Iterate on CST traversal

This commit is contained in:
Charlie Marsh 2022-09-02 15:34:51 -04:00
parent 21d39a9b44
commit 0cfa2d617a
7 changed files with 1016 additions and 732 deletions

View File

@ -4,7 +4,7 @@ use rustpython_parser::ast::{
PatternKind, Stmt, StmtKind, Unaryop, Withitem,
};
pub trait Visitor {
pub trait ASTVisitor {
fn visit_stmt(&mut self, stmt: &Stmt) {
walk_stmt(self, stmt);
}
@ -61,7 +61,7 @@ pub trait Visitor {
}
}
pub fn walk_stmt<V: Visitor + ?Sized>(visitor: &mut V, stmt: &Stmt) {
pub fn walk_stmt<V: ASTVisitor + ?Sized>(visitor: &mut V, stmt: &Stmt) {
match &stmt.node {
StmtKind::FunctionDef { args, body, .. } => {
visitor.visit_arguments(args);
@ -238,7 +238,7 @@ pub fn walk_stmt<V: Visitor + ?Sized>(visitor: &mut V, stmt: &Stmt) {
}
}
pub fn walk_expr<V: Visitor + ?Sized>(visitor: &mut V, expr: &Expr) {
pub fn walk_expr<V: ASTVisitor + ?Sized>(visitor: &mut V, expr: &Expr) {
match &expr.node {
ExprKind::BoolOp { op, values } => {
visitor.visit_boolop(op);
@ -399,7 +399,7 @@ pub fn walk_expr<V: Visitor + ?Sized>(visitor: &mut V, expr: &Expr) {
}
}
pub fn walk_constant<V: Visitor + ?Sized>(visitor: &mut V, constant: &Constant) {
pub fn walk_constant<V: ASTVisitor + ?Sized>(visitor: &mut V, constant: &Constant) {
if let Constant::Tuple(constants) = constant {
for constant in constants {
visitor.visit_constant(constant)
@ -407,7 +407,7 @@ pub fn walk_constant<V: Visitor + ?Sized>(visitor: &mut V, constant: &Constant)
}
}
pub fn walk_comprehension<V: Visitor + ?Sized>(visitor: &mut V, comprehension: &Comprehension) {
pub fn walk_comprehension<V: ASTVisitor + ?Sized>(visitor: &mut V, comprehension: &Comprehension) {
visitor.visit_expr(&comprehension.target, None);
visitor.visit_expr(&comprehension.iter, None);
for expr in &comprehension.ifs {
@ -415,7 +415,7 @@ pub fn walk_comprehension<V: Visitor + ?Sized>(visitor: &mut V, comprehension: &
}
}
pub fn walk_excepthandler<V: Visitor + ?Sized>(visitor: &mut V, excepthandler: &Excepthandler) {
pub fn walk_excepthandler<V: ASTVisitor + ?Sized>(visitor: &mut V, excepthandler: &Excepthandler) {
match &excepthandler.node {
ExcepthandlerKind::ExceptHandler { type_, body, .. } => {
if let Some(expr) = type_ {
@ -428,7 +428,7 @@ pub fn walk_excepthandler<V: Visitor + ?Sized>(visitor: &mut V, excepthandler: &
}
}
pub fn walk_arguments<V: Visitor + ?Sized>(visitor: &mut V, arguments: &Arguments) {
pub fn walk_arguments<V: ASTVisitor + ?Sized>(visitor: &mut V, arguments: &Arguments) {
for arg in &arguments.posonlyargs {
visitor.visit_arg(arg);
}
@ -452,24 +452,24 @@ pub fn walk_arguments<V: Visitor + ?Sized>(visitor: &mut V, arguments: &Argument
}
}
pub fn walk_arg<V: Visitor + ?Sized>(visitor: &mut V, arg: &Arg) {
pub fn walk_arg<V: ASTVisitor + ?Sized>(visitor: &mut V, arg: &Arg) {
if let Some(expr) = &arg.node.annotation {
visitor.visit_annotation(expr)
}
}
pub fn walk_keyword<V: Visitor + ?Sized>(visitor: &mut V, keyword: &Keyword) {
pub fn walk_keyword<V: ASTVisitor + ?Sized>(visitor: &mut V, keyword: &Keyword) {
visitor.visit_expr(&keyword.node.value, None);
}
pub fn walk_withitem<V: Visitor + ?Sized>(visitor: &mut V, withitem: &Withitem) {
pub fn walk_withitem<V: ASTVisitor + ?Sized>(visitor: &mut V, withitem: &Withitem) {
visitor.visit_expr(&withitem.context_expr, None);
if let Some(expr) = &withitem.optional_vars {
visitor.visit_expr(expr, None);
}
}
pub fn walk_match_case<V: Visitor + ?Sized>(visitor: &mut V, match_case: &MatchCase) {
pub fn walk_match_case<V: ASTVisitor + ?Sized>(visitor: &mut V, match_case: &MatchCase) {
visitor.visit_pattern(&match_case.pattern);
if let Some(expr) = &match_case.guard {
visitor.visit_expr(expr, None);
@ -479,7 +479,7 @@ pub fn walk_match_case<V: Visitor + ?Sized>(visitor: &mut V, match_case: &MatchC
}
}
pub fn walk_pattern<V: Visitor + ?Sized>(visitor: &mut V, pattern: &Pattern) {
pub fn walk_pattern<V: ASTVisitor + ?Sized>(visitor: &mut V, pattern: &Pattern) {
match &pattern.node {
PatternKind::MatchValue { value } => visitor.visit_expr(value, None),
PatternKind::MatchSingleton { value } => visitor.visit_constant(value),
@ -527,24 +527,24 @@ pub fn walk_pattern<V: Visitor + ?Sized>(visitor: &mut V, pattern: &Pattern) {
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_expr_context<V: Visitor + ?Sized>(visitor: &mut V, expr_context: &ExprContext) {}
pub fn walk_expr_context<V: ASTVisitor + ?Sized>(visitor: &mut V, expr_context: &ExprContext) {}
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_boolop<V: Visitor + ?Sized>(visitor: &mut V, boolop: &Boolop) {}
pub fn walk_boolop<V: ASTVisitor + ?Sized>(visitor: &mut V, boolop: &Boolop) {}
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_operator<V: Visitor + ?Sized>(visitor: &mut V, operator: &Operator) {}
pub fn walk_operator<V: ASTVisitor + ?Sized>(visitor: &mut V, operator: &Operator) {}
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_unaryop<V: Visitor + ?Sized>(visitor: &mut V, unaryop: &Unaryop) {}
pub fn walk_unaryop<V: ASTVisitor + ?Sized>(visitor: &mut V, unaryop: &Unaryop) {}
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_cmpop<V: Visitor + ?Sized>(visitor: &mut V, cmpop: &Cmpop) {}
pub fn walk_cmpop<V: ASTVisitor + ?Sized>(visitor: &mut V, cmpop: &Cmpop) {}
#[allow(unused_variables)]
#[inline(always)]
pub fn walk_alias<V: Visitor + ?Sized>(visitor: &mut V, alias: &Alias) {}
pub fn walk_alias<V: ASTVisitor + ?Sized>(visitor: &mut V, alias: &Alias) {}

