Some of you may already be aware that PNAELV have recently released
the first public beta of EL7, and you may thus be wondering if
you can expect a version of RedSleeve based on this in the near
future. I can confirm that is indeed the plan - I plan to spend at
least a week over the upcoming holiday season working on it to get
a basic minimal package set up and running. I cannot promise how
long this will take (it took a few man-months to get EL6 patched
up for ARM and working), and it is likely a similar effort will
be required for RSEL7.

Given the number of man-hours this is going to require, any further
EL6 work is going to get further diluted. Consequently, I am
looking for volunteers from the hundred or so subscribers to this
list to help carry on with the EL6 maintenance.

I will provide volunteers with access to a small Koji farm
consisting of a some armv5tel machines for ongoing EL6 package
builds and maintenance.

What is required is the following:

1) All the packages with a version tag like el6.0.armv5tel will
need to be investigated checked against the equivalent upstream
source, and the extra patch(es) that went into them extracted and
documented on the wiki.

2) Any further new versions of said packages will need to be
similarly patched. Note that I recall there being at least one
package that builds find without being patched, but unless it is
patched other packages that build against it will fail to build;
extra care will need to be taken.

Another thing I will deploy over the holiday season is a bug
tracking system. If you have a specific interest in a particular
package (or package set) please consider volunteering to become
the RedSleeve maintainer for those packages. This will involve
responding to bug tracker tickets filed against the package and
trying to push the patches upstream, to PNAELV, Fedora (release
version and rawhide) if the bug also manifests there, and further
upstream. Be advised that my experience shows that PNAELV's
general response to bugs that don't manifest on x86 is that
they don't care, even if you submit patches that fix the
problem. Similarly, my experience shows that bugs filed against
a particular Fedora release get left unchecked until the bug
zapper closes the tickets 6 months later when the support period
for that version of Fedora expires. So be prepared for a certain
amount of frustration and disappointment, hence why I also
mentioned pushing fixes further upstream, as that will at least
help make sure the same bugs don't drag on forever at distro
level, but have a real hope of being fixed further upstream and
permanently stop being a problem.

All of this will also apply to RSEL7 beta when that is released
in due time.

Unfortunately, Fedora appears to have dropped support for
armv5tel with F19, and EL7 appears to be based on packages of
approximately F19 vintage with some updates. That to some
extent means that there are likely to be unfixed issues
when targeting armv5tel in some of the packages and that
the whole process of getting a working armv5tel build of EL7
may be problematic. Therefore the initial RSEL7 work will be
aimed at getting a hard-float armv7hl build up and running,
and armv5tel will only be considered if time permits.

So if you really care about anything less than armv7hl (*Plug,
Pi, etc.), it is going to take volunteers prepared to put in
a non-trivial number of man-hours to make it happen; more
man-hours than it is likely to take to get an armv7hl build
in place, which is already very much non-trivial. If you are
interested in seeing this happen, now is the time to step up.

Best wishes to you all, and many thanks for using RedSleeve.

Gordan
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users

Reply via email to