Testing Summary, Auckland - 26 May 2012

Tom Parker tom at carrott.org
Sun Jun 3 07:53:49 EDT 2012

Do not adjust your clocks! This is last weeks report more than a week late. This week's report will be less tardy (hopefully) and will have include the power logs from last week.

Testing Summary, Auckland - 26 May 2012
Who: Fabiana, John, Oliver, Tabitha, Tom

We tried using the nandblaster to update our XO-1.75s and discovered that it is not suited to our use case as it’s just too slow. We only have a modest number of laptops to update, and we have a reasonably noisy wifi environment. 2 laptops managed to do the update after two full cycles through the image and the rest we did with USB sticks. Perhaps if we had requested more redundancy for error correction, but even if it had completed successfully on the first pass, it would still be slower than flashing from USB, including copying the file onto the USB. If we were doing more laptops or we weren’t actively sitting waiting for them to finish then it would be really useful.

olpc-au build 225 11.3.1-au on XO-1.5

Paint: OK. 
Record: Picture ok, movie on low res ok. Movie on high res - after a while it is jumpy with frozen images. Audio seems ok.
Memorise: Single player numbers ok . Created game, saved it and played it ok.
Measure First start with a lot of noise. Ok on second start. Clicking about does not seem to make it crash like the last time.
Robot. asked ‘who makes the babies’ and it returns “/random name/ colon” Why is it saying colon? otherwise seems fine
Physics Seems fine hand grab works fine. No surprises there.
Olli The sleep function does not seem to work when the activity is running. 

Maze works
Speak and robot works
Browse - crashed when trying to view the whitehouse.gov website to sign the petition “Require free access over the Internet to scientific journal articles arising from taxpayer-funded research.”
Memorize - played one sounds game but it was too quiet to hear if the guitar or drum was playing. 
Paint - works, can use the tools and add an image from Journal, the grayscale button has extremely different functionality to the rainbow button so I think they shouldn’t be placed together
Firefox - the screen looks a bit odd with all the different sizes of fonts, clicked a few links on the front page and they seemed to work (could see lists of books, chemistry and biology page, images list). From the world culture images list pages there are back buttons that don’t work (they say file not found). Music plays from the media section. Maps take a while to load.
on closing the firefox window it shows a firefox failed to start screen with a stop button
Moon - works
Clock - works
Tamtam mini - works
Stopwatch - works
Measure - starts, can adjust settings
Implode - works
Arithmetic - works
Numbers - works
GCompris Clickgame - works

12.1.0 os11 on XO-1.75
On Kiwi, write works, including formatting and inserting images from record.

On Kotare, started write, created a table and entered a few characters. Then shared within the neighbourhood and write stopped accepting data from the keyboard, cut and past and drag and drop still work, other applications still work with the keyboard, search box in write does not.

Had whio freeze after inserting a USB drive after stopping tcpdump

As reported last time and the time before, our yum problems persist. It seems the first invocation works but subsequent invocations do not. rpmdropbox.laptop.org is reporting what appears to be an incorrect content-type:

$ echo "GET /f17/repodata/primary.xml.gz HTTP/1.1
User-Agent: urlgrabber/3.9.1 yum/3.4.3
Host: rpmdropbox.laptop.org
Accept: */*

" | nc rpmdropbox.laptop.org 80 |head -n 12
HTTP/1.1 200 OK
Date: Sat, 26 May 2012 00:58:23 GMT
Server: Apache/2.2.8 (Ubuntu) DAV/2 SVN/1.4.6 proxy_html/3.0.0 mod_ssl/2.2.8 OpenSSL/0.9.8g mod_wsgi/1.3 Python/2.5.2 mod_perl/2.0.2 Perl/v5.8.8
Last-Modified: Thu, 24 May 2012 11:30:07 GMT
ETag: "1ac806b-25d5-4c0c692df2dc0"
Accept-Ranges: bytes
Content-Length: 9685
Content-Type: application/xml
Content-Encoding: x-gzip


Putting the same file on my own server yields an application/x-gzip content type and no content encoding.

It’s not clear whether or how this is related to the problems we see with yum, perhaps there is a transparent proxy which is sometimes re-serving the decompressed content?

More information about the Devel mailing list