Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Fri, 18 May 2007 19:48:54 +0200 maximilian attems <[EMAIL PROTECTED]> wrote: > tags 425050 wontfix > stop > > On Fri, May 18, 2007 at 07:32:45PM +0200, Tim Dijkstra (tdykstra) wrote: > > > > I created a patch to ask a debconf question (medium priority) if > > update-initramfs should update all initramfs or not. The idea is that > > other packages (like my uswsusp package) should check this question too. > > > > This way we can make both the people that want to keep old initramfses > > around and the people that want an up-to-date initramfs for several > > versions happy at the same time. > > > > grts Tim > > big no: > useless debconf proliferation is bad. Come on. `useless debconf proliferation'? The question has medium priority. I can also make it an configration option somewhere and use that, but it was just a convenient why to get info from a user. > second if you want to update all initramfs it is easy to do so. No, it is not. Well not in an automated way consistent over all packages that use it. Maybe you haven't noticed, but this has come up several times in several bug reports already. Now we have a situation where from several packages, some only update the initramfs only for the current one, others for all. This is IMHO bad. I presume (reading other bug reports against u-i) that there was a good reason not to update them for all. Well know you don't get that behavior. > third this does not belong to initramfs-tools at all. Of course it is if we want it to be consistent over several packages. grts Tim signature.asc Description: PGP signature
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Fri, May 18, 2007 at 09:26:25PM +0200, Tim Dijkstra wrote: > Come on. `useless debconf proliferation'? The question has medium > priority. I can also make it an configration option somewhere and use > that, but it was just a convenient why to get info from a user. I'd also say a debconf question is overkill. People who understand what the option means can edit a config file by hand. Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
also sprach Gabor Gombas <[EMAIL PROTECTED]> [2007.05.19.1832 +0200]: > > Come on. `useless debconf proliferation'? The question has medium > > priority. I can also make it an configration option somewhere and use > > that, but it was just a convenient why to get info from a user. > > I'd also say a debconf question is overkill. People who understand what > the option means can edit a config file by hand. They can set their debconf priority. It's not something to avoid, adding debconf questions. Anyway, I propose that update-initramfs just loses the -k switch or provides a new wrapper that packages like cryptsetup and mdadm just call without worrying whether it will update all initrds or just the current one. -- .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems signature.asc Description: Digital signature (GPG/PGP)
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
[Gabor Gombas] > I'd also say a debconf question is overkill. People who understand what > the option means can edit a config file by hand. Remember that using debconf allow the setting to be preseeded during installation. Editing a config file by hand is less convenient for preconfigured installations. Friendly, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 11:08:02AM +0200, martin f krafft wrote: Anyway, I propose that update-initramfs just loses the -k switch or provides a new wrapper that packages like cryptsetup and mdadm just call without worrying whether it will update all initrds or just the current one. I agree...use a debconf or config file option, then let packages call update-initramfs with a new option (let's call it "-p") and the option will be automatically honoured so that one or all initramfs images are rebuilt. -- David Härdeman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On 20-May-07, 04:08 (CDT), martin f krafft <[EMAIL PROTECTED]> wrote: > also sprach Gabor Gombas <[EMAIL PROTECTED]> [2007.05.19.1832 +0200]: > > > Come on. `useless debconf proliferation'? The question has medium > > > priority. I can also make it an configration option somewhere and use > > > that, but it was just a convenient why to get info from a user. > > > > I'd also say a debconf question is overkill. People who understand what > > the option means can edit a config file by hand. > > They can set their debconf priority. It's not something to avoid, > adding debconf questions. Yes, it is. Every question you ask is something more to confuse/distract/annoy a user. Debconf was *supposed* to be only for things for which there is no reasonable default. This is not one of those things. Certainly not at medium priority. Additionally, debconf breaks dpkg conffile handling -- if a setting is in debconf, the file can't be a conffile. Steve -- Steve Greenland The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 01:59:20PM +0200, Petter Reinholdtsen wrote: > Remember that using debconf allow the setting to be preseeded during > installation. Editing a config file by hand is less convenient for > preconfigured installations. But why would you want to install _multiple_ kernels in a preseeded install while _not_ building an initrd for some of them? That's the only case I can think of when a debconf setting would be relevant; if you install just one kernel image or you want to create an initrd for all installed kernels then debconf wouldn't buy you anything over a config file... Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 11:08:02AM +0200, martin f krafft wrote: > They can set their debconf priority. It's not something to avoid, > adding debconf questions. Debconf has a much higher maintenance cost than a config file - you have to provide translations, you must write maintainer scripts etc. Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
