[publican-list] Suggesting /usr/share/publican instead of /usr/share/Publican

[Jeffrey Fearn]

Jeroen van Meeuwen wrote:
> On 10/26/2009 11:30 PM, Jeffrey Fearn wrote:
>> Jeroen van Meeuwen wrote:
>>> No package xslth in Fedora, and no review request either (well, there
>>> is one but it was deferred by Jeff for no apparent reason).
>>
>> https://bugzilla.redhat.com/show_bug.cgi?id=455358
>>
>> Comment #6 From Jeff Fearn (jfearn at redhat.com) 2008-11-17 23:55:06 EDT
>> Packages rebuilt
>>
>> Comment #7 From Jeff Fearn (jfearn at redhat.com) 2009-01-04 23:51:45 EDT
>> Request withdrawn
>>
>> We now know that my tolerance for bureaucracy is less than 6 weeks.
>>
> 
> That, in my opinion, is a funny quote coming from someone @redhat.com.

True, but not at all unusual for me.

>>> What is wrong with this picture?
>>
>> The Fedora packaging process is ridiculously bureaucratic and completely
>> lacking in accountability?
>>
> 
> I'm not sure what this has to do with what I asked, but I'll bite;
> 
> Have you seen the packages that come from Red Hat? Make no mistake, 
> patches in there fail to apply in a simple rpmbuild --rebuild. Not to 
> mention the horror in Merge Review requests. They, *they*, are part of 
> the reason why the Fedora Packaging process is so bureaucratic, mind 
> you.

I've had to rewrite this reply numerous times, so it might be a bit 
disjointed :D

Poor spec files, from numerous, unspecified, sources, have led to the 
creation of the extremely useful tool rpmlint, this is good. The 
bureaucracy is much larger than validating that spec files are sane.

> This thing builds the *best* packages, too, in case you were 
> wondering what the use of all that paper shoveling was.

You just aligned bureaucracy with efficiency and good outcomes.

>>>>> - correct indentation of formalpara's in orderedlists (again mainly
>>>>> PDF)
>>>>>
>>>> ...but not for this, unfortunately :(
>>>>
>>>
>>> I hate repeating myself but why isn't the beta in rawhide again?
>>
>> A: Doing so would have made it impossible for those Red Hat people using
>> rawhide to do their daily work.
>>
> 
> Without it being in a current Fedora release or rawhide I cannot do my 
> daily work, how about that? It was fixed at one point, broken at 
> another, and you say it fixed in the mystical Beta, but nobody knows 
> where it is -until you just said it.

We have detailed how to get the beta repeatedly on this list and have 
instructions on our wiki on how to use it. There should be no issue with 
using the F11 beta on rawhide.

> I bet there's as many people within Red Hat using rawhide in their daily 
> jobs (AND using Publican) as there is people using Publican within my 
> company.

I use rawhide myself, I have both the release and the beta installed. 
The other people here who have a similar setup used the beta 
instructions off the wiki to install the beta. The initial instructions 
had a few errors, they were corrected on feedback from these users.

>> B: Publican is NOT a fedora project and our time lines are not
>> constrained by fedora or it's bureaucracy.
>>
> 
> Actually they are in many many ways but I guess you don't realize that 
> yet, so nevermind.

Believing this to be true will lead you to false conclusions as to what 
our goals and priorities are.

>> C: The Fedora packaging process is ridiculously bureaucratic and
>> completely lacking in accountability?
>>
> 
> Despite the feeling I get you are repeating yourself, all I can say here 
> is you either step up to improve it or you live with it.

Or ignore it and move on to distros that don't waste my time.

 > I'm not on this
> list to argue with you about what you think sucks so great you can't do 
> your job any more 

I am doing my job. Fedora packaging is not part of my job. It's 
something i thought I'd do as a community thing, but it's more onerous 
than it's worth.

 > well then you have -there's mailing lists for that
> (friday-list@, memo-list@).

It would not be appropriate to raise these fedora issues on internal Red 
Hat mailing lists. If I was going to fight that fight, I'd do it in public.

> I'm here to collaborate, and I'll tell you what I normally, but first:
>
> At several points, when I address an issue on this list, you just point 
> me to the fact there's some kind of mystical Beta out there, which isn't 
> available, and doesn't compile from SVN (didn't find a beta-like tag 
> that would build either), I gotta query a dozen resources to get the 
> packages (which I obviously have not), it uses different directories for 
> all kinds of foo (bye bye brands), and now you dare getting in my face 
> about it?

The beta has been extensively discussed on this list and we provided 
instructions on how to use the beta on the wiki.

The decisions to use a different directory and the affect on brands have 
also been explained on this list, it's the only way I could get internal 
people to test it.

>> The package requests have been in fedora's ridiculously bureaucratic and
>> completely lacking accountability package request system for 6 weeks, if
>> you want to vent at someone I suggest you direct it at the people who
>> advocate that system.
>>
>> And yes, if I'd opened those tickets I would have closed them by now and
>> not bothered with 1.0 on fedora. Bureaucracy is like a shell fish
>> allergy, your tolerance is lowered every time you encounter it :/
>>
> 
> Damn dude, your attitude sucks.

It's a simple cost/benefit analysis. The cost of dealing with the 
process is greater than the benefit I get out of it. If I luck out and 
it gets done in a timely manner, good, if not I cut my losses and move on.

> Here's what you should have done right from the beginning:
>
> Jeroen: formalpara's / tables blabla
> Jeff: Yeah, whatever, fixed in Beta but not in Fedora because of:
>   1) xslth
>   2) that other thing
>   3) whatever
>   (..)
>   389) ohw yeah, and this...
 >
> And you know... I would have gone through all the hoops and fixed up the 
> packages or whatever else needed fixing so that they would have been 
> accepted well within the fucking six weeks timelimit.

Why does the process you are defending require special dedicated 
attention to get it to work in a reasonable time frame?

Look at that xslthl bug again, it didn't spend 6 weeks getting changed, 
it spent 6 weeks sitting there after I made the changes. The two 
remaining deps packages spent most all of their 6 weeks sitting there 
with nothing happening.

> And you know what, if you had just told me your boss is putting pressure 
> on you to make sure all kinds of foo and bar departments within Red Hat 
> can use Publican "we-don't-care-about-our-core-competency"-style, I 
> would have *perfectly well* known what you were talking about -and 
> helped you get the shit done, just so that it would have helped *us*.

99% of the pressure on me come from me, it's why I learned to cut my 
losses, fighting the wrong fight costs me too much.

Cheers, Jeff.

-- 
Jeff Fearn <jfearn at redhat.com>
Software Engineer
Engineering Operations
Red Hat, Inc
Freedom ... courage ... Commitment ... ACCOUNTABILITY