An argument with name "class" already exists.

[code2prog]

• Laravel Version: 8.32.1
• PHP Version: 7.4.16
• Database Driver & Version: -

Description:

After upgrade laravel/framework from v. 8.31.0 to 8.32.1 throw error in console An argument with name "class" already exists. after dump-autoload. Reverting to a previous version fixes the problem

Steps To Reproduce:

composer upgrade laravel/framework

[driesvints]

I'm pretty sure that this is caused by #36513 but I don't see how that's breaking. Are you modifying the command in any way? We'll need a lot more info here. How are you calling your seeder classes? Are there internal artisan calls in your code/tests?

/cc @jasonmccreary

[code2prog]

@driesvints I didn't change anything. The only thing that might have to do with it is tenancy/multi-tenant. But this package was not updated at the same time

[driesvints]

@code2prog ah I wasn't asking if you changed anything. I'm asking how you're calling the db:seed command.

[code2prog]

tenancy/multi-tenant#988
It seems that here is the reason

[jasonmccreary]

@driesvints, if you're worried about conflicts with community packages, we could try to rename it to seeder. But what's to say another package doesn't call it that. 🙃

[driesvints]

Not entirely worried, was just wondering how your change broke certain use cases.

[driesvints]

In general that wasn't part of the public API. If people are overwriting it in that way then we can't protect against that.

[jonquintero]

That is not true at all, rather this new change overwritten something and broke a package

[driesvints]

@jonquintero we can't control the ways in how someone overwrites this. If you're overwriting a command and adding arguments of your own then you should be aware that an upstream change can break it. This is different from, for example, method signatures which are part of the public API of a codebase.

[jonquintero]

@driesvints I understand your point and it is totally valid, but my point is that this new change #36513 overwritten something that the mentioned package had been using for a long time and broke it, reverting to a previous version fixes the problem (8.31.0). What's great about all this? That forces developers to update their packages and not forget about them

[driesvints]

@jonquintero you can't expect us to keep in mind every single way someone overwrites something in the framework. If we had to do that we had to do new major releases each week just to make sure that no small change breaks something in a third party app.

[jonquintero]

@driesvints I know and you are right, that what I said: What's great about all this? That forces developers to update their packages and not forget about them

[driesvints]

ah sorry, I misunderstood you there 👍