maximilian attems wrote: > hello michael, > > On Fri, May 18, 2007 at 08:25:41PM +0200, Michael Biebl wrote: >> Hi Maks, >> >> I guess you followed the discussion on the d-d mailing list about this >> issue. Imho it actually makes sense, to have a consistent >> update-initramfs-behaviour between all packages modifying the initramfs. >> Tim's proposal is not that bad either imho: > > d-d is the wrong ml, this has already been x times discussed on > d-kernel. No, it's not imho. This discussion is about having a consistent behaviour across several packages, among them are uswsusp, udev and mdadm, usplash and a lot more. This is not d-kernel specific, and I doubt the maintainers of these packages read the d-kernel m-l. >>> useless debconf proliferation is bad. >> This debconf message would only be within update-initramfs. Other >> packages using initramfs-tools would only *read* the debconf setting. So >> talking about debconf proliferation is a mild exaggeration. >> You could also achieve the same via a setting in a configuration file >> but using debconf is more userfriendly. > > no debconf is _not_ userfriendly. > and the config makes _zero_ sense. Without giving more reason why you think this is so, this sentence is absolutely meaningless. Try to give arguments, not statements. > update-initramfs with -a it has a well defined meaning. I guess you mean -k all here. Still I'm not sure what you want to say with that. > >>> second if you want to update all initramfs it is easy to do so. >> Sure it is. But the point is to have a consistent behaviour between all >> packages modifying the initramfs. >> Having one package use "-k all" (as madm already does) defeats the idea >> of having a backup initramfs. > > -k all is risky due to several reasons. I agree it is. But if you followed the discussion, some packages already use -k all for various reasons (e.g. uswsusp needs to have / and initramfs in sync). So, while some packages might use a simple -u, if there is only one package that uses -u -k all, it will dominate this and regenerate all initramfs files. > the backup initramfs is a different issue. > > >>> third this does not belong to initramfs-tools at all. >> Sure it does. initramfs-tools should be the package which defines the >> default behaviour of update-initramfs. Other packages should *not* >> modify this setting but only read it. > > no. > initramfs-tools provides a framework it has no business in enforcing > a dubious policy. Well, this is the point. We don't have a consistent policy. Every package does it's own, which for the given reasons is not a good solution. If you have a better solution than the given one, please tell us. That the topic has come up several times, as you said yourself, should be and indicator that the current behaviour of initramfs-tools is not ideal. If you don't like the implementation via debconf, as Tim suggested, this is something that can be changed and discussed. But the idea in general is the right one, imho. Cheers, Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
Attachment:
signature.asc
Description: OpenPGP digital signature