Remounting rootfs somewhere else to reveal files hidden under mount points is the key, otherwise you can't be sure you have copied everything. At least with 770/n8x0 there were some 'hidden' nodes in /dev/ used early at boot time before it gets remounted and udev is started.
I've made exact copies of the pristine UBIFS rootfs (extracted from firmware image, mounted on-device) both using cp and tar - rsync shows no difference.
Then I created /mnt/initfs, removed fstab autogeneration from rcS-late and
added remounting of / as rw to rcS.
both boot but they hang a startup-pin-query
do you have a pre-release N900 or RD mode enabled?
i've read the full thread and, to be honest, i still don't get what you are attempting.
do you want to get a normal debian on the fr and for that purpose package everything not yet in debian (if possible patching things already there but modified by nokia)?
or do you want to continue the path of maemo, ie using the horribly crippled debian infrastructure employed by nokia, but .. yeah ... what?
sometimes it looks like you do not want to use debian but deb packages -- that's what maemo does now and does, imo, bad.
sometimes it looks like you, in fact, want to install a normal debian (using the available debian armel) and port missing parts and modifications to it, maybe rebuilding armel packages with optimizaition for the omap (what is it, armv7? v8?).
could someone, please, enlighten me?
i used debian armel for almost 2 years with the openmoko freerunner and it works well -- with a device far inferior to the n900, both in terms of available memory and speed.
clinging to "compatibility with maemo/meego" is imo the wrong way anyway -- those, who care for maemo won't switch to anything else, be it full fledged debian or debianized maemo and those who are interested in a full debian, won't hardly cling to maemo.
as long as maemo apps are available as source, it should be easy to build them for debian, too.
the only thing so far, after two weeks, i would like to see in debian, is skype. anything else is either available or has substitutes.
The purpose of this project is to integrate all Maemo5 (incl. proprietary) and future MeeGo components into a Debian base system running on the Nokia N900 and future devices, for the native installation of Maemo5, MeeGo and standard Debian software.
Picking your words: "want to install a normal debian (using the available debian armel) and port missing parts and modifications to it, maybe rebuilding armel packages with optimizaition for the omap".
So, is this project going to package open source Maemo 5 components for Debian and submit it to "upstream Debian" and maintain it there? This would be very welcome, as Maemo 4 packages already exist for Debian, and porting Maemo 5 stuff to Debian (or your Moebian) is easier when Hildon 2.2 and friends are available in the Debian repositories. Also, you can then build your project by simply starting from Debian and throwing the non-free packages in the mix.
After initial effort, I did not do much for it - both because bad lack of time, and unsurance that all this is really needed.
However, I'm really angry by all that PR1.2 story. That we had to wait for months after announcement, without access to current state as we have in free software world, without even any details on what the real problems are. And that after PR1.2 was released, many bugs are still there, and some things have been broken, without hope for a fix before some more months.
Because of all that, high chances I will try to resume this activity. Depending on situation: will there be some time windows, will any real alternatives arise, will somebody be interested or not.
same here!
The little time I have for Maemo activies I had devoted for the enhanced kernel (and a little bit for USB hostmode).
It's also an open question how many open-source replacements for the closed components we will see in MeEgo, possibly making our task much easier.
I still couldn't fix the major showstopper for me, i.e. start-pin-query hangs when booting from eMMC/SD, but I should try again with PR1.2.
Due lack of time and absence I won't be able to work on it before July.
After initial effort, I did not do much for it - both because bad lack of time, and unsurance that all this is really needed.
However, I'm really angry by all that PR1.2 story. That we had to wait for months after announcement, without access to current state as we have in free software world, without even any details on what the real problems are. And that after PR1.2 was released, many bugs are still there, and some things have been broken, without hope for a fix before some more months.
Because of all that, high chances I will try to resume this activity. Depending on situation: will there be some time windows, will any real alternatives arise, will somebody be interested or not.
If you want to help, I suggest you read the posts about installing a Maemo5
copy on eMMC or SD card on the mailing list https://garage.maemo.org/pipermail/moebian-devel/ and try it with the latest PR1.2.
It would be interesting to know whether the startup-pin problem persists.
