Re: have a mate-desktop CD was Re: magnet links and pushing them to meta-search engines.
[ Dropped the mate list - this isn't relevant there... ]
On Sat, Mar 17, 2018 at 12:38:33AM +0530, shirish शिरीष wrote:
>.<snipped all>
>
>Dear Thomas,
>
>In response to https://lists.debian.org/debian-cd/2018/03/msg00005.html
>
>Thank you for sharing the info. I probably will try to get a couple of
>usb sticks next week for work and probably will also try to see how
>rufus responds and if possible get in touch with Pete Batard to take
>the discussion forward.
>
>I found it pretty disconcerting especially for students or for that
>matter anybody who wants
>to take the plunge into Debian and get stuck like that.
>
>I also perhaps failed to mention that rufus tried to install syslinux
>instead of grub2 which at least to me feels more convenient and
>friendly but that's a discussion outside of this purview.
Then you were definitely using ISO mode in Rufus. The main problem
with that (and with unetbootin and other tools that work like this) is
that it's very easy to replace boot loader files with versions that
don't work well together, or don't work the way we expect and have
tested. If you're blindly modifying what the d-i and debian-cd scripts
have produced, then we have no way of knowing if reported bugs are our
fault or not. That's why we *really* don't recommend unetbootin, for
example. We have had a lot of bugs reported by novice users that we
just cannot reproduce.
For Debian installer images for both i386 and amd64, we create iso
hybrid media which will run 4 different ways:
* BIOS boot from CD/DVD (isolinux)
* BIOS boot from USB (isolinux)
* UEFI boot from CD/DVD (grub)
* UEFI boot from USB (grub)
and that's a deliberate choice for maximum flexibility from a single
image. If a tool like unetbootin changes the setup of the media, it's
very likely to break things here.
>During a discussion with Steve Mcntyre during debconf 2016, he had
>shared about both the unrealiability of the usb drives themselves as
>well as the data so now always use shasums to make sure the data
>stream is correct.
Yay!
>This https://www.debian.org/CD/faq/#write-usb I had already done
>before. I used the built-in
>certutil and Get-FileHash on MS-Windows powershell to make sure the
>sum was correct.
>
>Now that I'm on Debian GNU/Linux did as I had copied the same on one
>of my usb disks, the .iso image which I just transferred to the hdd.
>
>shirish@debian:~/games$ ./check_debian_iso SHA1SUMS
>debian-buster-DI-alpha2-amd64-xfce-CD-1.iso
>Piping 331776 blocks of 'debian-buster-DI-alpha2-amd64-xfce-CD-1.iso'
>through 'sha1sum'
>to verify checksum list item 'debian-buster-DI-alpha2-amd64-xfce-CD-1.iso'.
>331776+0 records in
>331776+0 records out
>679477248 bytes (679 MB, 648 MiB) copied, 1.35603 s, 501 MB/s
>Ok: 'debian-buster-DI-alpha2-amd64-xfce-CD-1.iso' matches
>'debian-buster-DI-alpha2-amd64-xfce-CD-1.iso' in 'SHA1SUMS'
>
>although trying to do the same with /dev/sdb or /dev/sdb1 results in -
>
>shirish@debian:~/games$ ./check_debian_iso SHA1SUMS
>debian-buster-DI-alpha2-amd64-xfce-CD-1.iso /dev/sdb
>Does not look like an ISO 9660 filesystem: '/dev/sdb' magic=' '
>
>That's the usb key/thumb drive on which debian iso image is done by unetbootin.
So unetbootin will not have written an exact copy.
>Although it appears to have worked though as was able to get a working
>Debian GNU/Linux system at the end, was able to get updates, get the
>net working and so on and so forth.
>
>FTR, I could have used cp or dd but that would have been only if I had
>taken my lappy that day. It should not be a requirement that we should
>have a debian laptop because in real-world scenarios many people not
>might have access to an installation nor should they require one,
>although do agree it's handy if it's there.
Absolutely, hence why we have recommendations for rufus in DD mode, or
before that win32diskimager.
--
Steve McIntyre, Cambridge, UK. steve@einval.com
< Aardvark> I dislike C++ to start with. C++11 just seems to be
handing rope-creating factories for users to hang multiple
instances of themselves.
Reply to: