> On Mon, Jun 18, 2012 at 12:59:16AM -0600, Theo de Raadt wrote:
> > Ariane wants to be involved as well, but is still waiting to
> > see how others in the project feel.
>
> I've changed from waiting to being involved.
>
> And in Theo's interest in breaking secrecy: I've stepped down from
> maintaining uvm. Why? Politics between me and Theo.
About 2 weeks ago ariane came to me privately to say this:
1) That 4 weeks ago she had become aware the fork did in fact exist,
counter to previous assertions all the people now working on bitrig
had gotten from their boss Marco.
2) Had _just_ now decided to mention it to me.
3) Also _just now_ had decided to on a desire to work on on both projects.
4) Want to know if this would be ok; if there would be consequences.
And finally:
5) For the last 4 weeks had been too busy to tell us, working 12 hour days.
I said I would not decide, but to ask all the developers.
However, instantly I recognized that the last part (5), about having
too busy to tell us, was not truthful.
The truth is, ariane did not tell us because of _fear_. And I can
understand that, there are many people upset to various levels about
these happenings. However, wrapping that up in a lie about having
been too busy is a not good. Those 4 weeks were spent mulling over
whether to be part of that fork and how to tell us, not by being too
busy with work.
Therefore, ariane, I do not believe that you found out something so
politically big, sat on it for 4 weeks because of being too busy, and
then suddenly decide to disclose it and the desire to be part of it.
And that is not political. I feel that I (and others in the project)
have been lied to in that part.
> I'm unhappy
> with how the situation of the fork was handled.
The situation of the fork was handled entirely by Marco Peereboom --
your boss.
> I've been collateral in
> the whole matter twice and taken it in stride.
Yes, we are all blameless. Especially people at that company, who all
claim they got too busy to tell others that they were too busy.
> I've expressed interest
> in the fork and am now suspect/tainted.
Certainly you are: You misled us.
> Third time's the charm.
> Discussions between me and Theo now trigger anger with both of us,
> which is not conducive to OpenBSD or our fellow developers.
I was not angry in my mail -- I was truthful and exact.
A diff was sent which adds a non-standard flag to mmap() to accelerate
realloc() performance. For years the project has had an attitude that
adding extensions to standardized system calls should be a last
resort. Rather than discuss this with developers, ariane went and
spent time, and then mailed in a diff -- asking only for an OK. Not
requesting the start of a larger discussion, but only asking for an
OK.
In a reply to that diff, I
(1) explained my continued reluctance for such non-standard flags.
(2) I also explained that this was a poor time to put such changes
into the tree with a coming hackathon, followed by the lock to our
next release soon after.
(3) I also then explained that due to recent events (recently two,
serious repairs had to be made to ariane's uvm changes without ariane
being around), I am pessimistic about the commitment level for such
big changes in the tree. Normally a way around this is to test them
as uncommited diffs in the snapshot builds, but I only do that for
people who I totally trust (one reason is that mistakes can be quite
costly, as I can damage 12 build environments in one go), and quite
frankly, I do not trust Ariane nearly as much as before.
At that point, Ariane got seriously angry, and has now resigned.
> I cannot commit to uvm under those circumstances.
Unfortunate.