[Gabor Gombas] > But why would you want to install _multiple_ kernels in a preseeded > install while _not_ building an initrd for some of them? I'm not talking about installing multiple kernels. I'm talking about setting the wanted configuration for how multiple kernels should be handled on the installed system after the installation. And that setting would be nice to be able to preseed during installation. Friendly, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
Please cc: the bug report On Sun, 20 May 2007 16:35:27 +0200 David Härdeman <[EMAIL PROTECTED]> wrote: > On Sun, May 20, 2007 at 11:08:02AM +0200, martin f krafft wrote: > >They can set their debconf priority. It's not something to avoid, > >adding debconf questions > > > >Anyway, I propose that update-initramfs just loses the -k switch or > >provides a new wrapper that packages like cryptsetup and mdadm just > >call without worrying whether it will update all initrds or just the > >current one. > > I agree...use a debconf or config file option, then let packages call > update-initramfs with a new option (let's call it "-p") and the option > will be automatically honoured so that one or all initramfs images are > rebuilt. That seems sensible to me. Maximilian, what do you think? Is this acceptable to you? grts Tim signature.asc Description: PGP signature
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 04:35:27PM +0200, David Härdeman <[EMAIL PROTECTED]> wrote: > On Sun, May 20, 2007 at 11:08:02AM +0200, martin f krafft wrote: > >Anyway, I propose that update-initramfs just loses the -k switch or > >provides a new wrapper that packages like cryptsetup and mdadm just > >call without worrying whether it will update all initrds or just the > >current one. > > I agree...use a debconf or config file option, then let packages call > update-initramfs with a new option (let's call it "-p") and the option > will be automatically honoured so that one or all initramfs images are > rebuilt. That sounds to me like addressing the wrong problem. Maybe what should be done is that update-initramfs updates *all* the images, whatever happens, but a copy of the kernel and initramfs images used to boot the system as it is at the moment should be kept somewhere and a special grub/lilo/whatever item like "Last successful boot setup" should be added using these. Mike -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
also sprach Mike Hommey <[EMAIL PROTECTED]> [2007.05.20.2102 +0200]: > That sounds to me like addressing the wrong problem. Maybe what should > be done is that update-initramfs updates *all* the images, whatever > happens, but a copy of the kernel and initramfs images used to boot the > system as it is at the moment should be kept somewhere and a special > grub/lilo/whatever item like "Last successful boot setup" should be added > using these. I would second this, but I think the two can be implemented separately as the initrd should be copied during early rcS and not when update-initramfs is called. -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems seen on an advertising for an elaborate swiss men's watch: "almost as complicated as a woman. except it's on time" signature.asc Description: Digital signature (GPG/PGP)
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, 2007-05-20 at 23:06 +0200, martin f krafft wrote: > also sprach Mike Hommey <[EMAIL PROTECTED]> [2007.05.20.2102 +0200]: > > That sounds to me like addressing the wrong problem. Maybe what should > > be done is that update-initramfs updates *all* the images, whatever > > happens, but a copy of the kernel and initramfs images used to boot the > > system as it is at the moment should be kept somewhere and a special > > grub/lilo/whatever item like "Last successful boot setup" should be added > > using these. > > I would second this, but I think the two can be implemented > separately as the initrd should be copied during early rcS and not > when update-initramfs is called. esp. since dpkg triggers aren't implemented yet and update-initramfs might be called a couple times between reboots and thus nuking the backup sean signature.asc Description: This is a digitally signed message part
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 09:02:04PM +0200, Mike Hommey wrote: > That sounds to me like addressing the wrong problem. Maybe what should > be done is that update-initramfs updates *all* the images, whatever > happens, but a copy of the kernel and initramfs images used to boot the > system as it is at the moment should be kept somewhere and a special > grub/lilo/whatever item like "Last successful boot setup" should be added > using these. For this the modules should also be saved, but dpkg has already overwritten them... And even if you save them, you need to patch modprobe etc. to detect that this is a special boot and modules should be loaded from the saved location instead of /lib/modules/. I think this method would be rather fragile... Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#425050: initramfs-tools: Ask if we should update all initramfses
On Sun, May 20, 2007 at 07:47:21PM +0200, Petter Reinholdtsen wrote: > I'm not talking about installing multiple kernels. I'm talking about > setting the wanted configuration for how multiple kernels should be > handled on the installed system after the installation. And that > setting would be nice to be able to preseed during installation. By that argument every config file should be converted to be managed by debconf. As this is not the case, automated installs should already handle replacing config files so this case should not be something special. IMHO settings that have a direct effect _during_ the install make sense to be settable using debconf, but I think this is not the case here. Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences - -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]