I have had a lot of issues with standard installation devices working and then breaking after a reboot because of driver/kernel changes that I am not initiating. Instead of hacking my install to bits and ending up with something other than LinuxMCE I wish to develop a feature. My plan will be to write a small shell script and perhaps a few supporting files. If they work without changing much else than I would like a dev to try it out and then figure out how to submit it.
Devices currently breaking themselves as examples:
Hybrid install w/ onboard ATI 4200 HD. Can be installed from snapshots or 10.04 Final and works amazing with onboard HDMI w/ audio etc, as long as the hybrid is off-line. If the machine has Internet during install or any reboot after it loads back ports and flgrx driver which breaks it.
Acer Revo 1600R as MD. First install as new MD works flawless and uses the nvidia driver. The MD continues to work unless it has to reboot, be reset. On any reboot after initial it removes the nvidia driver automatically and will not run again until the moon# is deleted and re-detected as a new device.
/usr/Pluto/bin/freeze_settings.sh & unfreeze_settings.sh. (This would also need to be present in /usr/Pluto/diskless/##/usr/Pluto/bin/
Replace sources.list with a blank or commented out version?
My problem, and where I need some assistance please:
LMCEs entire structure is built around detection, self testing and configuring new devices. I had tried everything I can to prevent a problem driver (and will need to try for the Revo) from making two working devices into non-working devices. Utils.sh rewrites itself, killing sources.list alone would not stop other detections from going off such as storage radar etc...
I am going to write it and test it but,
If you wanted to cement a configuration of your own, which files/scripts/initd/modprobe/blacklists would you change?