Re: Intent to hijack Bacula (END)
John Goerzen wrote:
> On Fri, May 12, 2006 at 03:13:31AM +0200, José Luis Tallón wrote:
>
> Jose,
>
> Before I comment on a few things, I want to make something clear to you.
>
> You have repeatedly accused me of having something personal against you,
> both in public and in private.
>
It definitively has looked like that over time.
NMUs are not personal attacks. I know that and appreciate them.
The attitude and comments which went along with them were offensive to
me however.
> I cannot recall ever having even *heard* of your existance prior to
> looking at Bacula, and didn't know a thing about you until later yet.
>
Don't worry, you have won.
Congratulations, mister "Official Maintainer of Bacula in Debian".
Time to step out for a while.
> [snip]
>
> What's more, I WOULD HOPE PEOPLE WOULD DO IT FOR ME. I have maintained
> my fair share of packages over the years, and I have also orphaned or
> given up packages for adoption over the years. I've been NMU'd, even
> recently.
>
> I am not mad at that. In fact, I am glad that these things have
> happened. (I wish it hadn't been necessary, but I'm glad that people
> have stepped up to do it when it was.)
>
I did thank you for your NMU, and hurried to request your patches.
That's a fact.
> A key difference is that I recognized when someone else would be better
> able to take care of a package, and amicably arranged with them to do
> so.
>
Well... I said "no"(for the time being) and you didn't respect that.
Instead, you simply proceeded with the hijack.
> [snip]
>
> I don't say this to try to prove that I'm some mighty Debian developer.
> I'm not, and I know it. I say this because I want you to understand
> something that you may not have had a chance to experience yet -- that
> is, that there is more to getting satisfaction from a project than doing
> everything yourself.
>
I do know that. It's the manners that I don't like.
> I think the best compliment for a Debian developer is for a *user* of
> Debian to say, "Wow, *Debian* is a solid OS that Just Works."
>
> NMUs make Debian better.
They definitively do.
> Hijacking of packages, as in this case, can make Debian better too.
Only when done in a proper way (my own POV, anyway)
> Remember, from the Social Contract, that Debian's priorities are its users and Free Software.
>
Indeed. Those have been mine, even though you and some others differ.
I cared about my users when I packaged Bacula for the first time almost
three years ago.
> Attitudes like this will ensure that there is always half a year left
> until Etch is released.
>
I beg to differ, but I can understand your point.
>> Still much more than two months left for the base freeze. A transition
>> takes 10 days at most.
>>
>
> If you get the package right to start with, and uploaded on time. Which
> history indicates is not likely.
>
Well... different conditions lead to different results.
I have had a Bacula-1.38.9-1 of my own ready for a couple days already,
but it hasn't been uploaded.
Before you say anything: the one who was going to make the upload seems
to prefer waiting until this thread ended instead of uploading the
fixes. And not, it wasn't because of technical concerns, and it wasn't
rover. Now it's too late. Thanks for nothing.
>>> * The last upload for Bacula was almost a year ago.
>>>
>>>
>> There were no upstream releases for over six months, either.
>>
>
> But there were RC bugs against it in Debian for over a year. You needed
> to make a new upload.
>
As has been recently recognized publicly, it wasn't over a year old; And
it had only been promoted to RC recently.
> Then your testing standards are insufficient for an upload to sid. You
> have uploaded packages that would not even *install from scratch* on
> most machines.
>
Well... I fixed the bugs I found before every upload. Definitively the
subset of machines you and I are talking about are different.
There's nothing wrong with that, but then...
>>> [snip]
>> The fact that you uploaded six versions of Bacula to NEW within one day
>> gives an idea of the level of testing you give them.
>>
>
> That is not correct. Here are the dates of the "NEW" acknowledgments
> from the ftp-master scripts for my uploads.
>
> Date: Wed, 03 May 2006 14:02:23 -0700
> Date: Mon, 08 May 2006 14:02:15 -0700
> Date: Tue, 09 May 2006 07:17:11 -0700
> Date: Wed, 10 May 2006 11:17:05 -0700
> Date: Wed, 10 May 2006 20:47:09 -0700
> Date: Thu, 11 May 2006 15:17:12 -0700
>
> BTW, you *KNEW* that your "six versions of bacula within one day" was
> inaccurate because the Debian Installer scripts CC'd you on every one of
> those messages whose dates I have just listed.
>
I failed to check the dates for those. Sorry.
Anyway, my point referred to the latest ones anyway.
> It only takes a few minutes to do a full build, and far less to do
> a partial build after making minor tweaks, and my machines aren't as
> powerful as that. But I did test my packages.
>
Well... so did I. It doesn't matter anymore, though.
> [snip]
> It is less justifiable to withhold fixes for an RC bug because I know
> the package is not perfect.
>
I don't have that many opportunities, unfortunately.
If I had, my development model would be completely different (or so I
expect). Neither of us can prove either possibility.
>> My users don't seem to agree with you (w.r.t. to "working" Bacula
>> packages)... I have *many* successful reports from my users, by the way.
>> Wanna check?
>>
>
> No. If they have been lucky enough to avoid all your bugginess, fine.
> I am glad that they have, and hope that their data really is as safe as
> they think it is.
>
Well... I use them myself in production, too. I can say that they do work.
> Users of Debian are not so fortunate, as your packages as they were in
> sid would not build, install, or remove correctly and stood a quite
> decent chance of taking down production services.
>
> And it is the users of Debian that are at issue here.
>
> It was trivial to see that the RC bugs against Bacula were valid. You
> even admitted as such to some of them.
>
Why would I not?
>
> The URL I sent you, referring to the Developer's Reference, documents
> the procedure, at least part of which you could have done (mailing
> -private). Due to your claimed nearly 2 years of being away from home
> and thus unable to maintain your package, you should have orphaned it 20
> months ago.
>
I was supposed to be able to maintain the package. Life was supposed to
be bearable for that time. Hardware was supposed to be in working state
when I needed it.
One too many suppositions, I guess.
Very few people did show any interest on the package. The few offers to
adopt it which I did receive before your hijacking (thank you to those
who did) happened only very recently.
Seeing the end that close, I wanted to fix my bugs; As you say, it was
my responsibility.
> No, I don't want your code running on my machine unless I've looked at
> it first.
>
Well... you know what packages to avoid, then. Easy.
>> You have yet to prove that your *sudden* interest in becoming Bacula's
>>
>
> I'm sure I can never prove that to your satisfaction, but it's not
> relevant anyway. I've told you it's not true, and it doesn't make any
> sense anyway. But it's not even relevant.
>
Not anymore, I have to concur. You got what you wanted. Congratulations.
> [snip]
>
>
> Then please, let it rest, and contribute to the Alioth project next week
> when it's ready.
>
I most probably will, since it was me who registered it in the first place.
You didn't even list me as an uploader in your latest upload (the one
with the hijacking).
That doesn't sound like wanting to maintain it collaboratively; I proved
I was willing to do that.
Go ahead and enjoy your victory and new package. Congratulations.
Thread ended. Nothing more to see here.
Bye,
J.L.
Reply to: