9.1-RC2 - could it be that the installer does not write the MBR?

[Chuck Burns]

On 10/19/2012 1:32 PM, Adrian Chadd wrote:
> Guys/girls/etc,
>
> I do suggest that someone actually spends some time coming up with a
> table of "what the current state is", "what we could do", "what would
> happen if we did that."
>
> Right now there's a lot of possibilities (new drive, drive with
> windows, drive with linux, drive with linux/windows, drive with legacy
> freebsd MBR, etc) and as an outsider trying to figure out what is
> actually the "right" sounding behaviour, it's difficult for me to sit
> down and digest all these emails that chip away at a bit of the
> problem at a time.
>
> So if you'd like to see this fixed, I really do suggest that one of
> you dumps some time into coming up with a basic table like I said
> above, work with others who can correct/flesh out the various options
> to take into account, and then we can come up with a real solution.
> Then 9.1 can go out the door.
>
>
>
> Adrian
>

In my opinion, I think that the installer should be able to determine 
whether or not there is an existing partition table, either MBR or GPT, 
and at that point, ask the user about replacing MBR.  But if there is no 
existing partition map on the drive, then there is no need for prompting 
about MBR over-writing.

Basically: detect if map exists, if not, proceed as usual, if 
part-table/map is detected, prompt user "Existing partitions detected, 
Should we replace any existing MBR boot code? Only advanced users who 
know what they're doing should say "No" here, as this could leave you 
with an unbootable system: Yes/No (Recommended: Yes)"

-- 
Chuck Burns <break19 at gmail.com>