Skip to content

Conversation

@nikolaszimmermann
Copy link
Member

No description provided.

@nikolaszimmermann
Copy link
Member Author

  • Wait for tag/24.04 to be finished
  • Wait for multi-arch builds of this PR to successfully finish
  • Prepare draft mail to technology-webkit, announcing the image change (new base image, new sysprof, etc.)

@nikolaszimmermann nikolaszimmermann force-pushed the nzimmermann/prepare-ubuntu-2504-image branch from f30d39e to cee4df4 Compare August 26, 2025 13:59
@nikolaszimmermann nikolaszimmermann force-pushed the nzimmermann/prepare-ubuntu-2504-image branch from cee4df4 to a96f716 Compare September 24, 2025 11:24
@TingPing
Copy link
Member

I guess my question is what is the main value of this bump?

24.04 is an LTS release, so it will have a lot of support for a while, and we want to target that version until April 2027

@nikolaszimmermann
Copy link
Member Author

Good question 😆 I vaguely remember that I could not easily get a newer sysprof version to build, without having to build additional components in new version that the system usually provides.

However, don't take my word on that, it's too long ago I would need to retry.

@TingPing
Copy link
Member

sysprof makes sense, I recently backported it in Yocto and it truly depends on the latest of everything...

I'm not sure that one project is important enough to move everything over though.

@nikolaszimmermann
Copy link
Member Author

I think alone the compiler versions that are more recent are a reason to update since this is meant to be a developer SDK.

It also simplifies using perf for the people that use Ubuntu 25.04 on their host - but that is rather a weak argument 🤣

@nikolaszimmermann
Copy link
Member Author

I discovered three distinct issues related to GCC 14 that I already fixed upstream in preparation for the bump. I'm already using this image since a month on my main development machine without a problem, just to give some additional data points.

I was under the impression that flatpak SDK was also using quite recent versions of everything - so I didn't give this much thought before starting the bump effort.

@TingPing
Copy link
Member

TingPing commented Sep 30, 2025

I think alone the compiler versions that are more recent are a reason to update since this is meant to be a developer SDK.

We have access to basically every clang version. So its just a newer GCC toolchain (losing the older version).

It's not my intention to fight against it, just thinking about reasons for it.

@nikolaszimmermann
Copy link
Member Author

I think alone the compiler versions that are more recent are a reason to update since this is meant to be a developer SDK.

We have access to basically every clang version. So its just a newer GCC toolchain (losing the older version).
It's not my intention to fight against it, just thinking about reasons for it.

That's correct, we can access newer clangs, but not gccs.
I'll bring this up on tech-webkit, for further discussion to make sure we don't miss any important pro/con.

@TingPing
Copy link
Member

TingPing commented Oct 3, 2025

Another pro: You cannot build epiphany main on 24.04 any longer also.

@clopez
Copy link
Collaborator

clopez commented Nov 6, 2025

Ubuntu 25.10 was released a month ago (9th October)

I guess that if we switch the base version, then it already makes sense to move to 25.10 instead of 25.04

@nikolaszimmermann
Copy link
Member Author

Actually we probably need to wait for the next LTS release -- otherwise the APT repos will disappear if supports ends, etc.

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants