[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Advocacy for Jari Aalto



Hello, and thank you for the cc:  I'm not currently subscribed to
d-newmaint.

On 05/22/2013 01:16 AM, Raphael Hertzog wrote:
> Hi,
> 
> On Wed, 22 May 2013, Jakub Wilk wrote:
>> * Thomas Goirand <thomas@goirand.fr>, 2013-05-21, 21:52:
>>> I think Jari would be a very valuable asset in Debian, and that he
>>> should be made DD.
>>
>> I very much disagree.
> 
> I'm with Jakub. My interactions with Jari have been rather on the bug
> reports front, but he tends to care a lot about issues which are not
> important at all. He doesn't seem to be able to weigh the usefulness of
> his reports towards the time it costs for others to deal with them.

It appears that I am in agreement with Thomas and disagreement with
Jakub and Raphael, at least to a certain extent.  I'll agree that Jari
has opted to focus on bug fixing and updating packaging - in fact, I
have an email folder with 80-odd sponsor requests for QA/MIA and
maintainer-ACK'd NMUs to complete getting rid of dpatch.  But I would
argue that, (a) these things are useful to the distribution (namely,
regularity and consistency in packaging practices as opposed to allowing
cruft to accumulate), and (b) I don't see other contributors spending so
much of their time on these questions of consistency.

> He made a number of dubious NMUs (just to upgrade the packaging to
> newer debhelper/source formats/etc) and he doesn't seem to be aware of
> the requirements for uploads during the freeze [1].

Reference [1] is about 2 and 1/2 years old, and I don't know what the
half-life is for mistakes, but it doesn't bother me that much.  His NMUs
were to fix bugs (often RC bugs) *and* upgrade packaging, never (to my
knowledge) merely to update packaging.  At least the NMUs of his that I
sponsored. And given that he doesn't have upload rights, they were all
sponsored.  So your beef is more properly with me.  (But let's save the
debate about the scope of NMUs for a separate thread.)

> He started a mass bug filing without any coordination [2].

This seems to happen from time to time anyway, often perpetrated by DDs
with good intentions.

> The amount of packages that he maintains alone, i.e. outside of any team
> is also worrying (in terms of willingess and ability to work with others).
> Quite a few packages are lagging behind new upstreams versions (I count 8
> on his DDPO page) but that's to be expected when one maintains so many
> packages.

Err... I don't buy either of these arguments at all.  He's certainly
been willing to work diligently and patiently with sponsors, with
developers (posting patches and debdiffs for NMUs), and recently he has
been helping others with their packaging (see the recent additions of
firehol and sanewall to the archive).  And if lagging behind upstream is
a problem, about 3/4ths of the Debian Java Team should be ejected.

If you doubt his contribution because he's trying to do too much or is
spread too thin, then I think many Debian contributors are guilty of that.

> I looked into a random package with bugs, I picked "zoo" to see what I could find:
> 
> A recent fix:
> http://anonscm.debian.org/gitweb/?p=collab-maint/zoo.git;a=commitdiff;h=fd1a30d151cc0b74558af41b09ddafbf9978c16d
> was only needed because his debian/rules rewrite (for dh support)
> completely dropped the differentiated target (make linux/linux64) thas was
> present when he took it over:
> http://anonscm.debian.org/gitweb/?p=collab-maint/zoo.git;a=commitdiff;h=f138eb3a46635849af0cb04b0099cc3f722203f8
> 
> The linked bug reports mention a cleaner solution (i.e. to rely on
> a built-in define instead of a manually added one) but he didn't
> update the patch accordingly. He just resurrected the old approach.
> 
> In the end, I'm far from being convinced that he has a sane technical
> judgment to be made a full uploading DD. He probably is able to maintain
> (simple) packages but given that I have serious doubts on his technical
> judgment, I would not like to see him having archive-wide unsupervised
> uploads rights.

I'm not going to contradict your assessment; that's what this discussion
is for.  But I will say that there many DDs of varying skill-levels, and
that social fit and willingness to contribute (even when it means doing
repetitive, thankless tasks like getting rid of dpatch or enabling
hardening rules) are also valuable to the community.

I have worked with Jari for a several years now and don't have any fear
that he will run amok with unsupervised upload rights.  However, given
the concerns expressed, would those who have stated their objections be
more comfortable if Jari was granted DM status?  To me, this seems the
next logical step if his advocacy is disputed.

> [1] https://lists.debian.org/debian-release/2010/12/msg00952.html
> [2] http://lists.debian.org/871undw15g.fsf@deep-thought.43-1.org
> 
> All in all, I think he's already at the right place in the spectrum of
> possible ways to contribute to Debian. He does a lot of polishing, trivial
> bug fixing and has probably helped many maintainers to update their
> packaging or to fix RC bugs while they were MIA.

I don't really agree that he's in the right place because he has spend a
non-trivial amount of time requesting sponsorship (and then sponsors
spend their time sponsoring).  This is time he (and they) could spend
contributing more directly.

DM status wouldn't alleviate that entirely, but would give him the
opportunity to maintain his own packages and participate in team
maintenance with other DDs, who can moderate and adjust his DM ACL as
they deem fit.

Thank you,
tony

Attachment: signature.asc
Description: OpenPGP digital signature


Reply to: