<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi, <br>
<br>
Is this proposal to make SOAS a live stick capable of installing
Sugar on conventional systems (Trisquel, ...)?<br>
<br>
We have a live version of this problem on the server side.<br>
<br>
Jerry Vonau wrote a script mkusbinstall based on live-cd. OLE Nepal
switched to unetbootin for NEXS 6_31 (OS-7 on CentOS 6.4).<br>
I have been trying to make this work cleanly with BERNIE - to no
avail.<br>
<br>
One problem is that the user needs to be root. This is not possible
for a script unless it is launched by a live user. Unetbootin is <br>
a gui implementation.<br>
<br>
What I am looking for is a way to make a bootable usb stick that is
ready to install XS without user having to supply any configuration
<br>
information (like path names to image or /dev for usb stick) - sort
of an all-in-one unetbootin.<br>
<br>
The steps require formatting the usb device (as would be true for
SOAS), copying the image to the disk, and running live_cd to create
the environment on the <br>
usb stick. <br>
<br>
In the SOAS case, the usb stick presumably runs live and has the
option to install for some target platforms.<br>
<br>
Tony<br>
<br>
<div class="moz-cite-prefix">On 03/20/2015 06:26 AM,
<a class="moz-txt-link-abbreviated" href="mailto:sugar-devel-request@lists.sugarlabs.org">sugar-devel-request@lists.sugarlabs.org</a> wrote:<br>
</div>
<blockquote
cite="mid:mailman.11474.1426803962.10755.sugar-devel@lists.sugarlabs.org"
type="cite">
<pre wrap="">Message: 6
Date: Thu, 19 Mar 2015 22:25:55 +0000
From: Iain Brown Douglas <a moz-do-not-send="true" class="moz-txt-link-rfc2396E" href="mailto:iain@browndouglas.plus.com"><iain@browndouglas.plus.com></a>
To: James Cameron <a moz-do-not-send="true" class="moz-txt-link-rfc2396E" href="mailto:quozl@laptop.org"><quozl@laptop.org></a>
Cc: <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:sugar-devel@lists.sugarlabs.org">sugar-devel@lists.sugarlabs.org</a>
Subject: Re: [Sugar-devel] GSoC projects
Message-ID: <1426803955.2592.56.camel@vey-waldorf>
Content-Type: text/plain; charset="UTF-8"
Hi James,
Thank you for taking the time to make a thoughtful contribution.
Perhaps you will forgive me if I brainstorm this a bit.
On Fri, 2015-03-20 at 08:48 +1100, James Cameron wrote:
</pre>
<blockquote type="cite" style="color: #000000;">
<pre wrap=""><span class="moz-txt-citetags">> </span>I've often thought of making such an application, because of the
<span class="moz-txt-citetags">> </span>difficulties that some people report with downloading files and
<span class="moz-txt-citetags">> </span>putting them on USB drive.
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>The problem with an application is one may end up having to explain
<span class="moz-txt-citetags">> </span>how to download the application; transferring the issue from the
<span class="moz-txt-citetags">> </span>original problem to an application that was supposed to fix the
<span class="moz-txt-citetags">> </span>problem.
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>In the meanwhile, I have been working the overall problem as a
<span class="moz-txt-citetags">> </span>training and experience issue, and maintaining a structured
<span class="moz-txt-citetags">> </span>document:
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span> <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://wiki.laptop.org/go/Download">http://wiki.laptop.org/go/Download</a>
</pre>
</blockquote>
<pre wrap="">Thanks for that - I believe that systematic approach would be great
backup for those experiencing difficulties downloading.
(Using curl is a sound idea from the point of view that one set of
instructions can cover a host of different OS)
</pre>
<blockquote type="cite" style="color: #000000;">
<pre wrap=""><span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>Some further ideas for what your application might do:
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>1. the initial download,
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>2. resuming an interrupted download,
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>3. verification of download using md5sum or other hashes,
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>4. media verification, reading back the files or image to check that
<span class="moz-txt-citetags">> </span>writing was successful and the media still works.
<span class="moz-txt-citetags">> </span>
</pre>
</blockquote>
<pre wrap="">I think I am right that 4 is covered already by livecd-iso-to-disk, so
(in my model) the user only has to write a bootable CD.
If one knew that a SoaS CD would always make a "Sugar stick", the
prospect of selling the CD, (by third parties ?) becomes more doable.
</pre>
<blockquote type="cite" style="color: #000000;">
<pre wrap=""><span class="moz-txt-citetags">> </span>I've no evidence of proportion of people who have problems with
<span class="moz-txt-citetags">> </span>downloading files and putting them on media; perhaps it is a
<span class="moz-txt-citetags">> </span>non-problem.
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>A more correct approach would be to do research and survey of people
<span class="moz-txt-citetags">> </span>before and after such an application is made available. A GSoC
<span class="moz-txt-citetags">> </span>project could be padded out with this research, and easily fill three
<span class="moz-txt-citetags">> </span>months.
<span class="moz-txt-citetags">> </span>
<span class="moz-txt-citetags">> </span>A systems engineering view would change the product so that the files
<span class="moz-txt-citetags">> </span>don't have to be written to media in any particular way. That's what
<span class="moz-txt-citetags">> </span>we did with the original XO laptops, but SoaS bootable images are
<span class="moz-txt-citetags">> </span>different because of the typical PC firmware being so exacting.
<span class="moz-txt-citetags">> </span>
</pre>
</blockquote>
<pre wrap="">I think this would be achieved if `liveinst` could be persuaded to write
<b class="moz-txt-star"><span class="moz-txt-tag">*</span>only<span class="moz-txt-tag">*</span></b> to an automatically confirmed target USB, with the host hard
drive locked out during install and during use of the stick, and grub
instructed to find only the USB SoaS system.
With reasonably priced availability of 8 GB sticks, this would seem a
preferable option to me.
Regards,
Iain
</pre>
</blockquote>
<br>
</body>
</html>