Linting Wish List

[bdemann]

Good referenc for all of the rules

• https://eslint.org/docs/latest/rules/ for eslint rules
• https://typescript-eslint.io/rules/ for typescript-eslint rules

Linting philosophy

With https://github.com/demergent-labs/azle/pull/1733 we were able to add the recomended linting rules and some of the below rules. From here we are going to add new linting rules in this manner. As we are going though PRs if there is something that keeps standing out to us and something we keep needing to call out, then we will set apart some time to enable the rule and fix all of the problems for that rule.

Rules we want but will take a long time to implement

• [ ] https://github.com/demergent-labs/azle/issues/1737 Explicit Any
• [ ] https://github.com/demergent-labs/azle/issues/1738 Explicit Return types

Consider the following rules

• [x] Can the linter force every function (stand alone, or in a map or filter or reduce or callback) have a return type?
  • [ ] @typescript-eslint/explicit-function-return-type should do the trick
  • [ ] But there are 1413 times where functions don't have return types that we would have to clean up.
  • [x] make issue for this one

Things that might be nice but require parsing https://github.com/demergent-labs/azle/issues/1739

• [ ] @typescript-eslint/consistent-return
• [ ] @typescript-eslint/consistent-type-definitions
• [ ] @typescript-eslint/strict-boolean-expressions (see bellow)
• [ ] @typescript-eslint/prefer-nullish-coalescing
• [ ] @typescript-eslint/prefer-optional-chain

Things Jordan wants (from https://github.com/demergent-labs/azle/issues/1535)

• [ ] No falsy checks (value !== null && value !== undefined instead of value)
  • [ ] @typescript-eslint/strict-boolean-expressions I think is the one we want
• [ ] No mutations of const variables
  • [ ] prefer-const?
  • [ ] We may need a plugin like eslint-plugin-immutable
  • [ ] We may need a more powerful plugin to let us mutate lets and catch the use of lets where it's not mutated

I'm guessing Jordan would like

• [ ] no-plusplus (15 problems)
• [ ] no-restricted-syntax
  • [ ] restrict switches, for loops, etc
• [ ] @typescript-eslint/prefer-readonly
• [ ] @typescript-eslint/prefer-readonly-parameter-types
• [ ] @typescript-eslint/prefer-reduce-type-parameter

Things to look into

• [ ] Can we have the camelcase rule if we have so many ic things that come in as snake case?
  • No not really
• [ ] consistent-return
• [ ] curly
• [ ] no-shadow
• [ ] no-throw-literal (11 problems)
• [ ] @typescript-eslint/only-throw-error
• [ ] no-undefined
• [ ] @typescript-eslint/consistent-indexed-object-style
  • [ ] ask Jordan
• [ ] @typescript-eslint/consistent-type-assertions
  • [ ] ask Jordan
• [ ] @typescript-eslint/consistent-type-exports
  • [ ] ask Jordan
• [ ] @typescript-eslint/consistent-type-imports
  • [ ] pairs with type exports
• [ ] @typescript-eslint/explicit-member-accessibility
• [ ] @typescript-eslint/naming-convention
• [ ] @typescript-eslint/sort-type-constituents
• [ ] @typescript-eslint/typedef
  • [ ] A combination of --noImplicitAny and --strictPropertyInitialization compiler options are actually what we want to get type defs where they are important

Help with declarative code

• [ ] max-depth
• [ ] max-lines
• [ ] max-lines-per-function
• [ ] max-nested callbacks
• [ ] no-nested-ternary

Probably Unrealistic

• [ ] no-magic-numbers (Almost 3000 problems, and sometimes a magic number is good enough)
• [ ] @typescript-eslint/no-magic-numbers

I'm not sure we need this but maybe we might as well

• [ ] no-unexpected-multiline
• [ ] no-constant-binary-expression
• [ ] no-array-constructor
• [ ] @typescript-eslint/array-type
• [ ] @typescript-eslint/explicit-module-boundary-types
• [ ] @typescript-eslint/no-non-null-assertion
• [ ] @typescript-eslint/no-confusing-non-null-assertion
• [ ] @typescript-eslint/no-confusing-void-expression
• [ ] @typescript-eslint/no-unnecessary-qualifier
• [ ] @typescript-eslint/no-unnecessary-type-arguments
• [ ] @typescript-eslint/no-unnecessary-condition
• [ ] prefer-object-spread?
  • [ ] Is only related to Object.assign, I am not sure if this is actually helpful
• [ ] prefer-spread
  • [ ] only applies in contrast to .apply(), not sure that this is helpful

Formatting

• [ ] Single line comments can get very long and it would be nice if prettier or eslint formatted those for us

Resolved with https://github.com/demergent-labs/azle/pull/1733

• [x] Add /*global functionName, variableName*/ to any file that has globals
  • We found that the tsc actually does this check for us so we can remove it
• [x] no-var
• [x] @typescript-eslint/prefer-for-of
• [x] prefer-arrow-callback
• [x] no-param-reassign
• [x] array-callback-return (related to the one above but not quite)
• [x] no-template-curly-in-string
• [x] eqeqeq
• [x] prefer-template

[bdemann]

Here are some rules that I tried out that didn't go so well

const proposedRules = {
    '@typescript-eslint/explicit-function-return-type': 'error', // 1413 problems // TODO https://github.com/demergent-labs/azle/issues/1576
    'consistent-return': 'off',
    '@typescript-eslint/consistent-return': 'error', // TODO this one isn't found?
    '@typescript-eslint/consistent-type-definitions': ['error', 'type'], // TODO seems nice but there are 17 places where I think we need interfaces over types
    '@typescript-eslint/strict-boolean-expressions': 'error', // TODO requires special parsing set up and more javascript heap
    '@typescript-eslint/prefer-optional-chain': 'error', // TODO same
    '@typescript-eslint/prefer-nullish-coalescing': 'error' // TODO same
};

[bdemann]

To do more complicated linting with different setting at different folders (which might enable being able to lint everything with the typescript compiler plugin) we may need to play around with the config file. It looks like they have recently updated the config file, and we are still on the new one. We may need to upgrade.

https://eslint.org/docs/latest/use/getting-started#configuration https://eslint.org/docs/latest/use/configure/migration-guide https://eslint.org/blog/2022/08/new-config-system-part-2/