G. Branden Robinson writes ("Re: Include git commit id and git tree id in
*.changes files when uploading?"):
> I personally find the disputes over this topic difficult to follow, and
> would prefer that the arguments be presented in the form of syllogisms.
This area is complicated. (Debian source packages are complicated and
weird, the ways that maintainers create them are complicated[1], and
Debian's internal systems and processes are complicated[2].)
The necessity for an actual transition - running both systems in
parallel - makes things even more complicated. As always, before we
can get rid of the unnecessary complexity, we need to add a layer
around it.
Worse, many of the people I'll call "dgit critics"[3] have absorbed
fundamental and critical misunderstandings. When challenged this
often leads to a vigorous defence of the misunderstanding, based on
more and more falsehoods.
I'm not surprised you find this difficult to follow. It often doesn't
make any sense.
I'm sorry to say that I see Otto doing this in this thread. You
should take anything he says with a great helping of salt, even if he
confidently asserts it. He writes a great deal, as well, so there is
no chance of refuting everything that's wrong. You might be better
off ignoring his messages.
I'm happy to try to explain things, for example if you would like to
try to pose questions.
Ian.
[1] https://wiki.debian.org/GitPackagingSurvey
[2] Debian is a large organisation trying to solve many fundamentally
difficult problems all at once. It's not surprising that this makes
things complicated.
[3] I do not intend to tar with this brush everyone who is sceptical
about dgit. Scepticism is the right approach to any new technology,
and dgit is only 12 years old.
--
Ian Jackson <[email protected]> These opinions are my own.
Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk,
that is a private address which bypasses my fierce spamfilter.
