Previous hackweeks spent on research (project/220), other alternatives. This time I'm going to try to package and polish the results of that and further research that happened in the mean time.

Goals:

  • long-term maintainability and compatibility with older kernels
  • keep compatibility at the commandline tool and function level where applicable
  • no feature rush
  • remove features that do not belong there or are sufficiently independent, possibly package separately
  • split systemd to parts that have eg. different update cycle, are beyond of percieved scope of what such system should do
  • change new defaults back (eg. network interface names)
  • deployable on commong distros, though I'm focusing on openSUSE 13.2 or Tumbleweed during testing
  • (more)

Why:

I don't like the direction in which the upstream systemd is going, but I'm not opposed to ideas implemented there. Goals of this project are gathered from my own experience with systemd-based systems, talks with other people or complaints by other internet citizens.

Unfortunatelly the things I'd like to change are not upstreamable, as they're directly against the 'innovation' features or have been dismissed already.

Where:

The git repositories are going to be published under here , preliminary versions are in my internal git.

Development plans:

Hopefully this hackweek I'll be able to get most of the stuff into shape so I will be able to continue in the future adding smaller parts due to time and other work reasons. Also to make it easier for potential external contributors.


Comments

  • dsterba
    over 3 years ago by dsterba | Reply

    Status: partial split to git repos done, still in one rpm package, update and boot is fine

  • dsterba
    over 2 years ago by dsterba | Reply

    Plan pool for HW15:

    • split the code to more git repos, library, shared code, so it would be possible to build standalone tools separately and possibly add the ones that were thrown away
    • write more docs, goals, mission statement, status, roadmap
    • maybe rebase on top of 210 from SLES, due to bugfixes and long-term support, OR make 210 backports to current base possible without much effort; 210 would be trimmed the same way as 208, the end result should be the same either way
    • research possibility of plain text journal backend

  • dsterba
    over 2 years ago by dsterba | Reply

    HW15 summary: * hwdb and units separated from the main package, lots of packaging, no coding or backports * docs writeups * tested on Tumbleweed with KDE5, it works; yeah systemd 208 + a few build fixes works on a recent distro * github project docs updated

Similar Projects

This project is one of its kind!