7:16 a.m.
I see very fair points in Alain's message. I found my way to become a
packager a month ago or so, and I perfectly recognise some resemblance
to the alluded chicken-egg issue in some procedures. It is true that I
managed to walk down this path in just a couple of weeks, but this is
because I had free time and I invested almost all that free time in
this endeavour.
On the one hand, if we want Fedora to continue to be reliable, then
the process has to be rigorous and demanding. And I have to say that I
found the extensive documentation extremely detailed (in a good sense,
because the process is explained step by step) and clearly written,
particularly the "Join the package collection maintainers" guidelines
and the packaging guidelines. But on the other hand, Fedora wants to
be "at the edge", so obviously there is some trade-off here.
I understand the need for the sponsorship model: someone has to check
that newbies like us don't break things. But at the same time, this is
clearly a bottleneck if nobody knows you. I take this opportunity to
describe the most difficult thing I found in the whole process: how to
prove yourself. In my case, I was the upstream maintainer of my first
package, which I understand that facilitates things. But still,
packaging is quite different from developing.
"Do informal reviews", the guide says, and everybody says. Ok, but
when you are new to this, you have to spend *a lot* of time in a
review, because all the time you have to check this or that guide
again, and read other related SPECs, and read other related reviews.
And then I found Robert-André Mauchin, among others, who does an
*impressive* reviewing work for Fedora, but man, you are simply too
fast for newbies! :) Even with all that free time in those weeks,
every time I saw a new review request, I was in a race against time to
provide some meaningful comments. Other options were to provide some
irrelevant comments to likewise irrelevant and forgotten old review
requests... but that is probably pointless.
That said, the process as a whole is quite good and well-explained.
But we can't afford losing contributors like Alain, so something has
to improve. I don't know what, I'm just a newbie, but something. And
for those of you who reached this point, sorry for this moan.
Iñaki
On Thu, 4 Oct 2018 at 13:13, Brian (bex) Exelbierd <bexelbie(a)redhat.com> wrote:
+packaging list
Thank you for taking the time to describe your challenges. I am
adding the packaging list as they are in a position to help with some
of these issues and to consider where more docs may be useful.
I am not sure how to help on the package reviews side. I personally
only have one package in Fedora and my role in the project made it
easy to find a reviewer. But, you shouldn't have to be the FCAIC to
make getting a reviewer easy. I know that we are working on more
automation in these areas to make review less human-labor intensive.
However, that doesn't help you today.
I hope we will hear some news from FPC about this.
regards,
bex
On Thu, Sep 27, 2018 at 10:40 PM Alain Vigne <alain.vigne.14(a)gmail.com> wrote:
>
> Hi, I am Alain, FAS: avigne
>
> As discussed today in #fedora-devel, here is some feedback about my experience
trying to join Fedora as a contributor -> packager.
>
> TLDR: Adding a new package and become a Fedora packager is NOT easy.
>
> I am not a computer scientist, but as an Integrated Circuit designer, I am using
eCAD proprietary tools heavily.
> With my years of experience, I came up to know how to use Fedora, and I like this
distro because it is reliable, and fairly up to date with software technologies.
>
> When Fedora FEL spin was alive, I picked up some tools, and slowly learn how to use
them. gEDA, PCB, NGspice, GerbV, etc...
> At some point, 2 years ago, I thought Open Source world gave me a lot, it was time
to give back... So I contacted the pcb-rnd project [1], and started to contribute code
around GTK, and GUI aspects of the application.
>
> Naturally, as I am developing with a Fedora system, I thought it could be nice to
have pcb-rnd for Fedora... I had no clue on how to proceed, and first, I tried to find
someone who can do that for me...
> Found no one. (I should have known :).
>
> Time passes by, and one day a pcb-rnd Mageia contributor showed me the .spec file he
wrote for Mageia. I was curious about what was behind all those commands and how this
"recipe" (the .spec file) can lead to a package.
> So I dug into the documentation (mainly Fedora wiki) learning how to first build an
RPM, then after a successful local "mock", my curiosity was satisfied. I thought
I understood the purpose of those tools (rpmbuild, rpmlint, mock).
>
> That is when I started to think about contributing this package to Fedora. "It
should be easy, I have the recipe, I just need to find where to check-in the .spec
file..." Easy thought, no ?
> Unfortunately, no, this is not easy.
>
> First, there are tons of pages describing the process, and what to do. In theory the
process is well described.
> In practice, I got stuck in the "need a sponsor" phase where I think there
is kinda chicken-and-egg problem for a new contributor.
>
> I might detail that, later, if someone is interested in this list.
>
> My feeling today, 6 months after I jumped in the unknown is not very much positive:
> I had to register, open accounts, leave traces on many systems before being able to
.... get nothing at the moment
>
> Bugzilla
> FAS
> COPR
> mailing list
> Freenode registration
>
> etc...
> I feel like someone who has a complicated map under the eyes, walk, try and error to
make sense of the map, up to a point where the map says: next step is "find a
sponsor" and I have no idea how this is being done.
> And time passes by... Slowly. I am silently ignored.
>
> Somebody says today : "Do informal reviews [a suggestion on the wiki, but what
can I suggest ? I have no experience ->chicken and egg problem], make some mails, fill
some bugs and you will get noticed".
> I think this is the problem: nobody noticed, it seems nobody cares having a new
volunteer.
>
> So, I am concluding: Fedora = too big ship, mainly automated, with a lot of
processes (procedures, way of working) and a community not open to new contributors [I
recall, my experience is only about contributing a new package], because this is too
complicated (which I agree and understand).
>
> That said, I am a patient man, and I have done all this travel not to being stop by
a wall. I spent my life trying to get around, over, across... so many walls, so, I
won't surrender here !
> Thanks for reading till that point, and let us open the debate.
>
> Kind regards
> Alain
> PS: I am French, not EN native speaker, pardon my language if it does not make sense
to you.
>
> [1] http://repo.hu/projects/pcb-rnd/
>
>
> --
> Alain V.
> _______________________________________________
> fedora-join mailing list -- fedora-join(a)lists.fedoraproject.org
> To unsubscribe send an email to fedora-join-leave(a)lists.fedoraproject.org
> Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives:
https://lists.fedoraproject.org/archives/list/fedora-join@lists.fedorapro...
--
Brian (bex) Exelbierd | bexelbie(a)redhat.com | bex(a)pobox.com
Fedora Community Action & Impact Coordinator
@bexelbie | http://www.winglemeyer.org
_______________________________________________
packaging mailing list -- packaging(a)lists.fedoraproject.org
To unsubscribe send an email to packaging-leave(a)lists.fedoraproject.org
Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproje...
--
Iñaki Ucar