Am 07.09.2014 um 19:49 schrieb Richard W.M. Jones: > On Sat, Sep 06, 2014 at 06:54:03PM -0400, Rahul Sundaram wrote: >> On Sat, Sep 6, 2014 at 5:36 PM, Richard W.M. Jones wrote: >> >>> We need to decide if just because you manage to get an important core >>> package into Fedora 4 years ago, that means you can forever more push >>> any old stuff you want into Fedora, without going back and consulting >>> with the community and FESCo. >>> >> >> I am puzzled. Upstream doesn't need to consult FESCo for developing new >> features. However it does need to consult FESCo for Fedora integration and >> it seems that systemd has. Can you point out any counter examples? > > There's been a lot of change between systemd-26 (Fedora 15 GA) and > system-216 in Rawhide. I'm just looking at the Fedora packages here, > not the upstream features, because as you say, upstream can develop > whatever they want and good luck to them. > > Anyway, systemd now does the following which it didn't do in F15: the first problem is that you only post a endless list without mentioning what is our problem with it and how if at all it affects something > - has its own network configuration system which does nothing as long it is not used even a own binary > - has a way to control firmware boot settings > - intercepts coredumps and introduces what problem? > - has the journal and introduces what problem? * it does not do any harm * you just can use rsyslog as all the time before * benefit: you get even stdout to rsyslog > - has tools for setting the system time and timezone, and locale > - has a firstboot mechanism and introduces what problem? > - detects virtualization (long story here, but a very bad idea to > encourage programs to do this) no it is *not* a bad idea to start services conditional the best example are VMware tools that way you can build appliances with the guest-tools enabled that images will run on physical and virtual machines unchanged on bare-metal the tools systemd-unit will skip start on http://comments.gmane.org/gmane.linux.redhat.fedora.devel/199113 *you* complain about systemd-readahead - guess what - if a virtual machine is detected it is skipped > - has the sysusers password mechanism > - has a tool for maintaining kernels in /boot > - has a login manager > - has a DBus inspection mechanism (useful, but does it need to be > in the init system?) > - has a 'top'-like program for control groups > - has a program for comparing /etc configurations > - has its own version of the FHS and a tool for managing it > > I question how many of these should be core concerns of an init system i question if you did recognize that systemd is more than a init-system one goals is to standarize things between distributions if it is done properly and careful it leads in less wasted time for upstream developers, pakcage maintainers and adopters
Attachment:
signature.asc
Description: OpenPGP digital signature
-- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
