Alexander M. Turek
Alexander M. Turek
I'm not comfortable with merging this change into 3.8, tbh. As you've already noticed, SQLite's implementation of integer PKs is a bit odd. The main issue is that SQLite does...
> throw an exception when a primary key is not an int in sqlite We probably don't want this. If someone uses BIGINT PKs on their entity model, e.g. because...
> Something like "You've already logged in somewhere else, please try again to log in here". That's not what an invalid CSRF token is about, but I guess that only...
> The Symfony 5.4 EOL is only in 1 year. The active support for PHP 5.4 ends this year in November. Beyond that is pure security fixing and all non-critical...
> How often have you _modified_ that code in the last year ? It's more like, how often do we add new polyfills and have to jump through more hoops...
> Let me close as I don't think we'd want to bump existing polyfills. … like never? I think that's a mistake. Really, what's the point in maintaining a polyfill...
Can you look into the static analysis and test failures, please? 🙃
What's the problem with those `setUp()` methods? I don't see this change as an improvement, tbh.
Again, I don't see the improvement. The tests are fine, I wouldn't change them.
Let's close this PR. Thank you for your proposal though.