Re: Rust packaging approaches workship, Debconf in Brest [and 9 more messages]
On Wed, Apr 23, 2025, at 7:19 PM, Ian Jackson wrote:
> Hi all. Thanks for the positive feedback on my workshop proposal.
>
> I've now registered for DC25 and submitted this proposal.
> Given that the early submission deadline is not until the 28th of
> April, I think probably no-one will look at it until then, so I can
> still edit it.
>
> Here is what I have put in. Any comments welcome.
>
> Also, I have *not* ticked the "video" tickybox, mostly because in
> practice I doubt remote participation would work well. But I don't
> know if that's the right thing to do. Please let me know what you
> think.
I think this makes sense for the intended format and content (with
notes being taken and published afterwards?) - unless somebody where
attendance is important can only make it online (and wants to do so!),
then of course enabling streaming + recording can make sense.
> I say "this is not the Rust Team BoF". I daresay that someone more
> central to the Rust Team would like to run a Rust Team BoF ?
Possibly - I am still not sure how many of "us" will be in attendance,
maybe we should figure this out first ;)
> New Directions in Rust packaging - workshop
>
> Track: Packaging, policy, and Debian infrastructure
> Type: BoF (45 minutes)
>
> There is considerable room for improvement in our Rust packaging
> workflows. In this Bof I hope to understand the constraints and
> consider options, outside of the usual day-to-day of packaging.
>
> Both the Rust Team tooling, and dh-rust, have important
> capabilities, but also significant downsides. In this session I hope
> to get together and explore, how we could do things better.
>
> Specific areas where I think we have opportunities for improvement
> include:
>
> * version dependency handling (automating and removing clerical work)
> * git-native approaches (ie, packaging based on upstream git)
> * workspace-per-a-source-package
I think all of these are desired for either (or both) debcargo and dh-cargo,
the main reason they don't exist yet is lack of people doing design and
implementation work.
> * abolishing feature virtual packages
I think this might be the most contentious point here ;)
> * making sure all the machinery and metadata is within the Debian archive
I am not quite sure what this means - could you maybe expand?
Something that I have pondered doing a session of some form for is
working on improving the Static-Built-Using (/dh-cargo-built-using)
tooling - do you think the Rust-specific part of that would fit as well
as part of the agenda for this workshop here?
The bigger issue there is probably out of scope though, namely getting
all the required stake holders together and start eliminating blockers
on the infra side for proper security (rebuilding) support based on that..
> This BoF may involve vigorous handwaving and/or scrawls on available
> pieces of paper. We probably won't come to complete agreement. But,
> we already have two approaches that coexist relatively well. I'm
> hoping that we can try to converge somewhat.
>
> The discussion will cover areas of technical disagremeent, but we
> will expect everyone to be friendly and constructive. We will make
> space for everyone's contributions, so that we can make the best use
> of everyone's expertise. And, NB: **this is not the Rust Team BoF**
> although of course Rust Team members are very welcome.
all of this seems sensible as well!
Reply to: