[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#603974: live-installer should depend and use console-setup-udeb



In data venerdì 19 novembre 2010 00:47:56, Samuel Thibault ha scritto:

> Marco Amadori, le Fri 19 Nov 2010 00:29:59 +0100, a écrit :
> > As it is now, live-installer fails in providing a proper keyboard
> > configuration in the installed system, in fact it fails in propagating
> > kbd-chooser debconf values.
> > 
> > Something changed in how console-setup-udeb is handled in d-i, and for
> > squeeze i found that depending on it and invoking the postinst would
> > succeed.
 
> I don't understand this. How was console-setup-udeb involved in
> live-installer before?

I didn't checked how it was handled before, maybe something changed in base-
installer, of which live-installer exploits some helpers.

> As I understand your sentences, live-installer
> seems to be using the kbd-chooser udeb, and installing console-setup in
> the installer system.

It does not do that directly, my guessing is that it was automatically cared 
of before.

> If so, then the bug is rather in console-setup,
> not in console-setup-udeb.

I'm pretty sure it isn't the case, I checked what console-setup does in the 
target and it is doing it right, but its debconf database it is empty (I mean 
default values).

>From what I understood kbd-chooser ask keyboard layout question and translate 
its results to console-setup debconf values, then console-setup-udeb 
propagates them to the target system.

auto-install udeb do something similar as my patch does. It depends on 
console-setup-udeb and explicitly calls its postinst.

> Depending on console-setup-udeb is just a
> convoluted way to fix things, as I guess what its postinst does to fix
> this in your case is just doing what console-setup is supposed to be
> doing in the installed system, already...

This is probably true. My patch seems hackish even to me although it works, I 
hope that someone with a deeper understanding of debian-installer internals 
could provide a cleaner fix if possible.

Anyway this fix is better than having an user unable to have its keyboard 
supported at all.

-- 
ESC:wq

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.




Reply to: