have at least one clause.
* Only function-like definitions are recursive.
* Non-recursive definitions are not mutually recursive either, and can
be used only after their definition. This necessitates rearranging some
definitions in existing Juvix code.
* Changes the scoping of identifiers in record updates. Now field names
on the right side don't refer to the old values of the record fields but
to identifiers in scope defined outside the record update. To refer to
old values, one needs to explicitly use record projections, e.g.
```
r@Rec{x := Rec.x r}
```
* Closes#3039
* Closes#3043
* Closes#2970
* Closes#3089
* Parser allows trailing semicolons for any kind of semicolon-separated
items:
- let-block statements,
- module statements,
- record declaration statements,
- record update fields,
- record pattern fields,
- named application arguments,
- list literal items,
- list pattern items,
- open statement using/hiding items,
- `syntax iterator` declaration parameters,
- `syntax fixity` declaration parameters.
* Formatter prints trailing semicolons if the items are displayed on
separate lines, removes them if on a single line.
* The formatting of multiline lists is changed to make it consistent
with other semicolon-separated blocks:
```
[
1;
2;
3;
]
```
instead of
```
[ 1
; 2
; 3
]
```
# Changes
1. Adds a new command `juvix dev anoma node`. This command runs the
anoma node.
2. Adds a flag `--anoma-dir` to `juvix dev nockma run`. When given, it
must point to the anoma clone. Then, it will run the nockma code in the
anoma node and report the result (with no traces).
# Prerequisites
1. An anoma clone at some specific commit.
```
git clone git@github.com:anoma/anoma.git
cd anoma
git checkout 98e3660b91cd55f1d9424dcff9420425ae98f5f8
# build anoma
mix deps.get
mix escript.install hex protobuf
mix compile
# build the client
mix do --app anoma_client escript.build
```
2. The `mix` command (elixir).
3. The [`grpcurl`](https://github.com/fullstorydev/grpcurl) command. To
install a single binary in `~/.local/bin` you can run:
```
curl -sSL
"https://github.com/fullstorydev/grpcurl/releases/download/v1.9.1/grpcurl_1.9.1_linux_x86_64.tar.gz"
| tar -xz -C ~/.local/bin --no-wildcards grpcurl
```
# Testing
I've not included any test. It can be tested locally like this:
```
cd juvix/tests/Anoma/Compilation/positive
juvix compile anoma test001.juvix
echo 20 > args.debug.nockma
juvix dev nockma run --anoma-dir ~/projects/anoma test001.nockma --args args.debug.nockma
* Updates the standard library to
https://github.com/anoma/juvix-stdlib/pull/130
* Also changes `null` to `isEmpty`, which required updating some tests
---------
Co-authored-by: Paul Cadman <git@paulcadman.dev>
* Closes#3079
* Closes#3086
* Depends on #3088
* Updates the coding style guidelines (CODING.md) to reflect issues not
foreseen originally
* Changes the unicode arrow printed in the REPL to `->`. This is to make
the output consistent with how function types are written in the
standard library.
---------
Co-authored-by: Paul Cadman <git@paulcadman.dev>
* Closes#2804
* Requires #3003
* Front-end syntax for side conditions was implemented in #2852. This PR
implements compilation of side conditions.
* Adds side-conditions to `Match` nodes in Core. Updates Core parsing,
printing and the evaluator.
* Only side-conditions without an `else` branch are allowed in Core. If
there is an `else` branch, the side conditions are translated in
`fromInternal` into nested ifs. Because with `else` the conditions are
exhaustive, there are no implications for pattern exhaustiveness
checking.
* Adjusts the "wildcard row" case in the pattern matching compilation
algorithm to take into account the side conditions.
* Specialization has become less effective after recent changes to the
codebase. This PR fixes issues with specialization.
* Closes#2939
* Closes#2945
Checklist
---------
- [X] Preserve pragmas for letrec and lambda in Stored Core
- [x] Remove the assumption that all type variables are at the front
(closes#2945)
- [x] Allow specialization when the argument is a constructor
application
- [x] Make renaming adjust pragmas
- [x] Allow pragmas for fields in record definitions (closes#2939)
- [x] Update standard library pragmas
- [x] Fix JuvixTree printing
This PR adds `Byte` as a builtin with builtin functions for equality,
`byte-from-nat` and `byte-to-nat`. The standard library is updated to
include this definition with instances for `FromNatural`, `Show` and
`Eq` traits.
The `FromNatural` trait means that you can assign `Byte` values using
non-negative numeric literals.
You can use byte literals in jvc files by adding the u8 suffix to a
numeric value. For example, 1u8 represents a byte literal.
Arithmetic is not supported as the intention is for this type to be used
to construct ByteArrays of data where isn't not appropriate to modify
using arithmetic operations. We may add a separate `UInt8` type in the
future which supports arithmetic.
The Byte is supported in the native, rust and Anoma backend. Byte is not
supported in the Cairo backend because `byte-from-nat` cannot be
defined.
The primitive builtin ops for `Byte` are called `OpUInt8ToInt` and
`OpUInt8FromInt`, named because these ops work on integers and in future
we may reuse these for a separate unsigned 8-bit integer type that
supports arithmetic.
Part of:
* https://github.com/anoma/juvix/issues/2865
- Closes#2923
This pr fixes a bug where all fields were assigned to be explicit
arguments in the NameSignature Builder. A single line change was enough
to fix it.
```diff
- RecordStatementField RecordField {..} -> addSymbol @s Explicit Nothing _fieldName _fieldType
+ RecordStatementField RecordField {..} -> addSymbol @s (fromIsImplicitField _fieldIsImplicit) Nothing _fieldName _fieldType
```
I've also added a compilation test for instance fields.
- Closes#2668
This pr migrates the old named application syntax to the new one. In
order to migrate a juvix file to the new syntax it suffices to run the
formatter.
After the next release, we should completely remove the support for the
old syntax.
## Other changes
I've improved Scope negative tests. Previously, when a negative test
failed, you could only see the title of the test and the message
"Incorrect Error", as well as the Haskell file and line where the test
is defined.
This is extremely incovenient because you have to go to the haskell test
file, go to the line where the error is defined, look at the name of the
file and then visit that file. Moreover, you need to manually run the
scoper on that file to see the error that was returned.
I've fixed that and it now shows all relevant information. Example:
I've implemented this only using the `Generic` instance for the
`ScoperError` type, so doing something similar for the rest of negative
tests should be straightforward.
This pr explores the option to implement error handling in Juvix à la
mtl. It adds the following as a test:
1. `MonadError` trait.
2. `MonadTrans` trait.
3. `ExceptT` monad transformer and its `Functor`, `Monad`, `MonadTrans`,
`MonadError` instances.
- This PR adds a temporary compiler error for when the bug #2247
happens. We do not have plans to fix this bug until we move the
typechecker to Core, so it makes sense to add a better error message.
This PR implements changes to make the `eval` command and internal
development commands fully Cairo-compatible.
* Change the default field size to Cairo field size
* Change the printing of "negative" field elements to be compatible with
the Cairo VM
* Quote function names in the Reg to CASM translation
* Closes#2571
* It is reasonable to finish this PR before tackling #2562, because the
field element type is the primary data type in Cairo.
* Depends on #2653
Checklist
---------
- [x] Add field type and operations to intermediate representations
(JuvixCore, JuvixTree, JuvixAsm, JuvixReg).
- [x] Add CLI option to choose field size.
- [x] Add frontend field builtins.
- [x] Automatic conversion of integer literals to field elements.
- [x] Juvix standard library support for fields.
- [x] Check if field size matches when loading a stored module.
- [x] Update the Cairo Assembly (CASM) interpreter to use the field type
instead of integer type.
- [x] Add field type to VampIR backend.
- [x] Tests
---------
Co-authored-by: Jan Mas Rovira <janmasrovira@gmail.com>
This PR fixes an issue with importing `syntax alias` and using them from
another module:
Say you have a module defining a syntax alias:
```
module SyntaxAlias;
import Stdlib.Prelude open;
syntax alias MyNat := Nat;
```
and another which imports / uses it:
```
module test073;
import SyntaxAlias open;
main : MyNat := 11;
```
The compiler crashed with the following error:
```
^?!): empty Fold
CallStack (from HasCallStack):
error, called at src/Lens/Micro.hs:711:28 in microlens-0.4.13.1-ARwI8t2x86cAxRs56XPcG1:Lens.Micro
^?!, called at src/Juvix/Compiler/Concrete/Translation/FromParsed/Analysis/Scoping.hs:565:29 in juvix-0.5.5-G7jC6MbkbsJGkMT9u4BkYQ:Juvix.Compiler.Concrete.Translation.FromParsed.Analysis.Scoping
```
We fix this by adding the aliases to the store's [scoper info
table](6649209d26/src/Juvix/Compiler/Store/Scoped/Data/InfoTable.hs (L14))
and then use it to initialise the scoper state when scoping a module.
This PR creates a new package that's bundled with the compiler in a
similar way to the stdlib and the package description package.
## The `package-base` Package
This package is called
[package-base](ab4376cf9e/include/package-base)
and contains the minimal set of definitions required to load a Package
file.
The
[`Juvix.Builtin`](ab4376cf9e/include/package-base/Juvix/Builtin/V1.juvix)
module contains:
```
module Juvix.Builtin.V1;
import Juvix.Builtin.V1.Nat open public;
import Juvix.Builtin.V1.Trait.Natural open public;
import Juvix.Builtin.V1.String open public;
import Juvix.Builtin.V1.Bool open public;
import Juvix.Builtin.V1.Maybe open public;
import Juvix.Builtin.V1.List open public;
import Juvix.Builtin.V1.Fixity open public;
```
`Juvix.Builtin.V1.Bool` is required to support backend primitive
integers `Juvix.Builtin.V1.Trait.Natural` is required to support numeric
literals.
## The `PackageDescription.V2` module
This PR also adds a new
[`PackageDescription.V2`](ab4376cf9e/include/package/PackageDescription/V2.juvix)
type that uses the `package-base`. This is to avoid breaking existing
Package files. The Packages files in the repo (except those that test
`PackageDescription.V1`) have also been updated.
## Updating the stdlib
The standard library will be updated to use `Juvix.Builtin.*` modules in
a subsequent PR.
* Part of https://github.com/anoma/juvix/issues/2511
This pr applies a number of fixes to the new typechecker.
The fixes implemented are:
1. When guessing the arity of the body, we properly use the type
information of the variables in the patterns.
2. When generating wildcards, we name them properly so that they align
with the name in the type signature.
3. When compiling named applications, we inline all clauses of the form
`fun : _ := body`. This is a workaround to
https://github.com/anoma/juvix/issues/2247 and
https://github.com/anoma/juvix/issues/2517
4. I've had to ignore test027 (Church numerals). While the typechecker
passes and one can see that the types are correct, there is a lambda
where its clauses have different number of patterns. Our goal is to
support that in the near future
(https://github.com/anoma/juvix/issues/1706). This is the conflicting
lambda:
```
mutual num : Nat → Num
:= λ : Nat → Num {| (zero : Nat) := czero
| ((suc n : Nat)) {A} := csuc (num n) {A}}
```
5. I've added non-trivial a compilation test involving monad
transformers.
This PR adds the `PackageDescription.Basic` module, available to
Package.juvix files.
```
module Package;
import PackageDescription.Basic open;
package : Package := basicPackage;
```
The `PackageDescription.Basic` module provides a Package type that is
translated to a Juvix Package with all default arguments. It is not
possible to customize a basic package.
A basic package does not depend on the standard library, so loads much
more quickly.
Additionally this PR:
* Adds `juvix init --basic/-b` option to generate a basic Package.juvix.
* Migrates Package.juvix files that only use default arguments, or only
customise the name field, to basic Package files.
* Closes https://github.com/anoma/juvix/issues/2508
This PR:
* Modifies entry point `_entryPointBuildDir` to use the `BuildDir` type
instead of `SomeBase Dir`. This allows delayed resolution of the default
build directory which was useful for the Package -> Concrete translation
point below.
* Modifies `juvix dev root` to render the current package as a
Package.juvix file.
* Modifies the Package -> Concrete translation to recognise default
arguments. So, for example, an empty `juvix.yaml` file will be
translated into the following (instead of the `name`, `version`, and
`dependencies` arguments being populated).
module Package;
import Stdlib.Prelude open;
import PackageDescription.V1 open;
package : Package := defaultPackage;
* Adds a temporary command (removed when juvix.yaml support is removed)
`juvix dev migrate-juvix-yaml` that translates `juvix.yaml` into an
equivalent `Package.juvix` in the current project.
* Adds a temporary script `migrate-juvix-yaml.sh` (removed when
juvix.yaml support is removed) which can be run in the project to
translate all Juvix projects in the repository.
* Actually translate all of the `juvix.yaml` files to `Package.juvix`
using the script.
* Part of https://github.com/anoma/juvix/issues/2487
* Closes#2365
* Implements the syntax `f@{x1 := def1; ...; xn := defn}` and `f@?{x1 :=
def1; ..; xn := defn}`. Each definition inside the `@{..}` is an
ordinary function definition. The `@?` version allows partial
application (not all explicit named arguments need to be provided). This
subsumes the old record creation syntax.
* Closes#2453
* Closes#2432
* Any nonnegative literal `n` is replaced with `fromNat {_} {{_}} n`
where `fromNat` is the builtin conversion function defined in the
`Natural` trait in `Stdlib.Trait.Natural`.
* Any negative literal `-n` is replaced with `fromInt {_} {{_}} -n`
where `fromInt` is the builtin conversion function defined in the
`Integral` trait in `Stdlib.Trait.Integral`.
* Before resolving instance holes, it is checked whether the type holes
introduced for `fromNat` and `fromInt` have been inferred. If not, an
attempt is made to unify them with `Nat` or `Int`. This allows to
type-check e.g. `1 == 1` (there is no hint in the context as to what the
type of `1` should be, so it is decided to be `Nat` after inferring the
hole fails).
If you have `juvix.yaml` file in a parent of the juvix source directory
in your filesystem then some Compilation tests will fail. This is
because the tests assume that the project root will be in the test
directory.
This PR just adds `juvix.yaml` files to those test directories to avoid
this issue.
Spotted by @agureev
Simplifies arithmetic expressions in the Core optimization phase,
changing e.g. `(x - 1) + 1` to `x`. Such expressions appear as a result
of compiling pattern matching on natural numbers.
* Closes#2154
* Evaluates closed applications with value arguments when the result
type is zero-order. For example, `3 + 4` is evaluated to `7`, and `id 3`
is evaluated to `3`, but `id id` is not evaluated because the target
type is not zero-order (it's a function type).
* Closes#2426
A coercion from trait `T` to `T'` can be declared with the syntax
```
coercion instance
coeName {A} {{T A}} : T' A := ...
```
Coercions can be seen as instances with special resolution rules.
Coercion resolution rules
-------------------------
* If a non-coercion instance can be applied in a single instance
resolution step, no coercions are considered. No ambiguity results if
there exists some coercion which could be applied, but a non-coercion
instance exists - the non-coercion instances have priority.
