- Issue created by @mstrelan
- ๐ฎ๐นItaly mondrake ๐ฎ๐น
Iโve closed ๐ Assert::isType() is deprecated Active that is now a duplicate.
The only question I have here is whether we should wait for PHPUnit 10 support to be removed, or add PHPUnit 10 polyfills for the new methods. PHPUnit 10 lacks them and we might still have someone testing locally core with it ATM.
- ๐ฆ๐บAustralia mstrelan
Oh I missed that one. A polyfill makes sense, but I wonder if it's worth it for 8 tests. Do we do still do test runs of 11.x with PHPUnit 10 or is it only 11 now? Do we need to keep BC for people running it in other environments?
- ๐ฎ๐นItaly mondrake ๐ฎ๐น
Do we do still do test runs of 11.x with PHPUnit 10 or is it only 11 now?
Core in GitLabCI is now consistently on PHPUnit 11 (has been so for just a couple of days though, very fresh), but we canโt say for contrib or people running core tests locally or via other CI tools.
- ๐บ๐ธUnited States smustgrave
Seems to be just 4 more instances inside ParamConverterManagerTest
- ๐ฆ๐บAustralia mstrelan
Updated ParamConverterManagerTest and rebased. Not sure who is best to answer #4.
- ๐ฎ๐นItaly mondrake ๐ฎ๐น
No more isType calls reported in the baseline. RTBC
Letโs see what core committers think of #4.
Now core is set on testing with PHPUnit 11 by default, so not sure itโs worth the effort.
- ๐บ๐ธUnited States xjm
Re: #4: It's been suggested that it was incorrect of us to drop support for PHPUnit 10 in 11.2, and that we should re-broaden core's constraint to be
^10 || ^11
. We also discussed changing one of the daily or weekly core environments back to testing on PHPUnit 10 for that (because if we don't run the test suite on PHPUnit 10 occasionally then the lower minimum will be a lie). That said, even if we do both those things, I'd still lean toward "no polyfill" given the low number of usages.