Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
feat(eslint-plugin): add no-unnecessary-condition rule (#699)
- Loading branch information
1 parent
b49dbfd
commit 5715482
Showing
7 changed files
with
457 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
64 changes: 64 additions & 0 deletions
64
packages/eslint-plugin/docs/rules/no-unnecessary-condition.md
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,64 @@ | ||
# Condition expressions must be necessary | ||
|
||
Any expression being used as a condition must be able to evaluate as truthy or falsy in order to be considered "necessary". Conversely, any expression that always evaluates to truthy or always evaluates to falsy, as determined by the type of the expression, is considered unnecessary and will be flagged by this rule. | ||
|
||
The following expressions are checked: | ||
|
||
- Arguments to the `&&`, `||` and `?:` (ternary) operators | ||
- Conditions for `if`, `for`, `while`, and `do-while` statements. | ||
|
||
Examples of **incorrect** code for this rule: | ||
|
||
```ts | ||
function head<T>(items: T[]) { | ||
// items can never be nullable, so this is unnecessary | ||
if (items) { | ||
return items[0].toUpperCase(); | ||
} | ||
} | ||
|
||
function foo(arg: 'bar' | 'baz') { | ||
// arg is never nullable or empty string, so this is unnecessary | ||
if (arg) { | ||
} | ||
} | ||
``` | ||
|
||
Examples of **correct** code for this rule: | ||
|
||
```ts | ||
function head<T>(items: T[]) { | ||
// Necessary, since items.length might be 0 | ||
if (items.length) { | ||
return items[0].toUpperCase(); | ||
} | ||
} | ||
|
||
function foo(arg: string) { | ||
// Necessary, since foo might be ''. | ||
if (arg) { | ||
} | ||
} | ||
``` | ||
|
||
## Options | ||
|
||
Accepts an object with the following options: | ||
|
||
- `ignoreRhs` (default `false`) - doesn't check if the right-hand side of `&&` and `||` is a necessary condition. For example, the following code is valid with this option on: | ||
|
||
```ts | ||
function head<T>(items: T[]) { | ||
return items.length && items[0].toUpperCase(); | ||
} | ||
``` | ||
|
||
## When Not To Use It | ||
|
||
The main downside to using this rule is the need for type information. | ||
|
||
## Related To | ||
|
||
- ESLint: [no-constant-condition](https://eslint.org/docs/rules/no-constant-condition) - this rule is essentially a stronger versison | ||
|
||
- [strict-boolean-expression](./strict-boolean-expressions.md) - a stricter alternative to this rule. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
192 changes: 192 additions & 0 deletions
192
packages/eslint-plugin/src/rules/no-unnecessary-condition.ts
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,192 @@ | ||
import { | ||
TSESTree, | ||
AST_NODE_TYPES, | ||
} from '@typescript-eslint/experimental-utils'; | ||
import ts, { TypeFlags } from 'typescript'; | ||
import { | ||
isTypeFlagSet, | ||
unionTypeParts, | ||
isFalsyType, | ||
isBooleanLiteralType, | ||
isLiteralType, | ||
} from 'tsutils'; | ||
import { | ||
createRule, | ||
getParserServices, | ||
getConstrainedTypeAtLocation, | ||
} from '../util'; | ||
|
||
// Truthiness utilities | ||
// #region | ||
const isTruthyLiteral = (type: ts.Type): boolean => | ||
isBooleanLiteralType(type, true) || (isLiteralType(type) && !!type.value); | ||
|
||
const isPossiblyFalsy = (type: ts.Type): boolean => | ||
unionTypeParts(type) | ||
// PossiblyFalsy flag includes literal values, so exclude ones that | ||
// are definitely truthy | ||
.filter(t => !isTruthyLiteral(t)) | ||
.some(type => isTypeFlagSet(type, ts.TypeFlags.PossiblyFalsy)); | ||
|
||
const isPossiblyTruthy = (type: ts.Type): boolean => | ||
unionTypeParts(type).some(type => !isFalsyType(type)); | ||
|
||
// isLiteralType only covers numbers and strings, this is a more exhaustive check. | ||
const isLiteral = (type: ts.Type): boolean => | ||
isBooleanLiteralType(type, true) || | ||
isBooleanLiteralType(type, false) || | ||
type.flags === ts.TypeFlags.Undefined || | ||
type.flags === ts.TypeFlags.Null || | ||
type.flags === ts.TypeFlags.Void || | ||
isLiteralType(type); | ||
// #endregion | ||
|
||
type ExpressionWithTest = | ||
| TSESTree.ConditionalExpression | ||
| TSESTree.DoWhileStatement | ||
| TSESTree.ForStatement | ||
| TSESTree.IfStatement | ||
