-
Notifications
You must be signed in to change notification settings - Fork 29.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
assert,util: always verify both sides #30764
Conversation
This veryfies that both input arguments are always of the identical type. It was possible to miss a few cases before. This change applies to all deep equal assert functions (e.g., `assert.deepStrictEqual()`) and to `util.isDeepStrictEqual()`.
This makes sure the assertion does not depend on the argument order. It also removes comments that do not apply anymore and verifies the behavior for the loose and strict implementation.
This comment has been minimized.
This comment has been minimized.
if (isDate(val1)) { | ||
if (DatePrototypeGetTime(val1) !== DatePrototypeGetTime(val2)) { | ||
} else if (isDate(val1)) { | ||
if (!isDate(val2) || |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
there's no failing tests that necessitate any change to the Date or RegExp or Error or boxed primitive logic - can you add those tests?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Boxed primitives already have such test case.
I could add more tests for the other types but I actually like things like these to keep them around for coverage tasks at Code and Learn events. I am also fine to add the tests though, if you think that should be done in this PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If there's that much of a shortage of tasks for those events, then so be it, but it seems unwise to make untested changes.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I just added multiple tests and I indeed forgot to add one check 😆
There are a couple more cases that are not yet completely covered: Set
, Map
, ArrayBuffer
and non-native errors.
@@ -1123,3 +1116,33 @@ assert.throws( | |||
// The descriptor is not compared. | |||
assertDeepAndStrictEqual(a, { a: 5 }); | |||
} | |||
|
|||
// Verify object types being identical on both sides. | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
these are the only tests that fail without your change
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, that is correct. The second commit is just there to prevent any further issues like these. It's somewhat independent but it fit together. Would you rather separate the two commits into two PRs?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nah this is fine to me, was highlighting here in reference to #30764 (comment)
if (!isEqualBoxedPrimitive(val1, val2)) { | ||
return false; | ||
} | ||
} else if (ArrayIsArray(val2) || |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why check isArray of val2 way down here, instead of only up by the place where val1 is checked (as in #30743)?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's less CPU work.
In #30743 all entries have to check both sides of both types. But if the arguments are e.g., sets, we'd only have to know that one side is not a set and then continue until we match sets. There we verify the other side.
If it's neither of the explicitly listed entries, we'll verify that the right side is none of the already tested types.
This LGTM as an alternative to #30743; once this lands, I'll either update my PR (if there's more tests worth adding) or close it (if there's no longer anything to add). |
This comment has been minimized.
This comment has been minimized.
CITGM https://ci.nodejs.org/view/Node.js-citgm/job/citgm-smoker/2110/ (everything seems normal). |
CI: https://ci.nodejs.org/job/node-test-pull-request/27406/ ✅ (yellow build with a single Windows flake. It passed in another CI run before) |
This veryfies that both input arguments are always of the identical type. It was possible to miss a few cases before. This change applies to all deep equal assert functions (e.g., `assert.deepStrictEqual()`) and to `util.isDeepStrictEqual()`. PR-URL: #30764 Refs: #30743 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: David Carlier <devnexen@gmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
This makes sure the assertion does not depend on the argument order. It also removes comments that do not apply anymore and verifies the behavior for the loose and strict implementation. PR-URL: #30764 Refs: #30743 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: David Carlier <devnexen@gmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
This veryfies that both input arguments are always of the identical type. It was possible to miss a few cases before. This change applies to all deep equal assert functions (e.g., `assert.deepStrictEqual()`) and to `util.isDeepStrictEqual()`. PR-URL: #30764 Refs: #30743 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: David Carlier <devnexen@gmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
Please have a look at the commit messages for details.
Refs: #30743
Checklist
make -j4 test
(UNIX), orvcbuild test
(Windows) passes