Alexander M. Turek
Alexander M. Turek
Looks like your change broke a lot of existing tests and you didn't add any new tests that would reproduce the issue that you're trying to solve. Are you planning...
Hello and thank you for your pull request. I haven't found the time to review it yet, just one thing: We don't accept new features on the 3.10.x branch anymore....
We probably need to define the class conditionally. Without `readonly` on PHP < 8.1 and with `readyonly` on higher versions. 😓
Can we please find a meaningful title for this PR? Thanks.
We have rejected PRs that only change code-style without being able to enfore the new style.
Thank you. I've merged a workaround for the broken DB2 CI. I'd like to see your changes tested against DB2 as well which is why I've rebased your branch. I...
Thank you for your PR. That is an interesting edge case indeed. What about the following scenario: I have a database with two identical foreign keys and compare it to...
> This change is safe for backward compatibility. It is not. You'll break downstream projects that override this very method.
> It would be interesting to take a look at the git history. I did and it seems it has always been this way for over a decade. And tbh,...
> Would it not be better to prevent inconsistencies and start returning NULL in the next major version as a backward incompatible change? No. That's a super subtle change that...