Alexander M. Turek

What if we stopped registering error handlers, exception handlers or shutdown functions of any kind in FrameworkBundle? We do need this stuff way before the framework is booted, so whatever...

> This would require deprecating the FrameworkBundle configuration settings that allow configuring the error handler though Would it? I mean, FWB can still connect to the error handler registered by...

> To overcome this Problem I needed to put this in the TestCase Class: So the bootstrap one-liner [I posted earlier](https://github.com/symfony/symfony/issues/53812#issuecomment-1962740145) did not help?

Please, no more me-toos here. We know the issue exists and a workaround has been posted. Let's focus on solutions.

> @derrabus what if `KernelTestCase::ensureKernelShutdown()` would restore PHPUnit exception handler as well? Please explore that idea if you like. But since you ask me, I believe that not fiddling with...

Thank you for the proposal. I've missed a baseline feature for a very long time. > In a large existing codebase, it usually not possible to fix all violations at...

Speaking for myself, it would be acceptable if PHPCBF would simply ignore the baseline completely. For me, the baseline is mostly interesting for non-fixable sniffs. If I run PHPCBF, I...

The proper way for Oracle would be to use either a sequence or UUIDs generated on the client.

> Everything is possible of course but it would cause the Oracle implementation within messenger to move far away from the default implementation. Yes.

Extracting that interface would result in a more formal contract however. But what makes this weird is that the _validator_ component would expose a service that estimates password strength. I...