With the new year already beginning I would like to sum-up the few
projects I'll try to achieve this year. Hopefully we could look
backward next year and see how it went.
- Cuda repository:
I plan to have dedicated koji target and a separate cuda namespace
(along with free/nonfree separation) This would allow to build cuda
enabled packages using the nvidia provided cuda compiler and software
collection if needed (as external repositories for theses build
target). The feature isn't fully backed yet so I'd like to have others
thought but this namespace would have branches based on cuda versions
(not fedora/el) and have a relaxed policy about replacement packages
(unless it can easily complement).
- libdvdcss: With the new "cuda-drivers repository", it's now possible
to have another "side repository" to handle this package, over using
livna. Last time we discussed this topic, I beleive this solution was
talked about but none volunteered to submit a libdvdcss package review
to start with...
The name of such "side repository" would need to be found, specially
if others packages could fall into such "forbidden item" category.
- Infra maintainers bring-up: I'm sure there are lot of people very
capable as sysadmin. Unfortunately I'm not able to evaluate and see if
we can work together as appropriate. This is a similar situation when
one wonders if he could give provenpackager rigths to a person that
haven't shown much packaging work yet. This is not possible as the
risk to have lot to recover is high.
Fedora wouldn't allow to default to trust either.
But there is also a need to break this circle to better evaluate
sysadmin candidates. One of them is to instantiate a cloud vm where
people could start installing packages and playing with ansible
(similar to the fedora infra apprentice group, but eventually with
some sudo rights).
- Less packages in the review queue:
Here is the list of the pending review (today there are 45 waiting packages):
There is a need to close non-responsive submitter and to bring new
maintainer into action.
- Drop old packages and concentrate on new:
I would prefer to concentrate on packaging Natron than cinelera-cv,
libquicktime should be nuked.
- RPM Fusion containers images:
It should be possible to provide containers for software from higher
fedora than what's currently installed. For example, ffmpeg 3.5
containers using a fedora 28 base. This would dramatically reduce the
need to provide newer softwares on older releases. And would leave
room to more work into a "deep stable" releases process along with
recent software. (as soon as it's a binary only package, not a
(see docker pull rpmfusion/ffmpeg as a first attempt).
- Better Multimedia for EL (scl-like):
Software Collection (scl) is a mean to provide an alternate version of
a given stack using a different path. RPM Fusion could use a dedicated
koji target to rebuild fedora/rpmfusion packages using such
- Flatpak/Module/Etc support.
Fedora tries to experiment in few directions to work towards better
support for alternates version.
I personally prefer native packages over flatpak by several order of
magnitude. But for toying and by popular demand, we might inevitably
"regress" to provide flatpak.
- RPM Fusion spins:
With the current infra, it should be easier to provide spins. Either
live or "ostree" images. Both of them could be provided as x86_64
armhfp and aarch64 (that would be fun to maintain a compatible hw list
for arm side).
* Multimedia Center: something about Koji (would need arm* fixes) and
mythtv or else.
* Arcade Gaming:
* Other: I know others are maintaining Spins from the community.
Thoses would be nice in a
- Drop the repoview from mirrors and use the "packages" applications.
repoview categories are outdated and relies on Group from spec files,
best would be to use the "packages" application and adapt for
rpmfusion. This would be a good candidate for the infra bring-up.
- Rework of the frontpage and wiki:
Right now the frontpage is part of the wiki. But it could be separate
components. There is a need to renew both of them. Either using the
same techno or something else. Suggestion ?
- Delegate more coordination tasks:
I would like to progressively phase down my involved in the project
with me acting as a infra maintainer. I specifically want to focus on
finding hw sponsor so we don't need to rely
(we have DNS domain bought for the next 5 years at least)
I'm pretty sure there are others items in my list, but at least that's
the main ones, of course if you don't know what you do, I have things
Thx for any help and other initiatives in these area.