Re: dgit advocacy again (Re: Survey results: git packaging practices / repository format)
Andreas Tille writes ("Re: dgit advocacy again (Re: Survey results: git packaging practices / repository format)"):
> On Wed, Jul 03, 2019 at 11:41:34AM +0100, Ian Jackson wrote:
> > People like you are precisely the intended users of "dgit push --gbp".
> >
> > I think if you used it you would help Debian users and contributors.
> > (You would slightly improve your own life too, but maybe not by enough
> > to be worth the bother of learning a new tool.)
>
> May be we should sit down next to each other at DebConf for one hour and
> you watch my workflow and tell me what can be enhanced by dgit.
Sure. But as I say, the *main* advantage is for other people, not for
you. You publish your git history in a better way.
> > As https://wiki.debian.org/DgitFAQ has it
> >
> > | Q: Why should I bother to learn how to use dgit?
> > |
> > | If you incorporate dgit push-source into your workflow, then you
> > | will be providing your git history to Debian's users and
> > | downstreams in a way that is more useful to them than adding a
> > | Vcs-Git header.
> > ...
> > | There are also minor conveniences for maintainers ...
>
> Sorry, but I do not understand this: I'm extremely picky about Vcs-Git
> headers (since some Blends tools are using these).
Certainly you should add a Vcs-Git header. dgit does not change that,
for packages already using git somehow. I have added the word "just"
so now it says:
| more useful to them than just adding a Vcs-Git header
Hopefully that helps clarify that adopting dgit does not mean
dropping the Vcs-Git header.
> I've read the FAQ (once, not recently) and was not convinced. As I said
> I think sitting down next to you at DebConf might be the best way to
> learn the advantages.
dgit push-source is slightly more convenient than the dput based
system and it catches slightly more errors and does slightly more
automatically. But, it is a fairly minor benefit *to the maintainer*.
As I say, the main advantage is for users and downstreams and certain
kinds of NMU campaigns. Mainly, you would help people other than you.
It is IMO quite easy to persuade users and downstreams of the value of
dgit clone. For the target audience it is simply a massive
improvement over all previous tools. Even if the maintainer didn't
use dgit push, so "git log" and so on give poor results.
The main difficulty now with getting dgit push adopted is that
maintainers don't see the benefit *to themselves* and it is always
harder to get someone to see a benefit *to someone else*.
Especially, it can be hard to convince people of the reality of a
benefit to someone else, which they perhaps ought to provide, if it
would involve them having to change the way they do things...
Anyway, I'm very happy to talk to you (or anyone) about your own
situation in detail IRL, and also open to hearing from anyone who has
good ideas about addressing this adoption problem.
> > I have said before that I think using "dgit push" (where possible) is
> > an ethical imperative. (I should clarify that I *don't* mean that
> > people who aren't using "dgit push" yet are bad people. Life is so
> > full of ethical imperatives that no real human could meet them all,
> > and of course Debian's right to call on volunteer effort is limited.)
>
> Question: I'm eating less and less meat (not real vegetarian but with a
> very positiv feeling that this might be an ethical imperative as well).
> Which ethical imperative would you consider stronger: Becoming
> vegetarian or using dgit? ;-)
I left this one to the end :-).
That is an interesting question. I doubt it is really possible to
compare things in that way. But certainly reading articles about
accelerating deforestation (and attacks on indigenous land rights
etc.) makes me think that I will try to avoid eating much Brazilian
beef.
Regards,
Ian.
--
Ian Jackson <ijackson@chiark.greenend.org.uk> These opinions are my own.
If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.
Reply to: