On 29/01/14 19:21, Richard Elkins wrote:
Elfy,
I am not trying to be picky but it is not just an issue of
"settings". There are issues about mixtures of MBR boot loaders, VBR
boot loaders, boot managers, their ages, and what order partition boot
chaining is performed. To whatever degree that folks have time and
energy and resources, there are subtleties that are worth verifying.
Of course, I respect that we all need to prioritize our time.
A useful if simplified reference:
http://www.ibm.com/developerworks/library/l-linuxboot/
More detailed description of the MBR:
http://en.wikipedia.org/wiki/Master_boot_record
and the partition-resident Volume Boot Record (VBR) i.e. partition
boot record: http://en.wikipedia.org/wiki/Volume_boot_record
For example, imagine a circa 2007 PC running Windows XP that had 12.04
added in 2012 as a dual-boot machine. Both OSes are stable. The MBR
is probably a 2005 bootloader created by Windows XP. The boot manager
is probably the Windows XP NTLDR. The user, based on the Windows
C:\boot.ini file, selects either Windows or 12.04 to boot. If Windows
is selected, then NTLDR initiates the loading of the rest of Windows
from the current partition. If 12.04 is selected, the 12.04 Linux
kernel in the destination partition is booted.
Depending on how 12.04 was installed, it is also possible that the
boot manager invoked by the MBR is the 12.04 Grub. In this case, the
user selects either Windows or 12.04 from the Grub menu. Grub boots
up Windows in the destination partition or 12.04 in the current partition.
So, in just this example, we can see the interaction of at least 3
different bootloaders and there are two different orders in which
either OS can be invoked.
Note that I inadvertently forgot the multi-boot with MACOSX scenario
in the previous mail. I have done this successfully in the past with
an iMac but dare not attempt this today with my wife's Macbook Air
without risking being exiled to the couch!
Richard
You can do as you wish - and the more testing you can do - the happier
we will be.
However, for the majority of people they don't have the time.
The testcases we use are in fact generic ones for Ubuntu and all the
flavours who use the tracker.
I can almost guarantee that if we made the upgrade testcases anymore
complicated with more variants that the numbers reported would drop from
the ~12 we'll get to close to zero.
There is no way that we can cover ever eventuality.
Elfy
--
Ubuntu Forum Council Member
Xubuntu QA Lead
--
xubuntu-devel mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/xubuntu-devel
