Manual merge of 3.0.x into master

[greg0ire]

Blocked by phpstan/phpstan#3335

[morozov]

Why do we need to merge two changes at once?

[greg0ire]

What do you mean by "2 changes"? I mean to merge 4 PRs at once. If you are referring to the 4 PRs, I feel safer pushing something and knowing for a fact there will be no up merge problems, which is guaranteed by the green builds and the --atomic flag.

[greg0ire]

@morozov only the 7 last commits are new and need to be reviewed

[greg0ire]

@morozov it's complicated to do the merge up with your changes, I'm not super sure. Can you please do the merge up of 74eca6b ? That way I only have my part to take care about…

[morozov]

@greg0ire not sure I understand the ask. 74eca6b is currently part of 3.0.x. Are you asking me to merge 3.0.x into master? Looks like this is what you're doing in this PR.

As for the change itself, it looks trivial: the method is moved from one interface to another and is implemented in the class which now needs to implement it. The same change already exists in master where it was backported from.

Please let me know how exactly I can help you.

[greg0ire]

@greg0ire not sure I understand the ask. 74eca6b is currently part of 3.0.x. Are you asking me to merge 3.0.x into master? Looks like this is what you're doing in this PR.

No I'm just asking you to merge 74eca6b into master. You don't have to handle the merge for my Psalm related changes, but only for your changes.

As for the change itself, it looks trivial: the method is moved from one interface to another and is implemented in the class which now needs to implement it. The same change already exists in master where it was backported from.

Please let me know how exactly I can help you.

I'll give it another try since it does look trivial, but 74eca6b might have ancestors that are not merged either, not sure if the conflict was about them or 74eca6b

[greg0ire]

Gave it another try, and I'm still uncomfortable with this. Please try the following on master:

git merge 74eca6ba7fdd6a48af0c450a486d1f77f950a041
git push fork
* review*
git push origin

EDIT: here is a comparison: master...74eca6b
As you see, it's far from being just that one trivial commit 😉

[morozov]

@greg0ire please see #4038.

As you see, it's far from being just that one trivial commit

Yeah, its the fetch mode rework and the corresponding deprecation backports.

[greg0ire]

Thank you!!!

[greg0ire]

@morozov ready for review (the last 5 commits are only on this branch)

[greg0ire]

@morozov can we merge as is? I handled your review in #4040, since it was already wrong on 3.0.x

[greg0ire]

@morozov sorry, please approve again, the last push contained the fixes that were intended for #4040 and missed the last commit without which I don't understand how the build could pass…