Difference between revisions of "TODO"
(→June 2011) |
(→June 2011) |
||
| Line 21: | Line 21: | ||
=== June 2011 === | === June 2011 === | ||
| − | * Linux 3.0-rc is out now, please test all kernel related packages and report issues (VirtualBox, Genkernel,....) to make sure nothing will break. | + | * '''Linux 3.0-rc is out now, please test all kernel related packages and report issues''' (VirtualBox, Genkernel,....) to make sure nothing will break. |
* Someone should check this boot-update bug with the scan command : http://forums.funtoo.org/viewtopic.php?pid=2706 (It's been there for a while now.) | * Someone should check this boot-update bug with the scan command : http://forums.funtoo.org/viewtopic.php?pid=2706 (It's been there for a while now.) | ||
* Investigate and fix mount-boot.eclass issue. While trying to unmerge openvz-rhel-* and sysrescuecd-std-sources the contect of /boot is cleared from kernel and initramfs images. | * Investigate and fix mount-boot.eclass issue. While trying to unmerge openvz-rhel-* and sysrescuecd-std-sources the contect of /boot is cleared from kernel and initramfs images. | ||
Revision as of 12:13, 31 May 2011
Contents |
Long term milestones
- One our our priorities is to provide a high-quality documentation, several wikis around are discussing about detailed setup and day-to-day Gentoo/Funtoo maintenance. Aside of what is directly Funtoo related, we *absolutely need* some articles on various Enterprise_class_subjects.
- Like any other distribution we are working a Funtoo specific boot media both on Intel and SPARC64 (netboot images are in the plan)
- Porting Funtoo on stragical non-Intel platforms like ARM (mobile devices which are more and more present in our lives)
- Continuous improvements like:
- Integrates bug-fixes/enhancement/security patches to existing packages
- Bring in Funtoo stalled packages posted in the bugzilla of our Gentoo fellows and not integrated to their main portage tree yet (rev bumps, new additions....)
- Quality Insurance on exiting things
- Document, document, document: Our documentation reflects our distribution, the richer/reliable it is the better Funtoo appears (we target synthetic approaches that facilitates the reader's global understanding of concepts behind and why he do something that way and what the impacts are.
Middle term milestones
- Funtoo 1.0 profiles is advancing well on x86/amd64 but Sparc64 lags a bit behind (mainly because of the very limited pool of resources working actually on the subject), however Sparc 1.0 should be due to summer 2011.
Past and present short term milestones
June 2011
- Linux 3.0-rc is out now, please test all kernel related packages and report issues (VirtualBox, Genkernel,....) to make sure nothing will break.
- Someone should check this boot-update bug with the scan command : http://forums.funtoo.org/viewtopic.php?pid=2706 (It's been there for a while now.)
- Investigate and fix mount-boot.eclass issue. While trying to unmerge openvz-rhel-* and sysrescuecd-std-sources the contect of /boot is cleared from kernel and initramfs images.
April 2011
For April 2011, I want the Core Team to start messing with the Funtoo 1.0 profile and getting it ready for use. I have a new alpha-quality profile tree profiles/funtoo/1.0/linux-gnu/ that needs some love. The older Funtoo 1.0 profile is still there but the tree in profiles/funtoo/1.0/linux-gnu/ is what I want to work on and refine and get production-ready.
Overhauling ebuilds - Core Team members who have completed their initial training of moving one key ebuild from masked to testing to current to stable can now proceed with the next step of the Core Team training - overhauling an ebuild. The way this works is that you find an ebuild that is a key part of Funtoo, but not too dangerous to improve (not gcc, glibc, etc.) and you look at all open Gentoo bugs for the ebuild. You try to fix as many as you can. Then you also look at ways to make the ebuild more elegant and simpler, so it is easier to maintain. You also look at Debian and maybe even CentOS packages for potential fixes that Gentoo is missing, and consider adding them to Funtoo. The idea here is to get an ebuild into top shape with all open bugs fixed, and to really spend some quality time making a particular ebuild a shining example of Funtoo. There is also a competitive aspect - we want to offer the best ebuilds we can. Details are important.
To start this overhauling process, find an ebuild you would like to work on and chat with Daniel about it.
Updating Ebuilds
For information on how to update ebuilds, see Updating Ebuilds.
The Fundamentals
Distros get better by focusing on the fundamentals. We'll be focusing on the install experience for a while, until it is great. The tasks below reflect this. If you have other ideas on how to improve the installation process, we're certainly open to them.
Test Installs
I need people to test a Funtoo install from stage3, and note any problems. We do this for several reasons - first, to identify bugs that we should fix; second, to identify weird issues that are not bugs but are areas that should be improved, and third, to look for general improvements to the install process, such as improving default config files, etc. which could result in a simpler install process. If you want, you can document your experiences on the wiki, here: Usability Testing.
Funtoo Quick Install Guide
We need to wrap up the Funtoo Quick Install Guide. I have edited quite a bit to get things simpler. GPT/GUID should be the default partitioning method. The instructions should be easy to understand and uncomplicated.
Genkernel Install
The Quick Install Guide should document how to install a kernel using genkernel. Right now, the funtoo-quebec guide shows manual kernel building which is not "quick". We should offer a genkernel option so that people who want a truly quick install can do it. [404 Error]
Default Kernel Sources/Binaries
We need to identify a default "desktop/general-purpose" kernel source tree as well as a general "server" kernel source tree that we need to offer as defaults to end-users. I encourage all core team members to look into this, and offer their suggestions, and start playing around here. Ideally these kernels should be relatively modern, but also leverage existing well-tested source trees so we get to leverage the testing that other distributions have performed, and benefit from their QA.