First, thank you very much for your reply.
Indeed, even if I do not write English fluently, I could have read it
again before posting. I now see my horrible mistakes, and I beg your
pardon.
Yes, this is a one more tool like ansible/chef/ puppet/propellor/…
The fact that there are still so many similar tools today may mean that
there is a big need, but that there is no simple and modular solution
to outperform others.
I'm not claiming that kt-update can do it, at best it could be used as
a module by these big platforms.
I developed it because I did not find any simple solution to this
simple problem: manage the list of expected packages on each of our
machines.
Second, I don't like to fork, and I don't want to fork. I like to
contribute and I want to contribute.
I don't think I forked cron-apt since I rewrote it more than 90%,
giving it functionality far beyond its original objective. The
conversation I had with Ola Lundqvist was pleasant, and he admitted
that he "can (almost?) not recognize any code that" he "have written",
but only a few lines from him and Marc Haber.
I wanted to do well. If you prefer to keep track (homeopathic?) Of the
other authors, I completely agree.
I think I am a fairly discreet person, who values ideas more than the
people who carry them. I would like the ideas/features brought by
kt-update to be available upstream. I don't care which names can be
associated with it. I would prefer mine not to be there, but you have
to assume and expose yourself a minimum to be contacted to advance
ideas.
Finally, I am also very bad at "marketing".
And making this tool available upstream, even if it allows people to
easily switch from a Debian configuration to a Gaussian configuration
(or vice versa), is absolutely not a priority for my company (which
could "market" it).
I'm glad I got an answer from you, an apt developer, and sad that
you don't want to sponsor the stuff.
Could you change your mind?
Could anyone else be interested in having a simple implementation of
these features in Debian?
Best regards & Happy new year everyone !
---
Jean-Jacques Brucker
Le Mon, 30 Dec 2019 15:32:45 +0100,
David Kalnischkies <david@kalnischkies.de> a écrit :
> Hi,
>
> On Mon, Dec 30, 2019 at 11:51:53AM +0100, Jean Jacques BRUCKER wrote:
> > What's next ? What did I do wrong ? May someone with super powers
> > consider this package, and tells me if it can be integrated
> > upstream
>
> First of all: I am likely the wrong person to reply to this mail.
> Second: Reading someone say bad stuff about your "baby" is hard, but
> try to not get defensive – it is not an attack!
>
>
> "lightweight distribution management" – so that is some tool to be
> used by the release team of a distribution? No, it isn't, but even
> though I am an APT developer, I have no idea what it does.
>
> I /think/ based on the snippets I read that it is an apt wrapper
> script to help in creating package installation profiles and to keep
> machines following/switching between these profiles. Sort of like
> ansible/chef/ puppet/propellor/… but for packages only – but I am not
> sure and I got that impression mostly from your reply to Paul who was
> also asking what this thing does although in a less blunt way than I
> am now.
>
> Every other sentence I read has at least one typo in it – and I say
> that as a non-native English speaker who is an expert in producing a
> lot of mistakes in his writing. Ask your users for help! There is
> also the debian-l10n-english@ mailinglist for the last few points of
> excellence. Having "Contain a tool than enhance cron-apt and manage
> an apt sources.list with some extensions." as the start of the long
> description of a package is wrong on so many levels that I can't
> enumerate them all.
>
> I mentioned already, I am APT developer, so I am super biased in this,
> but stating even in the RFS that whatever kt-update does (still not
> sure) should be done by apt, but you don't want to contribute to apt
> and therefore forked (= not contributing to that one either) another
> wrapper of apt is not instilling a lot of confidence in me.
>
> And I think you mean well, but what the heck means "AUTHORS files has
> then been cleaned"? That sounds like you removed names from that file
> which makes my stomach hurt as contributions by people never really
> fully disappear (even if no line they wrote survives, the idea they
> had might).
> BUT even if that is all good and fine: That is one of the first
> things you wanna tell a prospecting sponsor about your package? Not
> about some killer feature, but about a dubiously clean text file?
>
>
> So, long story short, this all heavily lacks a bit of quality in the
> documentation or if you prefer "marketing" – completely ignoring what
> it does on a technical level because nobody (at least not me) will
> bother looking any closer. Finding a sponsor in this state might be
> even harder than it usually is.
>
>
> (And no, even if that package would be in tip-top shape, I am not
> available for sponsoring it or in all likelihood for further reviews.
> Supercow of 'apt moo' fame would be furious with me… 😉)
>
>
> Best regards & Happy new year
>
> David Kalnischkies