Build & Packaging (and Docs) Working Group Minutes

I added a comment in github issue 2364 about using samovar , which is battle tested.

2 Likes

I ran the async meeting one day late in April and nobody showed up. Here are the minutes, which are essentially my own status update for rebar3 readiness on OTP-28.rc2 build-and-packaging-wg/minutes/2025-04-02.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

I will not be around for the May 6 meeting, but I’ve pre-scheduled messages kicking off async meetings the usual way for the working group to carry on.

2 Likes

I was on a month-long sabbatical for May and wasn’t around to run the async meeting. I sent a scheduled message to kick it off, but nobody showed up: build-and-packaging-wg/minutes/2025-05-06.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

We’ll see if tomorrow makes it 3 months in a row of nobody showing up.

Minutes for June: build-and-packaging-wg/minutes/2025-06-03.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

Highlights, for which you can get expanded notes in the link above:

  • how to package for MacOS?
    • a good discussion on what may be required to avoid security prompts when shipping an end-user app to MacOS.
  • rebar3_hex 7.0.11 released with a fix for the corner case that allowed for publishing with git dependencies.
  • Anon funs in ex doc config that require elixir scripts to work; can anything be done?
    • no solution is offered there, although there are traces of paths forward.

I’m glad to say we didn’t end up with a third dead month in a row!

Minutes for July, which it turns out got back to no participation:

August 2025 minutes: build-and-packaging-wg/minutes/2025-08-05.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

  • Rebar3 had breakage on Windows
    • The main thing that broke for us was related to two lines:
      1. erlang:system_info(system_architecture) stopped reporting ā€œwin32ā€ and now does ā€œwindowsā€ so we had to change some sub-calls to os:type()
        in places that can’t do that switch (eg. Rebar3 hooks) where we needed to update hooks (see Update escriptize hook to work with newer Windows versions by ferd Ā· Pull Request #2964 Ā· erlang/rebar3 Ā· GitHub)
      2. This latter one implies that a lot of compile hooks may be broken, but that’s due to OSes changing things afaict.
    • nobody else reported breakage (as nobody else was in the discussion)

Note that this is now the trend for monthly B&P participation:

  • April: nobody
  • May: nobody
  • June: 9 participants
  • July: nobody
  • August: 2 people, including myself, but a single person speaking.

At this point the trend is that we managed 1 meeting with content outside of my own as group-runner in the last 5 months. I’d consider the B&P group to be unhealthy, though I’m aware of the vacation effect of things.

September 2 meeting notes: build-and-packaging-wg/minutes/2025-09-02.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

We had attendees again, and 4 discussions:

  • Funding from Shiguredo, specifically rebar and hex
  • Rebar3 kickstarter campaign
  • Rebar3 call for more maintainers
  • Swift Interop

The rebar3 kickstarter campaign is at https://www.kickstarter.com/projects/peerstritzinger/rebar3-integrating-with-erlang-otp?ref=project_build

The Shiguro money as of now is not intended for the kickstarter, which has been judged too vaguely defined for them. Further talks are expected.

As for the call for maintainers, this is my message from the EEF Slack:

I’m doing a call for maintainers on Rebar3 (and Relx). Tristan has had limited time to contribute for a good while now, and my own contributions have only gotten sparser, while issue reports and calls for optimization keep happening.
At this point I feel it’s a disservice to users not to be more visible on the lack of time/maintainership available, even if I still plan to write patches when convenient for me personally, and be around to help debug stuff.
However, I’ve been at it for probably more than 10 years now and haven’t really been working with Erlang professionally for 5 years (it’s still my hobby language) but that makes me very much a lot slower to find and feel issues, and I do see in the recent tickets that I’ve been slipping and gradually losing the context on some elements.
Getting more people to lend a hand and for any sort of knowledge transfer to happen would be good for sustainability and definitely overlaps with the Kickstarter campaign as well.

Thanks for the participation!

Here are the minutes for October 2025: build-and-packaging-wg/minutes/2025-10-07.md at master Ā· erlef/build-and-packaging-wg Ā· GitHub

Highlights:

  • static linux builds
  • OTP build artifacts
    • the OTP team would like to find ways to centralize how build artifacts are done, to be outside of erlang/otp itself, but blessed by the team
    • there’s lots of targets and challenges to be had there, I’m assuming community collaboration is desired.

Note: I missed the meeting as the working group leader. I’ve later announced that I’ll be going back to school part time in the new year and will have to step down from working group leadership to be able to focus whatever is left of my OSS time on the rebar4 kickstarter support. Jonathan Otsuka has offered to become to new working group chair. I’ve sent a slack message to the EEF board, but haven’t heard back from them yet.

1 Like