Re: LILO

[Manoj Srivastava <srivasta@debian.org>]

>>"Itai" == Itai Zukerman <zukerman@math-hat.com> writes:

 Itai> Je 10 May 2001 16:02:36 -0500,
 Itai> Manoj Srivastava <srivasta@debian.org> scribis:
 >> And what would this script do, since it can't really edit the
 >> conffile? And you can choose to tell the kernel image not to put the
 >> symlinks in /; or not to do the symlinks at all; or not to run a boot
 >> loader, or to run an arbitrary script you specify.

 Itai> 1.  What conffile?  For grub, it *maybe* edits menu.lst and does
 Itai>     nothing else.  For lilo it does exactly what kernel-package does
 Itai>     now.

	kernel-package already does that. It ships with a example
 script that allows you to edit menu.lst, and it does the right thing
 for lilo-like systems with the symliunks. Indeed, even grub can be
 pointed to the symlinks, and that works as well. 


 Itai> 2.  Did you choose to ignore my suggestion to give update-boot-loader
 Itai>     some parameters to specify where the new kernel was put?

	That is an implementation detail irrelevant at this stage of
 discussion. 

 >> Given that we have all that already in place, why do we need
 >> yet another standards api that has to be implemented by all
 >> bootloaders of all architectures and hard coded in the postinst?

 Itai> Your way: Put scripts to update all the boot loaders after a kernel
 Itai> install on all architectures into kernel-package.  Make them
 Itai> centralized.

	Rubbish. kernel-package is no longer accepting scripts for
 bootloaders for inclusion; it has implemented a general purpose hook
 for *any* scripts to be run. Since the final decision has to be the
 sys admins, we make it easy (by setting a single variable
 appropriately) to run any script they wish.

 Itai> My way:  Distribute the scripts among the boot-loader packages.

	No, achieve the same functionality we already have by creating
 a whole lot of packages, creating a upload-boot-loader protocol,
 probably having to muck around with update-alternatives.

	A lot of hoopla for no discernible advantage.

 Itai> It's still pretty clear to me.

	It may be clear to you, but it is also wrong.

	manoj
-- 
 Before you ask more questions, think about whether you really want to
 know the answers. Gene Wolfe, "The Claw of the Conciliator"
Manoj Srivastava   <srivasta@debian.org>  <http://www.debian.org/%7Esrivasta/>
1024R/C7261095 print CB D9 F4 12 68 07 E4 05  CC 2D 27 12 1D F5 E8 6E
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C