git/git.github.io

Any comment about upcoming Git Rev News edition 118

Closed this issue · 24 comments

chriscool commented

A currently mostly empty draft is there:

https://github.com/git/git.github.io/blob/master/rev_news/drafts/edition-118.md

Feel free to comment in this issue, suggest topics, suggest persons to interview, or use the edit button (that looks like a pen) to edit and create a pull request with the changes you would like.

Let's try to publish this edition around the end of December 2024!

Thanks!

cc @jnareb @mjaix @sivaraam @gitster @stepnem

chriscool commented

Two Outreachy interns have been selected and should start working on their internship on December 9th. They have sent emails to the mailing list:

  • 👍1
  • 🎉1
jnareb commented

@jnareb maybe we could add a link to this LWN article about adding Rust to Git?

Yes, I agree that would be a good idea (perhaps with reference to Edition 82?).

  • 👍1
chriscool commented

@jnareb maybe we could add a link to this short one about upcoming Git 2.48 too?

jnareb commented

@jnareb maybe we could add a link to this short one about upcoming Git 2.48 too?

That's a good idea. Thanks @chriscool for the suggestion.

By the way, do you have any reference to discussion about choosing "Meson to be the primary build tool for Git in the future"?

chriscool commented

@jnareb I think most of the discussion about Meson happened in that thread:

https://lore.kernel.org/git/cover.1727881164.git.ps@pks.im/

Do you mean that I should cover this in an article?

jnareb commented

@jnareb I think most of the discussion about Meson happened in that thread:

https://lore.kernel.org/git/cover.1727881164.git.ps@pks.im/

Thanks, though that thread looks like quite far in the process, being v8 of "Modernize the build system" series...

I have just now remembered that the idea of Git switching to a modern build system was discussed on Git Contributor's Summit 2024: https://lore.kernel.org/git/Zu2E3vIcTzywWOx3@nand.local/

  • Patrick: three different build system; should get rid of at least one of them
    • Should delete autoconf because it's not really maintained
    • Think about a proper build system
    • Obvious choice: cmake
      [...]
  • Patrick: [Makefiles are] Non-standard
    • Meson is nicer than cmake as an alternative

I think I was one of people proposing autoconf as an addition to configuring Git via Makefile variables very early in the life of Git as a project - it made it much easier to create RPM spec file when compilation follows traditional 'configure && make && make install' steps.

But the Linux ecosystem have much changed since then.

For example Meson is 8 years younger (2013) than Git (2005). CMake is older than Git (2000), but not much.

I guess Build2 is too new for the consideration...

Do you mean that I should cover this in an article?

If you don't have a better idea for a topic, and have time for some research - it would be IMVHO interesting topic.

jnareb commented

I see that the following promise from Git Contributor's Summit 2024 (https://lore.kernel.org/git/Zu2E3vIcTzywWOx3@nand.local/)

  • Jonathan: Should we do a comparison of build systems in terms of what
    we need from them on the list? Similar to 'Documentation/technical/unit-tests.txt'
    • Patrick: I can write such a thing.

There is now Documentation/technical/build-systems.txt, written by Patrick Steinhardt (@pks-t).

chriscool commented

Actually I already started writing the blob post that GitLab will publish about the upcoming v2.48.0 and we might talk about the Meson build system if it is part of that release (otherwise it will hopefully be part of the next one). So I have started researching the subject a bit, and yeah around 2006 it looks like you introduced and worked on the configure.ac file to add autoconf build support to Git.

I will see if I have time to work on an article for Git Rev News about it...

jnareb commented

@chriscool , whether the information about modernizing the build system, choosing Meson, and adding support for it will be on GitLab Blog (perhaps in the future), or as an article in Git Rev News - in my opinion it does not matter, it is important that the information is or will be there.

  • 👍1
sivaraam commented

ve nudged Karthik to see if we could get an interview from him. Otherwise, I suppose we'll have no interview for this edition.

  • 👍1
chriscool commented

Thanks @sivaraam ! No worries if we have no interview for this edition.

chriscool commented

Draft email sent: https://lore.kernel.org/git/CAP8UFD3y_m1D4pF1rCZgmW+C_6JYrxjuyUnGAhKBgbek+zBdRQ@mail.gmail.com/

Sorry it took me a long time to write an article about the new Meson build system and related discussions.

chriscool commented

As I sent the draft email only a few minutes ago, I think it's fine if we publish this on Thursday January 2nd 2025 as I say in that email.

  • 👍1
jnareb commented

My links have landed in 2b88cb1 .

  • 👍1
jnareb commented

Sorry it took me a long time to write an article about the new Meson build system and related discussions.

No worry @chriscool . Thanks for a great and well-researched article about the planned use of Meson build system for Git!

I have made a few minor additions to it in c789c28 .

  • 👍1
chriscool commented

@jnareb thanks for the great links and the nice improvements to the Meson build system article!

chriscool commented

I added a small article about our Outreachy interns in 3cf8bf0. In it we say that we are still looking for funding, but I just learned that GitHub will fund the internships. I will change the article to thank GitHub.

chriscool commented

Small article about our Outreachy interns updated in bdfe3e4.

mjaix commented
chriscool commented

@mjaix thanks for the nice fixes and rephrasings!

mjaix commented
chriscool commented

Thanks @stenem fro your fixes in 90bbc5c.

I am going to publish this edition now. Further fixes can be made later if necessary.