[XSCE] Rendrering maps on XO1 at Silar's Orphanage Haiti

James Cameron quozl at laptop.org
Mon Sep 2 18:50:21 EDT 2013

If you mean by "blowing up en route" an interruption to the download,
then we have a set of instructions for people (like me) who must deal
with very unstable internet service.


Focus on using wget or curl, as these can recover gracefully from an
interruption.  So can some browsers.

If you mean by "blowing up en route" that the download completes but
contains wrong data, then I have news for you: an rsync server with
the released builds is now available, and may be used to _repair_
areas of damage in a download using tools already on an XO.  Only the
damaged part is sent over the internet.  See the "Wrong Data" section
of the "Download" page on the Wiki.  The rsync server URL is:


I look forward to Nancie and the battery of other volunteers engaging
with the development community.  Perhaps there hasn't been enough of
that leading up to the release of 13.2.0, and now you are paying for

On Mon, Sep 02, 2013 at 08:33:32AM -0400, Adam Holt wrote:
> We're trying Release 12.1.0 in Haiti later today, thanks to Anna's careful
> testing confirming IIAB maps work there, that is if our expensive OS download
> to Haiti actually succeeds, without blowing up en route etc :)
> PS Nancie Severs has done very extensive testing of Release 13.x usability on
> XO-1s and we cannot support that slowness.  She and a battery of other
> volunteers have determined Release 12.x is the most well-rounded solution for
> XO-1s for now.
> On Mon, Sep 2, 2013 at 4:31 AM, James Cameron <quozl at laptop.org> wrote:
>     I've reviewed the thread so far.  The problem is interesting.
>     In summary:
>       * some of the HTML content you have deployed uses features that do
>         not perform well on the browser renderer used on 11.3.1, and you
>         hadn't noticed this until now,
>       * the cause of the bad performance is not directly related to the
>         slower processor speed of the XO-1, since it is only 50% used,
>       * the content works flawlessly with the browser renderer user on
>         13.2.0-13.
>     What I didn't see was:
>       * whether memory was depleted; might add swap; again, as part of a
>         test, not as a proposed solution,
>       * whether the content performed well on 11.3.1 on XO-1.5, XO-1.75 or
>         XO-4; so that the possibility of graphics driver bugs might be
>         considered,
>       * a way for interested parties to reproduce the problem; without
>         which we are limited to speculation,
>       * how any bug fixes to the browser renderer, graphics driver, or
>         kernel could be practically deployed; can you afford the time to
>         install an RPM to every XO?  Probably not.
>       * why this has anything to do with what a _server_ does; on the face
>         of it the server only delivers the content, and the
>         incompatibility is between the content and the browser, so the
>         only thing a server might do is deliver _different_ content, or
>         _refuse_ to deliver the content at all,
>       * exactly what reasons other than prejudice lead to the persistent
>         claim that 13.2.0-13 cannot be used on XO-1; because without
>         addressing these reasons there's _no way_ that a new build can
>         help.  That's one for devel@ thanks, please get involved.
>     --
>     James Cameron
>     http://quozl.linux.org.au/
> --
> Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !

James Cameron

More information about the Devel mailing list