View File

@ -1,14 +0,0 @@
use crate::message::Message;
use crate::settings::Settings;
use crate::{cache, fs};
use libcst_native::parse_module;
pub fn autofix(contents: &str, messages: &[Message]) {
// Parse the module.
let mut m = match parse_module(&contents, None) {
Ok(m) => m,
Err(e) => panic!("foo"),
};
m.body
}

View File

@ -7,11 +7,11 @@ use rustpython_parser::ast::{
use rustpython_parser::parser;
use crate::ast_ops::{extract_all_names, Binding, BindingKind, Scope, ScopeKind};
use crate::ast_visitor;
use crate::ast_visitor::{walk_excepthandler, ASTVisitor};
use crate::builtins::{BUILTINS, MAGIC_GLOBALS};
use crate::checks::{Check, CheckCode, CheckKind};
use crate::settings::Settings;
use crate::visitor;
use crate::visitor::{walk_excepthandler, Visitor};
struct Checker<'a> {
settings: &'a Settings,
@ -37,7 +37,7 @@ impl Checker<'_> {
}
}
impl Visitor for Checker<'_> {
impl ASTVisitor for Checker<'_> {
fn visit_stmt(&mut self, stmt: &Stmt) {
match &stmt.node {
StmtKind::Global { names } | StmtKind::Nonlocal { names } => {
@ -255,7 +255,7 @@ impl Visitor for Checker<'_> {
_ => {}
}
visitor::walk_stmt(self, stmt);
ast_visitor::walk_stmt(self, stmt);
match &stmt.node {
StmtKind::ClassDef { .. } => {
@ -359,7 +359,7 @@ impl Visitor for Checker<'_> {
_ => {}
};
visitor::walk_expr(self, expr);
ast_visitor::walk_expr(self, expr);
match &expr.node {
ExprKind::GeneratorExp { .. }
@ -466,7 +466,7 @@ impl Visitor for Checker<'_> {
}
}
visitor::walk_arguments(self, arguments);
ast_visitor::walk_arguments(self, arguments);
}
fn visit_arg(&mut self, arg: &Arg) {
@ -478,7 +478,7 @@ impl Visitor for Checker<'_> {
location: arg.location,
},
);
visitor::walk_arg(self, arg);
ast_visitor::walk_arg(self, arg);
}
}

View File

@ -1,14 +1,11 @@
use std::collections::{BTreeMap, BTreeSet};
use std::fmt::format;
use std::collections::BTreeMap;
use rustpython_parser::ast::{
Arg, Arguments, Expr, ExprContext, ExprKind, Location, Stmt, StmtKind, Suite,
};
use libcst_native::{Expression, If, Module};
use rustpython_parser::ast::Location;
use crate::ast_visitor;
use crate::ast_visitor::ASTVisitor;
use crate::check_cst::ScopeKind::{Class, Function, Generator, Module};
use crate::checks::{Check, CheckKind};
use crate::cst_visitor;
use crate::cst_visitor::CSTVisitor;
use crate::settings::Settings;
enum ScopeKind {
@ -44,9 +41,6 @@ struct Binding {
struct Checker<'a> {
settings: &'a Settings,
checks: Vec<Check>,
scopes: Vec<Scope>,
dead_scopes: Vec<Scope>,
in_f_string: bool,
}
impl Checker<'_> {
@ -54,384 +48,27 @@ impl Checker<'_> {
Checker {
settings,
checks: vec![],
scopes: vec![],
dead_scopes: vec![],
in_f_string: false,
}
}
}
impl ASTVisitor for Checker<'_> {
fn visit_stmt(&mut self, stmt: &Stmt) {
match &stmt.node {
StmtKind::FunctionDef { name, .. } => {
self.push_scope(Scope {
kind: Function,
values: BTreeMap::new(),
});
self.add_binding(Binding {
kind: BindingKind::ClassDefinition,
name: name.clone(),
used: false,
location: stmt.location,
})
}
StmtKind::AsyncFunctionDef { name, .. } => {
self.push_scope(Scope {
kind: Function,
values: BTreeMap::new(),
});
self.add_binding(Binding {
kind: BindingKind::ClassDefinition,
name: name.clone(),
used: false,
location: stmt.location,
})
}
StmtKind::Return { .. } => {
if self
.settings
.select
.contains(CheckKind::ReturnOutsideFunction.code())
{
if let Some(scope) = self.scopes.last() {
match scope.kind {
Class | Module => {
self.checks.push(Check {
kind: CheckKind::ReturnOutsideFunction,
location: stmt.location,
});
}
_ => {}
}
}
}
}
StmtKind::ClassDef { .. } => self.push_scope(Scope {
kind: Class,
values: BTreeMap::new(),
}),
StmtKind::Import { names } => {
for alias in names {
if alias.node.name.contains('.') && alias.node.asname.is_none() {
self.add_binding(Binding {
kind: BindingKind::SubmoduleImportation,
name: alias.node.name.clone(),
used: false,
location: stmt.location,
})
} else {
self.add_binding(Binding {
kind: BindingKind::Importation,
name: alias
.node
.asname
.clone()
.unwrap_or_else(|| alias.node.name.clone()),
used: false,
location: stmt.location,
})
}
}
}
StmtKind::ImportFrom { names, module, .. } => {
for alias in names {
let name = alias
.node
.asname
.clone()
.unwrap_or_else(|| alias.node.name.clone());
if module
.clone()
.map(|name| name == "future")
.unwrap_or_default()
{
self.add_binding(Binding {
kind: BindingKind::FutureImportation,
name,
used: true,
location: stmt.location,
});
} else if alias.node.name == "*" {
self.add_binding(Binding {
kind: BindingKind::StarImportation,
name,
used: false,
location: stmt.location,
});
if self
.settings
.select
.contains(CheckKind::ImportStarUsage.code())
{
self.checks.push(Check {
kind: CheckKind::ImportStarUsage,
location: stmt.location,
});
}
} else {
self.add_binding(Binding {
kind: BindingKind::Importation,
name: match module {
None => name,
Some(parent) => format!("{}.{}", parent, name),
},
used: false,
location: stmt.location,
})
}
}
}
StmtKind::If { test, .. } => {
if self.settings.select.contains(CheckKind::IfTuple.code()) {
if let ExprKind::Tuple { .. } = test.node {
self.checks.push(Check {
kind: CheckKind::IfTuple,
location: stmt.location,
});
}
}
}
StmtKind::Raise { exc, .. } => {
if self
.settings
.select
.contains(CheckKind::RaiseNotImplemented.code())
{
if let Some(expr) = exc {
match &expr.node {
ExprKind::Call { func, .. } => {
if let ExprKind::Name { id, .. } = &func.node {
if id == "NotImplemented" {
self.checks.push(Check {
kind: CheckKind::RaiseNotImplemented,
location: stmt.location,
});
}
}
}
ExprKind::Name { id, .. } => {
if id == "NotImplemented" {
self.checks.push(Check {
kind: CheckKind::RaiseNotImplemented,
location: stmt.location,
});
}
}
_ => {}
}
}
}
}
StmtKind::AugAssign { target, .. } => {
self.handle_node_load(target);
}
_ => {}
}
ast_visitor::walk_stmt(self, stmt);
match &stmt.node {
StmtKind::ClassDef { .. }
| StmtKind::FunctionDef { .. }
| StmtKind::AsyncFunctionDef { .. } => {
self.pop_scope();
}
_ => {}
};
if let StmtKind::ClassDef { name, .. } = &stmt.node {
self.add_binding(Binding {
kind: BindingKind::Definition,
name: name.clone(),
used: false,
location: stmt.location,
