[Fedora-livecd-list] Development of biarch live CD script.

[Richard Shaw]

On Tue, Aug 31, 2010 at 10:33 AM, Jasper Hartline
<jasper.hartline at gmail.com> wrote:
> On Tue, Aug 31, 2010 at 5:33 AM, Richard Shaw <hobbes1069 at gmail.com> wrote:
>>>> One thing I noticed, it seems the GRUB menu options are backwards for
>>>> 64bit vs. 32bit.
>>> I'm not sure what you mean about them being backwards. Care you elaborate?
>>
>> When I selected the 32bit menu option I got a complaint that I was
>> trying to run a 64bit kernel on 32bit hardware (qemu, my actual
>> hardware/Fedora is x86_64). Now that I think about it, does the script
>> do any checking on the order the ISO's are supplied? If not we may
>> need to add some logic. I'll look at my bash history, I may have
>> specified them in the wrong order...
>
> Yeah it doesn't check. I'm not sure how to go about checking them
> right off either
> or how to make it better, but I would need more time to look at it.
>
> Of course this is why we're doing testing so I'll keep it in mind that
> this is an issue.
> If you have any ideas let me know.

The only thing I can come up with without mounting both the squashfs
images is to do a regex search through isolinux.cfg and do something
like:

def arch(file):
    """Accepts a text file object to search through and returns system
architecture as a flag."""

    x32_reg = re.compile("i[3-6]86")
    x64_reg = re.compile("x86_64")

    for line in <isolinux file object>:
        if x32_reg.search(line): return '32BIT'
        elif x64_reg.search(line): return '64BIT'

Then of course whatever function calls this function will have to do
something with that information. We'll need some logic to make sure we
have exactly one positive return for each arch type.

>>> As far as vmmouse you see on the screen, that means
>>> xorg-x11-drv-vmmmouse is not installed.
>>> This is not an issue with mkbiarch.py though.
>>>
>>> Does it fail to boot any further for you from this point or?
>>
>> Yup, get's that far and stops.
>
> Ok can you test it on real hardware with a real USB stick too?
> I did testing in Qemu and it doesn't fail.
> I also did testing on real hardware and it also does not fail on
> either x86 32 or x86 64.

I don't think I have a USB stick handy that's over 1GB...

Richard