On 11/08/17 19:30, Jaromil wrote:
> On Fri, 11 Aug 2017, Daniel Reurich wrote:
>
>> Ok, this is getting way out of control and it's turning into huge
>> energy vacuum. I thought working on Ascii was our priority now??
>> Let's not get so distracted by this and get peoples backs up.
>
> no, the priorities have been agreed by all developers participating in
> meetings and are being processed. the task I mention is a blocker to
> another priority, that of modernizing the CI to be able to:
>
> 1) activate mirrors as offered by various institutions (see notes)
>
Not related to CI at all. Amprolla3 and new packages server is already
up and will allow for rsync pull of the mirrors

> 2) switch to amprolla3 in production as its throughly tested and its
>    rewrite has been reviewed already by three programmers

Not related to CI either. All the servers for this are in place, just
waiting on direction from you and nextime re signing keys.

>

> 3) manage more transparently the needs of inclusion of developers
>    adopting package updates

Not related to CI. The existing functionality allows for this, we just
need more people spending time helping with this.

>

> 4) fix a current bug in the present CI which is not yet disclosed for
>    our own security

Well that's the first I've heard of an undisclosed bug, and unless it's
disclosed we can't fix it.

>

> 5) I may be forgetting other blockers this task has, but referring to
>    previous meeting notes should be easy

>
>> My understanding is that the ci system was mostly where I and nextime
>> lead by virtue of the fact nextime built and hosts it and I'm the most
>> experienced with it and have a similar understanding of how it works to
>> nextime.
>
> this is a false assumption. As of today there are more people who are
> very experienced with it, to the point they have rewritten the whole
> thing, drafted the first Devuan Improvement Proposal with extensive
> documentation on its aims and course of action, a sort of
> documentation was not even present before.

Who exactly?? KatolaZ and parazyd have both shown no interest in
actually understanding how it works in depth and this is the reason they
propose a solution that doesn't solve any real problem.
>
>> That doesn't preclude others being able to contribute enhancements
>> or make suggestions, but forcing a solution like scorsh on us isn't
>> a wise move. And in any case, talking to me and nextime about the
>> issues, before creating a solution such as scorsh might have avoided
>> this standoff.
>
> noone is forcing anything, the process is:
>
> 1- there is a problem

Then define it and provide evidence. In my response to KatolaZ I again
shoot down his misconceptions driving the development of scorsh

> 2- someone works hard on a valid solution
It's not a solution to a problem we have, and it's an ill conceived
solution due to it's NIH design

> 3- there are no alternatives to the solution
The problems haven't been clearly defined and investigated so how can we
have developed a solution??

>
> saying at this point that one "has to talk to you or nextime" who
> namely have no time, is blocking nonsense.

I have made myself very available, and blocked no one. I didn't tell
KatolaZ he couldn't write a solution, but I have continued to express he
is on the the wrong path due his lack of understanding of the existing
system

>
>> Jaromil your not helping. Stop throwing your weight around and putting
>> pressure on nextime. It won't solve things, just adds to his stress
>
> no, I know well enough Nextime.
>
> His first and only priority is Devuan's well being.
>
Then why do you arbitrate for a solution to imaginary and non-problems
when nextime says he will provide a solution??

> We are also lucky for that, since he would never-ever pull any plug on
> anything Devuan related, not even if one of us builds an hate-page for
> him. He is not a kid at all.
>
I agree, but this attitude against him is no doubt disheartening and
demotivating. Credit where credit is due, but show plenty of grace as
well, and stop calling to evict those who have done the hard yards and
have the experience because we point out the short sightedness in poorly
conceived solutions to non problems.

> But as Katolaz points and yourself, there is a single point of failure
> plus a nonsensical blocking attitude now which we need to deal in a
> way or another if it persist, for the well being of the Devuan
> project, which is a collective project.

It's neither a blocking or non-sensical attitude, but push back against
an ill conceived solution. I didn't push back against amprolla3, but
welcomed it, with fair criticism where I felt it needed minor tweaks.
I'm more of a sysadmin and therefore I look at the broader picture of
how things interact and push that perspective into the mix.

Are you talking of people or servers? Servers are relatively easy to
replicate and replace. It's the people that are of concern to me, and
you know this, and I've said this already many times to you in previous
discussions. Unfortunately I can't be cloned and I object to any
attempts to do so.
>
>
>> We are not dealing with a broken CI system here, it still works fine for
>> Devuan.
>
> no, it is broken.
>
>> KatolaZ, I'm sorry, but I still fundamentally disagree with scorsh, and
>> see it as a systemd like exploit being foisted on our CI system. It's
>> bad engineering in this context.
>
> this needs to be a specific comment on the DIP, with plenty of detail
> and the opportunity of Katolaz to reply in detail and/or adjust
> things.
>
I tried and just got shit on for it.
>
>> I strongly object to scorsh having it's own
>> authentication/permissions system built in. We already have a
>> sufficient system for that in gitlab and which do not belong in a
>> tool that should be a thin layer between gitlab and the ci system.
>> And if we were to replace gitlab, we should move to a separate tool
>> for providing federated authentication and permissions.
>
> I strongly disagree with this analysis.
On what grounds.

> We need to narrow down our dependencies to git.
What dependencies specifically?

> I won't explore this at lenght since we've done
> it already in meetings and since this e-mail is an emotional
> rabbit-hole.
And that's just an excuse to not lay out the problems perceived or real
in documentation, the one thing I have called for consistently ever
since KatolaZ came up with his solution. It is not an emotional rabbit
hole, it's just thorny because KatolaZ despite his claims is very
attached to his solution, and you back him because you want him to succeed.

Here lies the biggest issue. People that don't understand what
releasebot does, and how it's permissions model works yet want to
replace it with something of their own design.
>
>> I'll start:
>
> why should we wait you to do it?
>
> I find the quality of the work done by other developers here very
> high, perhaps higher than anything else that has been produced so far.
>
then get them to implement the simple fixes I've proposed.

> you are imposing a dependency on your actions to all of us without
> even giving a fair consideration to what has been done.

No, I'm simply saying don't acceot a solution born out of frustration
that came from ignorance of what already is there.
>
> I think that unless you let go with this attitude, we'll have to
> circumvent it. I wonder then if your choice will be to "pull the plug"?

Jaromil stop being so reactionary. You make it sound like you control
Devuan... sorry, but I think you try that sort of thing and you might
find the community kicks back quite significantly

What exactly would I pull the plug on? Besides which, I depend on
Devuan for my client systems these days. You think I'm going to walk
away over a spat on a piece of software?

I could always fork it though - after all I have enough knowledge to
pretty near build the entire stack from scratch - easily... but where is
the fun and what is the point in that?? :-D

The only choice you have is to build consensus of opinion, just like me.
And that is exactly why I write these emails in reaction, because I
know I can argue my case effectively and win on this issue.

Cheers,
    Daniel.

--
Daniel Reurich
Centurion Computer Technology (2005) Ltd.
021 797 722