impl CSTVisitor for Checker<'_> {
fn visit_If<'a>(&'a mut self, node: &'a If) -> &'a If {
if let Expression::Tuple { .. } = node.test {
self.checks.push(Check {
kind: CheckKind::IfTuple,
location: Default::default(),
});
}
}
fn visit_expr(&mut self, expr: &Expr) {
let initial = self.in_f_string;
match &expr.node {
ExprKind::Name { ctx, .. } => match ctx {
ExprContext::Load => self.handle_node_load(expr),
ExprContext::Store => self.handle_node_store(expr),
ExprContext::Del => {}
},
ExprKind::GeneratorExp { .. } => self.push_scope(Scope {
kind: Generator,
values: BTreeMap::new(),
}),
ExprKind::Lambda { .. } => self.push_scope(Scope {
kind: Function,
values: BTreeMap::new(),
}),
ExprKind::JoinedStr { values } => {
if !self.in_f_string
&& self
.settings
.select
.contains(CheckKind::FStringMissingPlaceholders.code())
&& !values
.iter()
.any(|value| matches!(value.node, ExprKind::FormattedValue { .. }))
{
self.checks.push(Check {
kind: CheckKind::FStringMissingPlaceholders,
location: expr.location,
});
}
self.in_f_string = true;
}
_ => {}
};
ast_visitor::walk_expr(self, expr);
match &expr.node {
ExprKind::GeneratorExp { .. } | ExprKind::Lambda { .. } => {
if let Some(scope) = self.scopes.pop() {
self.dead_scopes.push(scope);
}
}
ExprKind::JoinedStr { .. } => {
self.in_f_string = initial;
}
_ => {}
};
}
fn visit_arguments(&mut self, arguments: &Arguments) {
if self
.settings
.select
.contains(CheckKind::DuplicateArgumentName.code())
{
// Collect all the arguments into a single vector.
let mut all_arguments: Vec<&Arg> = arguments
.args
.iter()
.chain(arguments.posonlyargs.iter())
.chain(arguments.kwonlyargs.iter())
.collect();
if let Some(arg) = &arguments.vararg {
all_arguments.push(arg);
}
if let Some(arg) = &arguments.kwarg {
all_arguments.push(arg);
}
// Search for duplicates.
let mut idents: BTreeSet<String> = BTreeSet::new();
for arg in all_arguments {
let ident = &arg.node.arg;
if idents.contains(ident) {
self.checks.push(Check {
kind: CheckKind::DuplicateArgumentName,
location: arg.location,
});
break;
}
idents.insert(ident.clone());
}
}
ast_visitor::walk_arguments(self, arguments);
}
fn visit_arg(&mut self, arg: &Arg) {
self.add_binding(Binding {
kind: BindingKind::Argument,
name: arg.node.arg.clone(),
used: false,
location: arg.location,
});
ast_visitor::walk_arg(self, arg);
cst_visitor::walk_If(self, node);
node
}
}
impl Checker<'_> {
fn push_scope(&mut self, scope: Scope) {
self.scopes.push(scope);
}
fn pop_scope(&mut self) {
self.dead_scopes
.push(self.scopes.pop().expect("Attempted to pop without scope."));
}
fn add_binding(&mut self, binding: Binding) {
// TODO(charlie): Don't treat annotations as assignments if there is an existing value.
let scope = self.scopes.last_mut().expect("No current scope found.");
scope.values.insert(
binding.name.clone(),
match scope.values.get(&binding.name) {
None => binding,
Some(existing) => Binding {
kind: binding.kind,
name: binding.name,
location: binding.location,
used: existing.used,
},
},
);
}
fn handle_node_load(&mut self, expr: &Expr) {
if let ExprKind::Name { id, .. } = &expr.node {
for scope in self.scopes.iter_mut().rev() {
if matches!(scope.kind, Class) {
if id == "__class__" {
return;
} else {
continue;
}
}
if let Some(binding) = scope.values.get_mut(id) {
binding.used = true;
}
}
}
}
fn handle_node_store(&mut self, expr: &Expr) {
if let ExprKind::Name { id, .. } = &expr.node {
// TODO(charlie): Handle alternate binding types (like `Annotation`).
self.add_binding(Binding {
kind: BindingKind::Assignment,
name: id.to_string(),
used: false,
location: expr.location,
});
}
}
fn check_dead_scopes(&mut self) {
// TODO(charlie): Handle `__all__`.
for scope in &self.dead_scopes {
for (name, binding) in scope.values.iter().rev() {
if !binding.used && matches!(binding.kind, BindingKind::Importation) {
self.checks.push(Check {
kind: CheckKind::UnusedImport(name.clone()),
location: binding.location,
});
}
}
}
}
}
pub fn check_ast(python_ast: &Suite, settings: &Settings) -> Vec<Check> {
pub fn check_cst(python_cst: &Module, settings: &Settings) -> Vec<Check> {
let mut checker = Checker::new(settings);
checker.push_scope(Scope {
kind: Module,
values: BTreeMap::new(),
});
for stmt in python_ast {
checker.visit_stmt(stmt);
for node in &python_cst.body {
checker.visit_Statement(node);
}
checker.pop_scope();
checker.check_dead_scopes();
checker.checks
}

