[It has been suggested that I CC @devel with this weekly update.
Feedback from the list members would be appreciated.]
1. Indianapolis: Benjamin Mako Hill and David Cavallo gave keynote
talks and led a session on the laptop at the OpenMinds conference in
Indianapolis this week. Indiana is in the vanguard in the US on
laptops for learning (over 110,000 already deployed) and in using free
and open-source software (FOSS) for learning. The conference brought
together educators and developers to discuss issues and share
experiences. OLPC was highlighted for making laptops more affordable
everywhere and for our commitment to FOSS. In attendance were various
governmental entities about to begin 1:1 laptop initiatives.
2. Suspend/resume: John Watlington has written a long description of
the B3/B4/C1 suspend/resume problems, along with what it takes to
modify a B4 to correct the problems can be found at in the wiki (See
http://wiki.laptop.org/go/B4_Suspend_ECR). A small pre-build will be
assembled next week to test the circuit changes introduced since the
C1 build.
3. Schedules: The Trial-3 Open Firmware (OFW) first code-drop is
scheduled for Monday. Testing of the Q2C28i is happening this weekend,
and a final drop will be available for Quanta next Wednesday. Trial-3
is essentially complete, but we do not need to drop it to Quanta for
another week or two, so we will consider critical bug fixes—if there
is adequate time for testing. Everyone should please be focusing on
First Deployment bug fixes, minor features, and, most especially,
testing.
4. Test: Alex Latham spent most of the week performing suspend/resume
testing. We now have a setup that is pretty easy to get running and
keep running. Yani Galanis has spent the week documenting and testing
various network configurations. There were a number of
bugs/enhancements found this week that will help people who have
recently been experiencing problems connecting to their home access
points; for example, now that we support multiple key types, it is
necessary to type $: in front of a hex key for a WEP connection.
Michael Stone is spear-heading a Test Sprint day to review test
plans, automation, and finding ways to make it easy others community
to help out. Next Wednesday will the the test sprint day. Please join
in. (Details will be sent to devel, sugar, and testing mailing lists.)
SJ Klein will be getting the wiki to produce inline diffs of watched
pages in response to changes to those pages so that we can more
efficiently track the progress of the sprint.
5. Mesh view: Polychronis Ypodimatopoulos has developed a new
activity, Space, which displays an alternative mesh network
neighborhood; it offers a sense of space by placing you in the
center and everyone else in the mesh network at a distance proportional
to link quality between you and the node that is being displayed (See
http://web.media.mit.edu/~ypod/mesh/).
6. Kernel: Andres Salomon spent much of the week debugging
suspend/resume patches related to the display controller (DCON). He
also worked with upstream, massaging patches in, getting more patches
ready, and helping others with their patches.
7. Sugar activities: Simon Schampijer set up a page in the wiki for
the activity template (See
http://wiki.laptop.org/go/Activity_Template) in order to set a
standard by which activity developers communicate about their
projects. (Now that loading new activities is as easy as clicking on
an .xo bundle from the browser, there is certain to be more
activity-related traffic in the wiki.) Simon also implemented the
standard control for providing in-activity alerts (See
https://dev.laptop.org/ticket/2822); these can be used in the
activities and can be placed at the top or bottom of the window. He
has also begun work on a Sugar control-panel window.
8. X Window System: Bernardo Innocenti has gotten Xorg 1.4 fully
packaged and available for general testing; while we haven't done any
benchmarking yet, it seems to be quite a bit faster. There is still
need for a kludge in the kernel to help the hardware abstraction
layer (HAL) and evdev_drv see the glide sensor as a normal mouse, but
that will soon be fixed.
Bernie also reports that we have a fix for the glibc problems
affecting Ethiopian, but it requires upgrading to the F8 version of
the library. Replacing glibc at this later stage isn't as
destabilizing as it may seem: the only fallout Bernie can see is the
exposure of a latent memory allocation bug in the olpc-dm program,
which he has already fixed. Of course, we have more testing to do. Rob
Savoye may be helping us with the Geode specific optimizations in
glibc, for the benefit of Gnash and all other applications that rely
heavily on memcpy() and similar functions.
Walter Bender has been working with Bernie, Tushar Sayankar, Jens
Petersen, Parag Nemade, Manusheel Gupta, and Rosh Kamath on a
Devanagari keyboard for the laptop that will be deployed in Mumbai
(See
