Rhino 1.9 and `core-js` compat tests

[zloirock]

I found that Rhino 1.9 has been released and tried to update core-js compat data.

Since binaries are no longer attached to Rhino releases, this doc should be updated,

npm run compat-rhino YOUR_PATH_TO_RHINO

So, I tried to run core-js on Rhino 1.9 and found a strange error:

The compat data tests runner gets an object with tests, enumerates properties of this object, and runs them.

On the third property of this object, Rhino throws an error, no matter what it contains.

Rhino 1.8 (and all the rest JS engines) works fine here.

[aardvark179]

Okay. I think we need to update the docs, maybe they should point to instructions in the core-js repo since it can change between releases.

Running the tests with npm run compat-rhino /Users/duncan.macgregor/git/mozilla/rhino/rhino-all/build/libs/rhino-all-1.9.0.jar is producing the error you see. I can fix that but then we immediately fail with a different error.

@gbrail do we have a procedure to follow before doing the release, these feel like things we should have caught.

[zloirock]

maybe they should point to instructions in the core-js repo

https://github.com/zloirock/core-js/blob/master/CONTRIBUTING.md#how-to-update-core-js-compat-data

[aardvark179]

Okay, the other problem is that at some point global has stopped being set in whatever context this is run in. I'll bisect that later to work out when it broke. Fixing the script for that does allow me to run the script on 2.0 snapshot and shows the following changes

es.error.is-error                            | not required
es.array.concat                              | not required
es.array.filter                              | not required
es.array.map                                 | not required
es.array.slice                               | not required
es.array-buffer.is-view                      | not required
es.array-buffer.detached                     | not required
es.array-buffer.transfer                     | not required
es.array-buffer.transfer-to-fixed-length     | not required
es.function.has-instance                     | not required
es.math.f16round                             | not required
es.object.assign                             | not required
es.object.proto                              | not required
es.promise.try                               | not required
es.promise.with-resolvers                    | not required
es.regexp.exec                               | not required
es.string.match                              | not required
es.string.replace                            | not required
es.string.search                             | not required
es.string.split                              | not required
es.typed-array.copy-within                   | not required
es.typed-array.every                         | not required
es.typed-array.fill                          | not required
es.typed-array.filter                        | not required
es.typed-array.find                          | not required
es.typed-array.find-index                    | not required
es.typed-array.for-each                      | not required
es.typed-array.includes                      | not required
es.typed-array.index-of                      | not required
es.typed-array.join                          | not required
es.typed-array.last-index-of                 | not required
es.typed-array.map                           | not required
es.typed-array.reduce                        | not required
es.typed-array.reduce-right                  | not required
es.typed-array.reverse                       | not required
es.typed-array.slice                         | not required
es.typed-array.some                          | not required
es.typed-array.sort                          | not required
es.typed-array.subarray                      | not required
es.typed-array.to-locale-string              | not required

[zloirock]

@aardvark179 thanks -)

[aardvark179]

Okay. global not being defined is all my fault. It broke in c1ee781, I'll fix that tomorrow.

[gbrail]

Thanks for checking out core-js and trying to update it! Yes, there are some steps on the release checklist that aren't done yet and TBH some of them need to be updated. I'll take a look at that stuff, but at this point it will be probably some time next week when I really spend time on it. We have similar issues with the "compat table" that I DID update for this release -- I'd love to automate running that and checking for regressions, since sometimes stuff gets broken there between releases. Some sort of test that checks what core-js is relying on would be great as well, so that we don't regress and rely on manual checks at the last minute to catch these kinds of things.

…

On Tue, Dec 23, 2025 at 12:45 PM Duncan MacGregor ***@***.***> wrote: *aardvark179* left a comment (mozilla/rhino#2239) <#2239 (comment)> Okay. global not being defined is all my fault. It broke in c1ee781 <c1ee781>, I'll fix that tomorrow. — Reply to this email directly, view it on GitHub <#2239 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAD7I23S2JTXNONLNT53ASD4DGSU7AVCNFSM6AAAAACP2WVSQSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMOBXHEZDANBWGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[aardvark179]

Okay. I've now narrowed down the issue with the method name. We started to infer correct method names in more places, meaning method names end up reaching the class compiler which had previously been erased. I'll make a PR for that, and a separate one for global.

[aardvark179]

PRs created.