[META] TSLint Compatibility

[dannyfritz]

Please view the ROADMAP instead.

---

Outdated List:

• array-type (Add rule array-type #211 )
• adjacent-overload-signatures (New: Add adjacent-overload-signatures #37 )
• ban-types (Add ban-types rule #181 )
• callable-types
• class-name (New: class-name-casing rule (class-name from TSLint) #44 class-name-casing)
• interface-name (New: Add rule interface-name-prefix #10 interface-name-prefix)
• interface-over-type-literal (New: Add rule no-type-literal #15 no-type-literal)
• member-access (New: Add rule explicit-member-accessibility #4 explicit-member-accessibility)
• member-ordering (New: Add rule member-ordering #23 )
• no-angle-bracket-type-assertion (New: Add rule no-angle-bracket-type-assertion #13 )
• no-any (New: Add rule no-any #8 no-explicit-any)
• no-empty-interface (New: Add no-empty-interface #39 )
• no-inferred-empty-object-type
• no-inferrable-types (add no-inferrable-types rule #124 )
• no-internal-module (New: Add rule prefer-namespace-keyword #22 prefer-namespace-keyword)
• no-mergeable-namespace
• no-misused-new
• no-namespace (New: Add rule no-namespace #21 )
• no-non-null-assertion (New: Add no-non-null-assertion #108 )
• no-object-literal-type-assertion
• no-parameter-properties (New: Add no-parameter-properties #53 )
• no-reference (New: Add rule no-triple-slash-reference #6 no-triple-slash-reference)
• no-reference-input
• no-redundant-jsdoc
• no-unsafe-any
• no-unnecessary-type-assertion
• no-var-requires (New: Add no-var-requires #114 )
• type-literal-delimiter (New: Add member-delimiter-style (fixes #17) #48 member-delimiter-style)
• typedef
• typedef-whitespace (type-annotation-spacing)
• unifed-signatures

[dannyfritz]

@weirdpattern Any of the unchecked items are up for grabs 😄 . My work got unexpectedly busy and I won't be active on this for a little while.

[weirdpattern]

@dannyfritz alright, I will start working on some of these... still deciding which ones though...
I will let you know, thanks!

[weirdpattern]

I'm working on no-parameter-properties and no-internal-modules...

[dannyfritz]

Huge shoutout to @weirdpattern for taking us over the 50% line! 🎉 🎉 🎉 🎉 🎉 🎉

[weirdpattern]

Update

no-parameter-properties is currently blocked due to TypeScript Parser bug 143. I have a solution, but I'm waiting on @nzakas to confirm how he wants to proceed.

no-inferrable-types is currently blocked due to TypeScript Parser bug 146. PR is being reviewed.

So, working on member-ordering for now =]

[dannyfritz]

@weirdpattern can you add the compatibility section into #23?

[weirdpattern]

of course, not in front of my computer at the moment, but will do it tonight...

[weirdpattern]

@dannyfritz done!

[weirdpattern]

Hi @dannyfritz can you mark member-ordering as completed?
It was merged this afternoon... Thanks!

[dannyfritz]

Thanks for your work!

[j-f1]

How about class-name (TSLint rule source)?

[weirdpattern]

All, I'm going to start working on no-parameter-properties.

[weirdpattern]

Now working on no-mergeable-namespace and no-non-null-assertion...

[corbinu]

@weirdpattern sweet thanks so much for all the work your doing!

[weirdpattern]

I was reviewing the remaining rules, I believe #48 is a superset of type-literal-delimiter... I mean it does something similar, and includes interfaces too... thoughts?

[dannyfritz]

Yes, I do agree the functionality is met with member-delimiter-style. Thanks!

[dannyfritz]

It is worth someone else going through the TSLint rules and seeing if you agree I grabbed all the relevant TS specific ones. https://palantir.github.io/tslint/rules/

They change pretty frequently themselves. So it is a moving target. 🎯

[weirdpattern]

Hey @j-f1, this is great! One question though, some of these are common to both JavaScript (es5, es6, etc) and TypeScript, so not sure if they should be defined here or at the eslint level... I know eslint already offers some rules to validate jsdoc and to restrict syntax like for-in...

By the way, I haven't really looked at all the rules, this was something that just popped into my head... still needs some analysis before we can decide anything ☺️

[j-f1]

The difference between ESLint’s valid-jsdoc and TSLint’s jsdoc-format is that TSLint just validates the syntax; whereas ESLint requires defining parameter and return type info.

[dannyfritz]

It might be worth adding those to this list and opening a pr in eslint to close it?

[dannyfritz]

I think the line that should be drawn for including in this repo is: can this rule be used without typescript?

[macklinu]

no-non-null-assertion can be checked in the original list now that #108 has been merged.

[nevir]

no-floating-promises would be a huge win, and seems like it would require TypeScript's knowledge of an expression's return type

[dannyfritz]

TSLint has changed since I originally made this issue a few times. If someone wants to do the research to corroborate where we are with the moving target I'd gladly update the original list.

[weirdpattern]

Just a thought...

Would it be better to move all these pending rules to the README.md or maybe ROADMAP.md file?

I think it will give users a clear idea of where we are heading without having to search this info in a bug?

[armano2]

i'm going to work now on no-misused-new

[armano2]

Updated list

• type-literal-delimiter
• no-reference-import
• no-redundant-jsdoc
• no-parameter-properties
• no-angle-bracket-type-assertion
• interface-over-type-literal
• interface-name
• callable-types
• array-type
• no-mergeable-namespace
• no-object-literal-type-assertion
• no-misused-new
• unified-signatures
• typedef-whitespace
• typedef
• no-var-requires
• no-non-null-assertion
• no-namespace
• no-internal-module
• no-inferrable-types
• no-empty-interface
• no-any
• member-access
• ban-types
• adjacent-overload-signatures

With Type Info

• no-unnecessary-type-assertion
• await-promise
• no-floating-promises
• no-inferred-empty-object-type
• no-unbound-method
• no-unsafe-any
• no-unused-variable
• strict-boolean-expressions
• strict-type-predicates
• use-default-type-parameter
• prefer-readonly
• match-default-export-name
• no-boolean-literal-compare
• no-unnecessary-qualifier

generator

// go to https://palantir.github.io/tslint/rules/ and tun this script

function generate () {
var items = Array.from(document.querySelectorAll('.rules-list li'))
    .map((item) => {
        return {
            tsOnly: Boolean(item.querySelector('.rule-features .feature-ts-only')),
            fixer: Boolean(item.querySelector('.rule-features .feature-fixer')),
            requireType: Boolean(item.querySelector('.rule-features .feature-requires-type-info')),
            name: item.querySelector('strong')
        }
    })
    .filter(item => item.tsOnly && item.name)
	.sort(item => item.requireType ? 1 : -1)
	.map((item) => { item.name = item.name.innerHTML; return item })

return `
${items.filter(item => !item.requireType).map((item) => { return `- [ ] [${item.name}]` }).join('\n')}

### With Type Info
${items.filter(item => item.requireType).map((item) => { return `- [ ] [${item.name}]` }).join('\n')}

${items.map((item) => { return `[${item.name}]: https://palantir.github.io/tslint/rules/${item.name}/` }).join('\n')}
`
}
generate()

[j-f1]

See #228

[bradzacher]

Thanks to @j-f1, we can now track this via https://github.com/bradzacher/eslint-plugin-typescript/blob/master/ROADMAP.md

Unless anyone has any objections, I'll close this issue.

[dannyfritz]

No objection from me