I've just had a somewhat difficult experience following an upgrade that went awry from 5.2.001 to 5.2.11. Not sure what went wrong exactly, but my ET8500 got into a state where it wouldn't boot into OpenVix and for some reason my usual USB stick for updating wouldn't work either. I've got it back up and running again after a couple of days of headscratching.

The main blocker to getting booted seemed to be this message in the logs. I've not got a copy of the exact text, but the gist is this:
Code:
[Harddisk] Unable to determine structure of /dev
[Harddisk] new Harddisk ->'->
...
undefined variable card in Harddisk.py in bus(self) at line 132
Looking at the code in Harddisk.py, this suggests that self.type is not set, and sure enough, there doesn't seem to be a /dev/.udev or a /dev/.devfsd on the box. This causes self.type to be set to -1. Later the code checks for the 2 defined values, so the -1 falls through, ultimately leading to a crash due to card not being set.

Now things are up and running properly again, I can see a /dev/.udev. Google suggests that .devfsd is a bit "old school". With this in mind, would choosing udev be a sensible to default so that this crash would be avoided? I'm not familiar with any of the other boxes that OpenVix runs on, so can't say whether this is a silly idea or not!

The other blocker to getting running was screensaver.py kicking in when there was no screen size available (did I mention that my system was *really* screwed :-D) with doMovePicture crashing when random.randint(1,self.maxx) caused an exception, due to maxx being 0.

Once I'd patched both these, I was able to get into the OpenVix image restore menus and reflash from there. Easy enough to protect against, but very fringe.

So the question is whether these changes are of general use and worth submitting a patch for?