File diff suppressed because it is too large Load Diff

View File

@ -1,6 +1,5 @@
mod ast_ops;
mod ast_visitor;
mod autofix;
mod builtins;
mod cache;
pub mod check_ast;

View File

@ -5,6 +5,7 @@ use log::debug;
use rustpython_parser::parser;
use crate::check_ast::check_ast;
use crate::check_cst::check_cst;
use crate::check_lines::check_lines;
use crate::checks::{Check, LintSource};
use crate::message::Message;
@ -24,19 +25,26 @@ pub fn check_path(path: &Path, settings: &Settings, mode: &cache::Mode) -> Resul
// Aggregate all checks.
let mut checks: Vec<Check> = vec![];
// Run the AST-based checks.
if settings
.select
.iter()
.any(|check_code| matches!(check_code.lint_source(), LintSource::AST))
{
let path = path.to_string_lossy();
let python_ast = parser::parse_program(&contents, &path)?;
checks.extend(check_ast(&python_ast, settings, &path));
}
// Run the CST-based checks.
let python_cst = match libcst_native::parse_module(&contents, None) {
Ok(m) => m,
Err(e) => panic!("Failed to parse CST."),
};
checks.extend(check_cst(&python_cst, settings));
// Run the lines-based checks.
check_lines(&mut checks, &contents, settings);
// // Run the AST-based checks.
// if settings
// .select
// .iter()
// .any(|check_code| matches!(check_code.lint_source(), LintSource::AST))
// {
// let path = path.to_string_lossy();
// let python_ast = parser::parse_program(&contents, &path)?;
// checks.extend(check_ast(&python_ast, settings, &path));
// }
//
// // Run the lines-based checks.
// check_lines(&mut checks, &contents, settings);
// Convert to messages.
let messages: Vec<Message> = checks
@ -47,6 +55,7 @@ pub fn check_path(path: &Path, settings: &Settings, mode: &cache::Mode) -> Resul
filename: path.to_string_lossy().to_string(),
})
.collect();
cache::set(path, settings, &messages, mode);
Ok(messages)