* If no non-coercion instance can be applied in a single resolution
step, all minimal coercion paths which lead to an applicable
non-coercion instance are considered. If there is more than one,
ambiguity is reported.
Examples
----------
The following type-checks because:
1. There is no non-coercion instance found for `U String`.
2. There are two minimal coercion paths `U` <- `U1` and `U` <- `U2`, but
only one of them (`U` <- `U2`) ends in an applicable non-coercion
instance (`instU2` for `U2 String`).
```
trait
type U A := mkU {pp : A -> A};
trait
type U1 A := mkU1 {pp : A -> A};
trait
type U2 A := mkU2 {pp : A -> A};
coercion instance
fromU1toU {A} {{U1 A}} : U A :=
mkU@{
pp := U1.pp
};
coercion instance
fromU2toU {A} {{U2 A}} : U A :=
mkU@{
pp := U2.pp
};
instance
instU2 : U2 String := mkU2 id;
main : IO := printStringLn (U.pp "X")
```
The following results in an ambiguity error because:
1. There is no non-coercion instance found for `T Unit`.
2. There are two minimal coercion paths `T` <- `T1` and `T` <- `T2`,
both of which end in applicable non-coercion instances.
```
trait
type T A := mkT { pp : A → A };
trait
type T1 A := mkT1 { pp : A → A };
trait
type T2 A := mkT2 { pp : A → A };
instance
unitT1 : T1 Unit := mkT1 (pp := λ{_ := unit});
instance
unitT2 : T2 Unit := mkT2 (pp := λ{_ := unit});
coercion instance
fromT1toT {A} {{T1 A}} : T A := mkT@{
pp := T1.pp
};
coercion instance
fromT2toT {A} {{T2 A}} : T A := mkT@{
pp := T2.pp
};
main : Unit := T.pp unit;
```
The following type-checks, because there exists a non-coercion instance
for `T2 String`, so the coercion `fromT1toT2` is ignored during instance
resolution.
```
trait
type T1 A := mkT1 {pp : A -> A};
trait
type T2 A := mkT2 {pp : A -> A};
instance
instT1 {A} : T1 A :=
mkT1@{
pp := id
};
coercion instance
fromT1toT2 {A} {{M : T1 A}} : T2 A :=
mkT2@{
pp := T1.pp {{M}}
};
instance
instT2 : T2 String :=
mkT2@{
pp (s : String) : String := s ++str "!"
};
main : String := T2.pp "a";
```
* Introduces the `inline: case` pragma which causes an identifier to be
inlined if it is matched on. This is necessary to support GEB without
compromising optimization for other targets.
* Adapts to the new commits in
https://github.com/anoma/juvix-stdlib/pull/86
* Adapts to https://github.com/anoma/juvix-stdlib/pull/86
* Adds a pass in `toEvalTransformations` to automatically inline all
record projection functions, regardless of the optimization level. This
is necessary to ensure that arithmetic operations and comparisons on
`Nat` or `Int` are always represented directly with the corresponding
built-in Core functions. This is generally highly desirable and required
for the Geb target.
* Adds the `inline: always` pragma which indicates that a function
should always be inlined during the mandatory inlining phase, regardless
of optimization level.
* Closes#2416
* Closes#2401
* Avoids generating identical specialisations by keeping a
specialisation signature for each specialised function application.
* Allows to specialise on a per-trait or per-instance basis:
```
{-# specialize: true #-}
trait
type Natural N := mkNatural {
+ : N -> N -> N;
* : N -> N -> N;
fromNat : Nat -> N;
};
```
or
```
{-# specialize: true #-}
instance
naturalNatI : Natural Nat := ...
```
* The above `specialize: bool` pragma actually works with any type or
function. To be able to simultaneously specify the boolean
specialisation flag and specialisation arguments, one can use
`specialize-args: [arg1, .., argn]` which works like `specialize: [arg1,
.., argn]`.
