Package Review Stats for the week ending January 18th, 2009

[Behdad Esfahbod]

Robert Scheck wrote:
> On Wed, 28 Jan 2009, Behdad Esfahbod wrote:
>> Can we please be done with this kind of blanket statements as well as "Red Hat
>> vs the community" comments?  They are derogatory, demoralizing, and border on
>> offensive.  And after the third time, simply frustrating.  Red Hat has some of
>> the best engineers in the world, perhaps some of the worst if you wish, and
>> most probably a lot of average ones.  So what?
> 
> Such comparisons must happen. And if you would have read my mail correctly
> or even checked what I'm saying, you would have seen, that my examples with
> the Merge Reviews are true. In that cases it was even blocking/refusing the
> Merge Reviews (or parts of it) - the old story from my December 2008 mail.

One of my complaints was against blanket statements.  Sure you may have
positive samples of 1, 10, or even 100 Red Hat employees misbehaving.  That
does NOT extend to all Red Hat employees.  Avoid blanket statements.  Avoid
discriminating.  If someone's violating something, point that out,
specifically, and politely.


> When looking e.g. to ethtool, the maintainer was allowed, not to show up
> ever on the Merge Review, once I asked Tom Callaway to give him a clout. As
> co-maintainer, I've solved the Merge Review, fixed all open bug reports the
> maintainer did not take care about, but I can't co-maintain every package.
> The ethtool maintainer being upstream as well, did after Tom's clout only
> one time reply to my e-mails and to a bug report. So one time in ~ 6 years,
> that's neither a good behaviour for a downstream maintainer, nor even for
> an upstream one. And when talking in #fedora-devel, I got told, that I am
> not alone with this issue.

So what?  There's this small package that I co-maintain.  I'm upstream for
(hint, it's the only upstream package I originally wrote).  My co-maintainer
originally got it in Fedora.  He got past the reviews, he went over the Debian
patches and cherry-picked them for Fedora, and was generally very nice,
helpful, and a very good packager.  Then one day I released a new version of
the package upstream, and it happened to bring machines down under certain
circumstances.  I didn't know.  He didn't know.  I pushed the package in
Rawhide.  Against my will, he pushed the package in F-9 and F-8
updates-testing.  When it became clear that the package eats babies I
desperately asked him to retract the updates.  He stopped responding and as
far as I know he disappeared from Fedora.  I did whatever I could to retract
those updates.  They escalated from updates-testing to updates.  Bug reports
poured in.  I filed bugs against rel-eng.  Etc, etc.

Do I walk around saying "all community guys are irresponsible"?  No.


>> Most of what is seen as "Red Hat guys are bad at packaging" is simply baggage
>> held from the Core days.  Unfortunately (for us "Red Hat guys") most Core
>> packages were simply thrown upon us as part of our job.  I would never
>> *volunteer* to do packaging, as that's not what I'm best at.  If someone wants
>> to go ahead and take maintainership of all my packages, I very well appreciate
>> it.  I'm sure many other "Red Hat guys" are in the same boat.
> 
> Taking co-maintainship or even the package doesn't solve 'em all, as I
> pointed out above already. If the "Red Hat guy" is upstream of software,
> it's more hard to work around.

Why's that?  I don't understand.  Now you're attacking the upstream
maintainers too?


>> proud Red Hat employee, poor package maintainer, RPM n00b.
> 
> Last of it could be easily solved by getting a RPM training, isn't there a
> training department inhouse of your employer?

I'm smart enough to not need training.  I can train myself.  The question is,
should I learn RPM?  I don't think so.  My time is better spent doing upstream
work.  As I've been saying repeatedly, RPM is not my (and many "RH guys"')
strength, and I'm honest about it.  I know how to update my packages to the
new version of the upstream package I just released.  And I hope you don't
have any problem with that.  For anything more complicated, I'd be happy to
let more experienced Fedora packages jump in.  Such collaboration has happened
between me and Nicolas Mailhot already.  He oversees font packaging, and I fix
upstream issues he wants to see fixed.  That's constructive IMO.

Cheers,
behdad

> Greetings,
>   Robert