Print RUBY_DESCRIPTION when Puma starts

[JacobEvelyn]

This commit changes the runner to print RUBY_DESCRIPTION instead of a manually-generated string for the Ruby version. Other popular gems like Sidekiq similarly print RUBY_DESCRIPTION. One specific motivation for this change is that RUBY_DESCRIPTION indicates whether YJIT is enabled in MRI.

This commit also deprecates the Puma::Runner#ruby_engine method, which will be removed in v7.

Description

My team was testing YJIT and was frustrated by the fact that Puma does not indicate whether YJIT is running, while other gems like Sidekiq do.

I welcome any suggestions for this PR.

Your checklist for this pull request

• I have reviewed the guidelines for contributing to this repository.
• I have added (or updated) appropriate tests if this PR fixes a bug or adds a feature.
• My pull request is 100 lines added/removed or less so that it can be easily reviewed.
• If this PR doesn't need tests (docs change), I added [ci skip] to the title of the PR.
• If this closes any issues, I have added "Closes #issue" to the PR description or my commit messages.
• I have updated the documentation accordingly.
• All new and existing tests passed, including Rubocop.

[dentarg]

Mind showing some before and after examples?

[MSP-Greg]

Current:

[172745] Puma starting in cluster mode...
[172745] * Puma version: 6.4.2 (ruby 3.4.0-p-1) ("The Eagle of Durango")

PR:

[172898] Puma starting in cluster mode...
[172898] * Puma version: 6.4.2 ("The Eagle of Durango") (ruby 3.4.0dev (2024-06-05T20:27:00Z master 0b31986909) [x86_64-linux])

After the PR, the line is 127 characters wide. Splitting it to another line might be an idea, but doing so could (maybe) mess up code that is parsing it...

[dentarg]

Ideas/alternatives:
Should Ruby has its own line instead?
Should we only add YJIT yes/no?

[ioquatix]

RUBY_DESCRIPTION is the correct choice, and it should probably be on its own line. I recently adopted it in Falcon too.

[dentarg]

This method is public, so we should deprecate it, not remove it.

$ ruby -rpuma -rpuma/runner -e 'p Puma::Runner.new(Puma::Launcher.new(Puma::Configuration.new)).ruby_engine'
"ruby 3.2.4-p170"

Unless this change should happen in Puma v7, but even if so, it is nice to deprecate things we intend to remove.

[JacobEvelyn]

$ ruby -I lib -rpuma -rpuma/runner -e 'p Puma::Runner.new(Puma::Launcher.new(Puma::Configuration.new)).ruby_engine'
Puma::Runner#ruby_engine is deprecated; use RUBY_DESCRIPTION instead. It will be removed in puma v7.
"ruby 3.3.2-p78"

[JacobEvelyn]

@dentarg let me know what I should do to move this forward. Thanks!

[dentarg]

I think this makes sense. Is there any existing tests for this part of Puma that should be extended?

[dentarg]

I recall some change to the Puma logging in the past and I remember @nateberkopec being conservative about it. It would be great if he signed off on this.

If I have time I can try to look for that past discussion.

[JacobEvelyn]

Is there any existing tests for this part of Puma that should be extended?

I didn't see any, but let me know if you think I should try to add one.

[JacobEvelyn]

I recall some change to the Puma logging in the past and I remember @nateberkopec being conservative about it. It would be great if he signed off on this.

@nateberkopec any thoughts on this?

[JacobEvelyn]

This change feels fairly uncontroversial to me; what can we do to move it forward?

[dentarg]

I recall some change to the Puma logging in the past and I remember @nateberkopec being conservative about it. It would be great if he signed off on this.

If I have time I can try to look for that past discussion.

Maybe I misremembering, I can only find this comment #2817 (comment) and it is probably questionable to call this logging 'behavior'.

We have changed the logging before in a minor release, ee3341d67, so this should be fine. I'll merge it if you rebase @JacobEvelyn

[dentarg]

I'll merge it if you rebase @JacobEvelyn

I clicked the Update branch button so no need

[nateberkopec]

Yeah thanks for checking w/me but in this case I think we're good.