[Pkg-zfsonlinux-devel] Current status?

Carlos Alberto Lopez Perez clopez at igalia.com
Tue May 7 01:46:25 UTC 2013

On 07/05/13 02:05, Darik Horn wrote:
> On Mon, May 6, 2013 at 4:03 PM, Carlos Alberto Lopez Perez
> <clopez at igalia.com> wrote:
>> We agreed to use your repository as base (which I didn't liked) and to
>> ask you for review each commit before merging it with the idea that you
>> could feel comfortable and you will engage in collaborating with me.
> Do you know how I pounced on your ITP so quickly?
> ZoL is high profile, and I was notified that you copied my work
> without attribution into your initial submission.  I like to think
> that Debian preserved my git history to give me some credit and
> concession for that slight oversight.

I'm asking you right now to stop making false accusations without any proof.

>> The one that was supposed to do code reviews was you, but you just
>> ignored all merge request I sent to you. Aron just limited himself to
>> act as a mediator because of your lack of response.
> Let me disabuse you about those code reviews:
> 1. Your early submissions failed to build or upgrade cleanly.  I
> usually insist that a junior developer submit binaries after doing
> that to me a few times, but you were identified as a recruiting
> prospect and were handled with kid gloves.
> 2. I let you ride for eight weeks hoping that you would notice the
> dkms or configuration bugs.  You failed that test too because you
> weren't using the software.  (ie: Dogfood principle.)
I always tested the package to build and always updated the test machine
with the built packages. If you noticed some error that I didn't, maybe
you could have try to communicate with me instead of ignoring my emails.
Don't think?

Also we didn't released (yet), so any error on the package isn't a big
deal. Only before the upload extensive QA should be done.

> 3. You left global search-and-replace artifacts at 661c4a4b, which
> indicates unsafe coding habits.
$(MAKE) is the recommend way to call sub-makefiles. Go read:

> 4. Your rationale in fbc1bfef is demonstrably false, which means that
> you invent bugs or fix things that aren't broken.  (In both dash and
> bash, the logical && and || operators provide errexit protection.)

I though the commit message was clear. In any case I will tell you again:

$ false && true
$ echo $?

(1 means false if you didn't know)

> 5. You added an unnecessary awk at e6fd8d0f and described it yourself
> as "hellish".  I agree, that is low quality.  Never use a regex where
> easily understandable code suffices.
Well, this is a question of tastes.

My solution involves one big awk line on debian/rules

Your solution is 5 patches of >1500 lines in total.

I think many people would agree with me that my idea looks cleaner and
easier to maintain.

> 6. The hostid issues pursuant to 2cc7293f would be termination
> offenses in most software shops. You do not break customer systems.
> Full stop; no discussion.

The way you handle /etc/hostid is broken in many ways:

* You can't ship that file on the package configuration files (this is
already by policy)
* If you create it on the postinst of some package, the earlier the
better. So the right place to do this is in spl-dkms *before* compiling
the spl module for the first time.

> Aron was acting as middle-management here because you were scoring
> nearly zero in code review, and I didn't want to embarrass you.

This is actually funny :)

>> Finally you just told us that you were not longer interested in working
>> in this project inside Debian.
> More accurately, I am no longer interested in covering for your
> mistakes.  I just couldn't think of a simultaneously polite and honest
> way to provide additional feedback or exit gracefully.

And I'm not longer interested in hearing about your fantasies of

>> I was offered to co-mentor the project because of my experience with the
>> ZoL package inside Debian. If you hadn't tell us that you were not
>> longer interested in packaging ZoL inside Debian, then you would have
>> had the same chance to be a co-mentor of the project.
> Your promotion from trainee to GSoC mentor was certainly rapid.

I'm remembering you that I'm the owner of the ITP on Debian, therefore
I'm the de-facto maintainer of the package.

And I have contributed to Debian more than you did in your whole life.

>> And this is not a "product". This is Debian. If you don't know how open
>> source communities work neither how Debian works, please go to learn it,
>> and come back when you have learned the lesson. In the meanwhile I would
>> thank you if you can stop spreading FUD.
> Okay, sure, I need to learn how this works.  In your capacity as
> project mentor and professional software developer, tell me more about
> how Debian is not a product.

Darik.. to be honest, I'm just freaking out with all of this situation.
What's wrong with you? Why do you come here to make false accusations?
What are you trying to do?

I'm not interested in this childish discussion. I have better things to
do with my time than trying to argue with you who is "better developer".

And contributing to Debian is not only about dealing with code, but also
with people.

If you tried to communicate with me about the above points when I sent
you the merge requests, we would have get to understand each other. The
whole idea of sending you merge request was precisely that (discussing
the code _before_ merging it)

Coming here after months have passed of repeatedly ignoring my merge
request and resigning to this project, to say the above says nothing
good about yourself.

I suggest you to pick your inflated ego and go find another site to
claim your superiority.

I've had enough of this nonsense.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 900 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-zfsonlinux-devel/attachments/20130507/0c8e95b6/attachment.pgp>

More information about the Pkg-zfsonlinux-devel mailing list