Skip to content

Conversation

@beyhan
Copy link
Member

@beyhan beyhan commented Dec 17, 2025

@stephanme
Copy link
Member

LGTM: the only thing we could add is a sort of release plan for expectation management:

  • v5.0.x - experimental release intended to get broad feedback by users, incompatible changes may happen
  • v5.1.0 - first non-experimental release (some call it GA)

Having a 5.0 release (instead of just a branch) may help to get adoption.

@beyhan
Copy link
Member Author

beyhan commented Dec 18, 2025

LGTM: the only thing we could add is a sort of release plan for expectation management:

  • v5.0.x - experimental release intended to get broad feedback by users, incompatible changes may happen
  • v5.1.0 - first non-experimental release (some call it GA)

Having a 5.0 release (instead of just a branch) may help to get adoption.

@stephanme thank you for the feedback! It is addressed with f31def3

@beyhan beyhan changed the title [WIP] RFC for the Java Buildpack reimplementation in Golang RFC for the Java Buildpack reimplementation in Golang Dec 18, 2025
@beyhan beyhan added the rfc CFF community RFC label Dec 18, 2025
@beyhan beyhan requested review from a team, cweibel and rkoster and removed request for a team December 18, 2025 11:46
stephanme
stephanme previously approved these changes Dec 18, 2025
@mymasse
Copy link

mymasse commented Dec 18, 2025

I think releasing the 5.0.x as an experimentla version is risky, some people might automatically update to this version without knowing this is an experimential version. RC classifiers should probably be used. Just think for example CF-deployment might autobump to version 5 is this really what we want?

@stephanme
Copy link
Member

I proposed the experimental 5.0 version for transparency. We can name it as we want - the first releases are risky. Of course they will pass cf-deployment validation and CATS - but customers will find gaps and things that don't behave as before.

An experimental 5.0 makes clear what to expect and probably there will adjustments to the the list of supported features and deprecations. If it helps, we can also name it beta instead of experimental.

@mymasse
Copy link

mymasse commented Dec 18, 2025

I proposed the experimental 5.0 version for transparency. We can name it as we want - the first releases are risky. Of course they will pass cf-deployment validation and CATS - but customers will find gaps and things that don't behave as before.

An experimental 5.0 makes clear what to expect and probably there will adjustments to the the list of supported features and deprecations. If it helps, we can also name it beta instead of experimental.

From the discussion on Slack I now get where this is coming from. Where would that name show up however?

@Gerg Gerg moved this from Inbox to In Progress in CF Community Jan 6, 2026
@Gerg
Copy link
Member

Gerg commented Jan 6, 2026

Given that there don't seem to be any open issues with this RFC and work for this is progressing at great speed, the TOC decided to move this to the final comment period. You now have two weeks to leave any comments or forever hold your peace.

@stokpop
Copy link

stokpop commented Jan 13, 2026

Hi, thanks for updating the Java Buildpack. This might also be a good time to review the contents of the build pack as it contains many dependencies... for example the takipi agent where the github does not show activity for the last 12 years: https://github.com/takipi/debugAgent. We also looking forward to include the newest java releases, including Java 25. Is there a way to see what is a good "lean" starting point?

@beyhan
Copy link
Member Author

beyhan commented Jan 14, 2026

Hi, thanks for updating the Java Buildpack. This might also be a good time to review the contents of the build pack as it contains many dependencies... for example the takipi agent where the github does not show activity for the last 12 years: https://github.com/takipi/debugAgent. We also looking forward to include the newest java releases, including Java 25. Is there a way to see what is a good "lean" starting point?

Hi @stokpop,

Thanks so much for bringing this up! We discussed your suggestion during yesterday's TOC meeting, and we really appreciate you highlighting this. The Takipi agent has already been removed from the Go-based Java Buildpack. For a complete overview of what's changed, please check out this comparison document between the Ruby and Go-based implementations. The Go-based version has cleaned up several unmaintained dependencies as part of the modernization effort. Regarding Java 25 support we're definitely looking into including it into the first Go-based release.

Since we want the Go-based buildpack to be a drop-in replacement, we'd love to hear your suggestions on any other unmaintained components that should be removed for the Noble version of the Java Buildpack. Your feedback would be valuable you can provide it for example by opening an issue for the buildpack. Also, if you get a chance to test the Go-based version, we'd greatly appreciate any validation feedback you can provide! For this you could use the java buildpack Slack channel.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

rfc CFF community RFC

Projects

Status: In Progress

Development

Successfully merging this pull request may close these issues.

6 participants