[clug] Anyone want to give a talk for this week's CLUG meeting? [SEC=PERSONAL]

[tridge at samba.org]

Hi Antti,

 > I will bring a digital photo frame and cheap MP3 player once I finish at
 > a committee meeting.

thanks!

 > If you would like to provide a tarball of the subdirectory before/after
 > the meeting, out of meeting testing is also a possibility.

A tarball isn't sufficient unfortunately. I can provide a raw
filesystem image, but that would require wiping everything on the
device to use it.

So the technique I use is to run a script which loads a varity of test
patches, creating subdirectories FATTEST/$p where p is the patch
name. The script takes a name of a device to attach to.

You can see one version of the script here:

  http://samba.org/tridge/testdev

It creates MP3 files using a voice synthesiser (espeak), and JPG files
using raw postscript and pnmtools. The MP3 files contain the
synthesiser saying the name of the patch and a tag indicating what
file it is, while the JPG has the same information encoded as an
image.

So after you run it something like this:

  ./testdev /dev/sdg1

it will create /mnt/vfat/FATTEST/* with a bunch of MP3s and JPGs. You
then use the device, and you can tell exactly what patches work in
what way by just viewing images and listening to the 'music'.

So if you end up hearing the synthesiser saying 

   "This is a test 1 ORIG tagged"

then you know that the device can correctly see MP3 files using an
unpatched Linux system when the files have id3 tags on them.

If it fails to play a file with the following message

   "This is a test 2 PATCH2 another" 

then you know it can't see MP3 files created with PATCH2 in a
subdirectory with a long name.

and so on for JPG files etc.

This technique allows me to fully test devices even if they don't
display the full file name in their LCD (or even if they have no real
UI).

btw, I'm going to need someone to keep a record of the testing that is
done, and the results of the tests. Any volunteers?

Cheers, Tridge