[CentOS-devel] https://blog.centos.org/2020/12/future-is-centos-stream/

Fri Dec 11 16:30:01 UTC 2020
Leon Fauster <leonfauster at googlemail.com>

Am 11.12.20 um 16:55 schrieb Michael Scherer:
> Le jeudi 10 décembre 2020 à 12:39 +0000, Dan Seguin a écrit :
>> Who exactly do you expect will contribute to RHEL from the CENTOS
>> community?
>>
>> Why would *anybody* spend their own time, even their IT reputations
>> rolling any of this mess?
> 
> I would, even if I wasn't paid by RH.
> 
> My job involve being a sysadmin, and so I run Centos servers in
> production, along some Fedora, some Debian and others, for myself and
> for work.
> 
> Each time I want to get a small new feature or fixes, not being able to
> easily contribute to RHEL (and so Centos) despites working at RH is a
> problem.
> 
> I have a few examples:
> a few years ago, I bought a Yubikey for myself. I want to use it on my
> RHEL 7 (my laptop) as a smartcard, but it was not supported.
> 
> I searched, found the fix is 2 lines in the usb-ids database. I opened
> a bug in may 2016, I provided the patch, went to complain on internal
> IRC to the right people, who escalated that in meeting, and it still
> took ~6 months:
> https://bugzilla.redhat.com/show_bug.cgi?id=1157226
> 
> I knew what to do, because I had the same problem in 2014:
> https://bugzilla.redhat.com/show_bug.cgi?id=1157226 where it took 1
> year to get the fix in RHEL 7 (and so Centos 7).
> 
> On Fedora (or others faster distros) where there is a documented way to
> contribute, I would have waited a few weeks at best, maybe more in
> worst case (depending how persistant I am into getting a fix). With
> streams, I hope that at least, I wouldn't need to wait on "next minor
> release", given the delay it had.
> 
> And keep in mind that while I did that using my profesionnal
> relationship, I did it for myself, and others Centos users did benefit
> from it.
> 
> 
> Another example, SELinux. When I see a unconfined service (last one,
> knot, for my personal DNS server, but also synapse or gitea, personal
> matrix and personal forge), I try to write a policy, and while on it,
> get it usable for others. Usually, I have not much trouble to get
> things in the Fedora package.
> 
> I have no hope to get it backported to RHEL (and so Centos), but at
> least, with Stream, I would have a path to try to get it done.
> 
> Again, I would write the policy for me anyway, so why shouldn't I try
> to share it with others ?
>   
> 
> And I do not think I am the only one with a ethos of sharing bugfixes
> with the rest of the community. The whole free software movement is
> built on that.
> 


I see it the other way arround. I am not sure if this component
would get rebased

EL8: https://bugzilla.redhat.com/show_bug.cgi?id=1835210

or fixed

EL8: https://bugzilla.redhat.com/show_bug.cgi?id=1836024

if I (the community) had not reported it (at least not so early 
relatively to the point release).

Especially for early point releases (you known that they have baby 
diseases) such community contributions contribute to a faster stable 
RHEL release.

More examples

https://bugzilla.redhat.com/show_bug.cgi?id=651780

https://bugzilla.redhat.com/show_bug.cgi?id=1623692

The last one was while RHEL6 was already in Maintenance Support 2 phase.

Sure, the vendor must focus on the functional business requirements but
as we all known the non-functional ones are also very important -> 
having a community.


--
Leon