Re: Usability bugs which aren't bugs?

[Jordan Uggla]

On Wed, Jan 9, 2013 at 11:36 AM, Jago Pearce <address@hidden> wrote:
> Are these bugs?
> I notice it's common in technical projects such as grub that
> useability is often the non-bug, since you can argue that it's not a
> bug. Is this the case here?
> Let's be honest, grub2 is a much bigger learning curve from the days Lilo.
>
> I thought I'd mail here first to discuss and check before jumping in
> with bugzilla.
>
> 1) In CLI mode you should be able to pass "kernel" to boot rather than
> just "linux". The justification I make for this is that this is the
> syntax the grub.cfg file follows so this is what a new user expects it
> to be

That is not the syntax that the grub.cfg follows. Exactly the same
interpreter is used for the interactive terminal as for the grub.cfg,
and "linux" is the command used to load a linux kernel. The "kernel"
command is from grub legacy.

>
> 2) grub2-install gives no output. In my case somehow it had written a
> blank menu to my bootsector. I ended up having to rescue with both LVM
> and an encrypted volume which was no walk in the park. Some verbosity
> here is needed.

I don't understand what you mean by "written a blank menu to my
bootsector". The file that defines grub's menu is the grub.cfg, and
that is neither stored in the boot sector nor does grub-install ever
touch the grub.cfg. grub-mkconfig is the tool for automatically
generating a grub.cfg, or you can write your own manually.

>
> 3) I noticed in the grub2-mkrescue that the syntax is described as
> "grub2-mkrescue SOURCE". Surely the syntax should be "grub2-mkrescue
> DESTINATION"? It's very confusing to think you are reading a file in
> rather than writing a file out. Or am I mistaken?

You are mistaken, the argument is the source, the directory from which
files will be copied into the generated iso. --output is the argument
that takes the output file name as a parameter.

-- 
Jordan Uggla (Jordan_U on irc.freenode.net)