Optimise post rectors

[carlos-granados]

When I was investigating rectorphp/rector#8793 yesterday I noticed that the NameImportingPostRector post rector could be optimised because it was unnecessarily calculating the use statements for every node so I decided to add a fix. I took the opportunity to look at the other post rectors and found some other possible optimisations. They are small optimisations, nothing really very noticeable, but, as they say, every little helps!

See the PR comments for more info

[carlos-granados]

@samsonasik code updated

[TomasVotruba]

Thanks for the improvement 👍

Could you share few time measurements, as on the other PR?

[carlos-granados]

@TomasVotruba time improvements are not really noticeable, I see the same timings more or less with the previous version and this one (the results vary a little but they vary equally in both versions). It is more a question of trying to be optimal overall so that all these little increments don't add up

[TomasVotruba]

We're adding a state to readonly service here, which I don't see as a good practise.
I'd prefer to have measurable improvement, to balance the added complexity. We've added few "optimizations" in the past, but they added more complexity than real value so they were removed later.

[carlos-granados]

I have tested this without using parallel, which gives more stable results that don't vary too much and I can see numbers similar to these:
Without the optimizations:
bin/rector --clear-cache 97.41s user 2.47s system 98% cpu 1:41.25 total
With the optimizations:
bin/rector --clear-cache 97.26s user 2.41s system 98% cpu 1:41.06 total
This represents an improvement of around 0.3%, so not very significant.
I would say that we should probably keep the changes to UnusedImportRemovingPostRector and UseAddingPostRector as they don't really add more complexity and I'll let you decide about the other two

[TomasVotruba]

Thanks for checking 👍

I would say that we should probably keep the changes to UnusedImportRemovingPostRector and UseAddingPostRector as they don't really add more complexity and I'll let you decide about the other two

You're right. Let's ship this. Could you rebase the PR?

[carlos-granados]

@TomasVotruba rebased

[TomasVotruba]

Thank you 🙏

[samsonasik]

another PR similar for this, for ClassRenamingPostRector

• [AutoImport] Optimize ClassRenamingPostRector with stop on enterNode() #7295

[github-actions]

This pull request has been automatically locked because it has been closed for 150 days. Please open a new PR if you want to continue the work.