| TSESTree.WhileStatement; | ||
|
||
export type Options = [ | ||
{ | ||
ignoreRhs?: boolean; | ||
}, | ||
]; | ||
|
||
export type MessageId = | ||
| 'alwaysTruthy' | ||
| 'alwaysFalsy' | ||
| 'literalBooleanExpression' | ||
| 'never'; | ||
export default createRule<Options, MessageId>({ | ||
name: 'no-unnecessary-conditionals', | ||
meta: { | ||
type: 'suggestion', | ||
docs: { | ||
description: | ||
'Prevents conditionals where the type is always truthy or always falsy', | ||
category: 'Best Practices', | ||
recommended: false, | ||
requiresTypeChecking: true, | ||
}, | ||
schema: [ | ||
{ | ||
type: 'object', | ||
properties: { | ||
ignoreRhs: { | ||
type: 'boolean', | ||
}, | ||
}, | ||
additionalProperties: false, | ||
}, | ||
], | ||
messages: { | ||
alwaysTruthy: 'Unnecessary conditional, value is always truthy.', | ||
alwaysFalsy: 'Unnecessary conditional, value is always falsy.', | ||
literalBooleanExpression: | ||
'Unnecessary conditional, both sides of the expression are literal values', | ||
never: 'Unnecessary conditional, value is `never`', | ||
}, | ||
}, | ||
defaultOptions: [ | ||
{ | ||
ignoreRhs: false, | ||
}, | ||
], | ||
create(context, [{ ignoreRhs }]) { | ||
const service = getParserServices(context); | ||
const checker = service.program.getTypeChecker(); | ||
|
||
function getNodeType(node: TSESTree.Node): ts.Type { | ||
const tsNode = service.esTreeNodeToTSNodeMap.get(node); | ||
return getConstrainedTypeAtLocation(checker, tsNode); | ||
} | ||
|
||
/** | ||
* Checks if a conditional node is necessary: | ||
* if the type of the node is always true or always false, it's not necessary. | ||
*/ | ||
function checkNode(node: TSESTree.Node): void { | ||
const type = getNodeType(node); | ||
|
||
// Conditional is always necessary if it involves `any` or `unknown` | ||
if (isTypeFlagSet(type, TypeFlags.Any | TypeFlags.Unknown)) { | ||
return; | ||
} | ||
if (isTypeFlagSet(type, TypeFlags.Never)) { | ||
context.report({ node, messageId: 'never' }); | ||
} else if (!isPossiblyTruthy(type)) { | ||
context.report({ node, messageId: 'alwaysFalsy' }); | ||
} else if (!isPossiblyFalsy(type)) { | ||
context.report({ node, messageId: 'alwaysTruthy' }); | ||
} | ||
} | ||
|
||
/** | ||
* Checks that a binary expression is necessarily conditional, reports otherwise. | ||
* If both sides of the binary expression are literal values, it's not a necessary condition. | ||
* | ||
* NOTE: It's also unnecessary if the types that don't overlap at all | ||
* but that case is handled by the Typescript compiler itself. | ||
*/ | ||
const BOOL_OPERATORS = new Set([ | ||
'<', | ||
'>', | ||
'<=', | ||
'>=', | ||
'==', | ||
'===', | ||
'!=', | ||
'!==', | ||
]); | ||
function checkIfBinaryExpressionIsNecessaryConditional( | ||
node: TSESTree.BinaryExpression, | ||
): void { | ||
if ( | ||
BOOL_OPERATORS.has(node.operator) && | ||
isLiteral(getNodeType(node.left)) && | ||
isLiteral(getNodeType(node.right)) | ||
) { | ||
context.report({ node, messageId: 'literalBooleanExpression' }); | ||
} | ||
} | ||
|
||
/** | ||
* Checks that a testable expression is necessarily conditional, reports otherwise. | ||
* Filters all LogicalExpressions to prevent some duplicate reports. | ||
*/ | ||
function checkIfTestExpressionIsNecessaryConditional( | ||
node: ExpressionWithTest, | ||
): void { | ||
if ( | ||
node.test !== null && | ||
node.test.type !== AST_NODE_TYPES.LogicalExpression | ||
) { | ||
checkNode(node.test); | ||
} | ||
} | ||
|
||
/** | ||
* Checks that a logical expression contains a boolean, reports otherwise. | ||
*/ | ||
function checkLogicalExpressionForUnnecessaryConditionals( | ||
node: TSESTree.LogicalExpression, | ||
): void { | ||
checkNode(node.left); | ||
if (!ignoreRhs) { | ||
checkNode(node.right); | ||
} | ||
} | ||
|
||
return { | ||
BinaryExpression: checkIfBinaryExpressionIsNecessaryConditional, | ||
ConditionalExpression: checkIfTestExpressionIsNecessaryConditional, | ||
DoWhileStatement: checkIfTestExpressionIsNecessaryConditional, | ||
ForStatement: checkIfTestExpressionIsNecessaryConditional, | ||
IfStatement: checkIfTestExpressionIsNecessaryConditional, | ||
WhileStatement: checkIfTestExpressionIsNecessaryConditional, | ||
LogicalExpression: checkLogicalExpressionForUnnecessaryConditionals, | ||
}; | ||
}, | ||
}); |
Oops, something went wrong.