feat!: detect duplicate test cases #17955
Conversation
eslint-github-bot
bot
added
breaking
✅ Deploy Preview for docs-eslint canceled.
|
|
We are waiting on #17654 for this, correct? |
|
No I think we can treat them as independent. |
* main: chore: Split Docs CI from core CI (eslint#17897) fix: Ensure config keys are printed for config errors (eslint#17980) chore: delete relative-module-resolver.js (eslint#17981) docs: migration guide entry for `no-inner-declarations` (eslint#17977) docs: Update README feat: maintain latest ecma version in ESLint (eslint#17958) feat!: no-inner-declaration new default behaviour and option (eslint#17885) feat: add `no-useless-assignment` rule (eslint#17625) fix: `no-misleading-character-class` edge cases with granular errors (eslint#17970) feat: `no-misleading-character-class` granular errors (eslint#17515) docs: fix number of code-path events on custom rules page (eslint#17969) docs: reorder entries in v9 migration guide (eslint#17967) fix!: handle `--output-file` for empty output when saving to disk (eslint#17957)
* main: docs: Update note about ECMAScript support (eslint#17991) chore: update `markdownlint` to `v0.33.0` (eslint#17995) docs: Update release blog post template (eslint#17994) docs: Add sections on non-npm plugin configuration (eslint#17984) 9.0.0-alpha.1 Build: changelog update for 9.0.0-alpha.1 chore: package.json update for @eslint/js release
|
This is ready for review. |
lib/rule-tester/rule-tester.js
Outdated
| * @private | ||
| */ | ||
| function checkDuplicateTestCase(item, seenTestCases) { | ||
| if ((Array.isArray(item.options) && item.options.some(i => typeof i === "object")) || item.settings || item.languageOptions?.parser || item.languageOptions?.parserOptions || item.plugins) { |
There was a problem hiding this comment.
there is a room to improve it, but I think it's fine to just check these properties 👍 - as it's described in the rfc.
There was a problem hiding this comment.
It's a start.
However, I would love it if I could replace this bail-out logic with a simple function call to see if a given object is serializable. Surprisingly, I haven't found a third-party library to do this. Another option is to try to piece together such a function based on StackOverflow snippets (https://stackoverflow.com/questions/30579940/reliable-way-to-check-if-objects-is-serializable-in-) etc and add unit tests for it. Thoughts?
It might be best to solve this the right way now (in a way that avoids a lot of false negatives) since changing this later could be a breaking change.
There was a problem hiding this comment.
IMHO, it's not worth adding a new dependency: rule tester is shipped with eslint, but most eslint users don't really use it. A simple implementation to cover 90%+ cases is good enough for me. 😄
There was a problem hiding this comment.
I went ahead and added a isSerializable() function which uses only the tiny lodash.isplainobject as a dependency. While the previous approach could have been sufficient as you mentioned, this new approach is a lot more robust (no false negatives) and the code is cleaner.
|
@christian-bromann looks like our browser tests are failing again, can you take a look? |
|
@nzakas on it |
@nzakas @bmish I've discovered a performance regression with the recent release which forced me to find the culprit which I did successfully. The browser tests should run much faster now and should be more reliable too. Please re-trigger the build and the fix should be picked up automatically. |
mdjermanovic
added
the
accepted
It works now, thanks! |
lib/shared/serialization.js
Outdated
| /** | ||
| * Check if an object is serializable. | ||
| * Functions or objects like RegExp cannot be serialized by JSON.stringify(). | ||
| * Inspired by: https://stackoverflow.com/questions/30579940/reliable-way-to-check-if-objects-is-serializable-in-javascript | ||
| * @param {any} obj the object | ||
| * @returns {boolean} true if the object is serializable | ||
| */ | ||
| function isSerializable(obj) { | ||
| if (!isSerializablePrimitiveOrPlainObject(obj)) { | ||
| return false; | ||
| } | ||
| for (const property in obj) { |
There was a problem hiding this comment.
Does "object" mean that this function expects only objects? If that's not the case, then it might be better to rename obj to val to clarify that it accepts any value, and also do for-in only on objects so that it doesn't, for example, iterate over string characters (currently, for string test cases, it calls isSerializablePrimitiveOrPlainObject for each character of the string).
There was a problem hiding this comment.
Good point, changed to val, and only looping for objects.
Prerequisites checklist
What is the purpose of this pull request? (put an "X" next to an item)
[ ] Documentation update
[ ] Bug fix (template)
[ ] New rule (template)
[ ] Changes an existing rule (template)
[ ] Add autofix to a rule
[ ] Add a CLI option
[x] Add something to the core
[ ] Other, please explain:
What changes did you make? (Give an overview)
Fixes #15104, alongside:
Link to this feature in the RFC:
Inspired by:
Is there anything you'd like reviewers to focus on?
Are we okay with adding this
isSerializable()function? Surprisingly, I couldn't find any third-party library that offers it. I had an alternative approach visible in b400d1e that simply checked for the presence of any properties that might not be serializable, but this could have false negatives. See discussion in #17955 (comment).TODO: