DebiaN900 - A set of scripts to facilitate the installation of native Debian (WIP) - Page 9 - maemo.org

wicket | # 81 | 2015-02-10, 23:51 | Report

Originally Posted by baranovskiykonstantin View Post

Hi!

First of all want say thank you all for your hard work!

I want to share the experience of the assembly system using these scripts.

Host system: Ubuntu 14.04 32bit.
Configuration files are left unchanged.

1. build_kernel.sh
Kernel build will fail (circular dependency between libphy and of_mdio).
Solved the problem with a patch.

I know about the circular dependency problem with Linux 3.16. It sounds like the problem was caused by this commit. I'll look into it.

Originally Posted by baranovskiykonstantin View Post

2. install_debian.sh
The first installation of the system also failed. As it turned out the problem was established scratchbox. Later, repeated the installation on another computer using the same operating system without scretchbox, is making progress, but met another problem. To solve this problem need to disable automatic mounting of the card reader and mount the card manually.

Just to make it clear to anyone else, these scripts are not intended to be run inside Scratchbox.

Originally Posted by baranovskiykonstantin View Post

3. configure_u-boot.sh
Here, everything went without problems.

Finally managed to run the system on N900

I hope it will help others avoid similar problems.

Thanks for the tips!

The Following User Says Thank You to wicket For This Useful Post:
AapoRantalainen

wicket | # 82 | 2015-02-10, 23:57 | Report

Originally Posted by sulu View Post

Depending on your location some mirrors will be faster than others and some might even have an unstable connnection.

For example I come from Germany (Berlin region), so ftp.de.debian.org should be best in my case. But in fact ftp.cz.debian.org is usually faster and more stable for me.

So if you live closer to New York than to Indianapolis (that's where debian.org is registered) in terms of server hops, usually also but not necessarily geographically chosing the Columbia University mirror would make sense.

http.debian.net should act as load-balancer/redirector and select the best mirror automatically. You can test it here. In my experience, some mirrors do occasionally encounter the odd problem during package updates.

Originally Posted by saponga View Post

EDIT: Its running now.

Glad you got it working.

wicket | # 83 | 2015-02-11, 00:02 | Report

Originally Posted by Android_808 View Post

wicket: any updates lately?

Unfortunately not. I still have lots of plans for this project but as a post-grad student with no income right now, I've found it hard to justify dedicating any significant time to it. It's not helped by the fact that I only have one N900 which serves as my everyday mobile phone/computer but doubles as my device for testing so booting into Debian becomes a problem given that calls don't currently work in Debian.

I've recently been thinking that maybe I need to change the direction of the project and evolve it into a full Linux distribution based on Debian (and/or Devuan). It's becoming less and less like Debian with the stuff that gets added to it (e.g. non-Debian kernel) and with the decisions I make about what packages should or shouldn't be included by default. My thoughts are that it currently needs to be bleeding-edge to deliver the best possible experience but the long term objective would be that many of the differences could eventually be merged back into Debian. My personal focus would still be for usability on the N900 and Neo900 but I would also like to look into getting libhybris running on it so that it will be usable on Jolla and supported Android devices.

I would hope that this would generate much more interest and external contributions to the project than it currently has. In turn I could raise funds via donations, hopefully enough to be able to justify being able to dedicate up to 20 hours per week of my time. Raising funds would also enable me to buy a second N900, and cover hosting costs for a website and APT repositories. In the longer term, I will look for a sponsor to make it sustainable.

The DebiaN900 scripts will form the basis for building the distro so they won't suddenly disappear and I will continue work on them in the meantime.

To those reading this, what are your thoughts? Is this a crazy plan or do you think it could work?

The Following 7 Users Say Thank You to wicket For This Useful Post:
AapoRantalainen, klinglerware, nokiabot, realu, reinob, sulu, Xagoln

DDark | # 84 | 2015-02-11, 07:16 | Report

Originally Posted by wicket View Post

I've recently been thinking that maybe I need to change the direction of the project and evolve it into a full Linux distribution based on Debian (and/or Devuan). It's becoming less and less like Debian with the stuff that gets added to it (e.g. non-Debian kernel) and with the decisions I make about what packages should or shouldn't be included by default. My thoughts are that it currently needs to be bleeding-edge to deliver the best possible experience but the long term objective would be that many of the differences could eventually be merged back into Debian. My personal focus would still be for usability on the N900 and Neo900 but I would also like to look into getting libhybris running on it so that it will be usable on Jolla and supported Android devices.

I personally don't like the idea of bleeding-edge approach, but I like the idea of debian-based distro with voice-calls support.

Originally Posted by wicket View Post

Raising funds would also enable me to buy a second N900

What about getting in touch with Neo900 team and obtain one of those Mozilla donated phones?

The Following 5 Users Say Thank You to DDark For This Useful Post:
klinglerware, nokiabot, reinob, sulu, wicket

sulu | # 85 | 2015-02-11, 08:18 | Report

I agree with DDark. You should get in contact with the Neo900 devs.
Maybe you can team up in your efforts to bring Debian to the N(eo)900.

Besides phone support, power consumption seems like a major issue to me. I get about 6 hours of battery runtime with DebiaN900 (idle with backlight off but not suspended).
I know it's a huge leap but I'd like to have four times as much (trading suspend with still working calls for actually doing something on the phone).

The Following 5 Users Say Thank You to sulu For This Useful Post:
DDark, klinglerware, nokiabot, reinob, wicket

wicket | # 86 | 2015-02-11, 16:47 | Report

Originally Posted by DDark View Post

I personally don't like the idea of bleeding-edge approach, but I like the idea of debian-based distro with voice-calls support.

Just to clarify what I mean by bleeding-edge. The distro will be built on top of Debian so my current thinking is to provide separate images based on Debian stable and Debian testing with sid being available to build via the scripts. Bleeding-edge refers to the extra stuff on top of Debian that will be provided by my distro. This will help the project to work at an accelerated pace. I don't have the resources right now for the thorough testing that goes into producing a stable distro so I don't see any other way of going about it.

Originally Posted by DDark View Post

What about getting in touch with Neo900 team and obtain one of those Mozilla donated phones?

Thank you for the suggestion. I will contact them.

The Following 3 Users Say Thank You to wicket For This Useful Post:
DDark, klinglerware, reinob

wicket | # 87 | 2015-02-11, 16:59 | Report

Originally Posted by sulu View Post

I agree with DDark. You should get in contact with the Neo900 devs.
Maybe you can team up in your efforts to bring Debian to the N(eo)900.

Good idea. I will contact them to find out if they have anything planned with regards to Debian support.

Originally Posted by sulu View Post

Besides phone support, power consumption seems like a major issue to me. I get about 6 hours of battery runtime with DebiaN900 (idle with backlight off but not suspended).
I know it's a huge leap but I'd like to have four times as much (trading suspend with still working calls for actually doing something on the phone).

Me too. A lot of this stuff is new to me so it will take me some time to investigate and develop solutions. I have however recently started an experiment which should allow us to run much of Fremantle on top of Debian with little overhead and hopefully it will bring its power saving features with it.

The Following 5 Users Say Thank You to wicket For This Useful Post:
DDark, klinglerware, reinob, saponga, sulu

pharmakon | # 88 | 2015-02-11, 20:11 | Report

Originally Posted by wicket View Post

I would hope that this would generate much more interest and external contributions to the project than it currently has. In turn I could raise funds via donations, hopefully enough to be able to justify being able to dedicate up to 20 hours per week of my time. Raising funds would also enable me to buy a second N900, and cover hosting costs for a website and APT repositories. In the longer term, I will look for a sponsor to make it sustainable.

It's a shame that you don't have a spare to work on. I have a number of N900s in various conditions - one has a dead microphone, one is one of those HK refurb/forgeries w/ the missing magnet on a flimsy back cover, one is my currently daily phone (used and abused) which has suffered some wicked drops and has dead USB disease (I've been too lazy to repair it because I just rsync and I have spares to use as battery chargers), and finally another used one from ebay that I'm planning to make my daily driver.

I'm in the middle of a apartment move (which is incidentally good for taking account of where everything I own is), so it may take a short while, but I have a N900 surplus and I'd like to fix your N900 deficit.

PM me so I have a reminder.

pharmakon
(bought N900 on release day, lurker on tmo since 2010, Neo900 purchaser.)

The Following 8 Users Say Thank You to pharmakon For This Useful Post:
endsormeans, klinglerware, nokiabot, peterleinchen, reinob, saponga, sulu, wicket

Android_808 | # 89 | 2015-02-11, 22:17 | Report

Kind of similar to your aim, I'm trying to get use Debian packages where possible without modification (only starting dbus earlier at the moment), stick in some Nemo base packages (mce, dsme, statefs etc) and then stick a gtk3 port of hildon on top. Currently just fighting a load of bashisms in the script files they use...with no real shell programming experience

The idea was that the Debian and Mer/Nemo packages (with exception of packaging) would be maintained upstream, reducing the amount of work needed to maintain the distro. The Maemo>Meego>Mer/Nemo heritage of some of the packages should allow some of the other open source Maemo packages to be updated/ported. Basically, a modernised Fremantle clone.

The Following 7 Users Say Thank You to Android_808 For This Useful Post:
AapoRantalainen, endsormeans, klinglerware, nokiabot, reinob, sulu, wicket

sulu | # 90 | 2015-02-12, 09:22 | Report

Originally Posted by wicket View Post

I have however recently started an experiment which should allow us to run much of Fremantle on top of Debian with little overhead and hopefully it will bring its power saving features with it.

"Easy Maemo"?

Maybe you should also contact Aapo!
He tried to rebase Maemo on Squeeze [1] and Wheezy [2] and must therefore have some idea of what awaits you.

edit:
@Android_808:
Did you consider checking whether your efforts are compatible with those of FPTF?

[1] http://talk.maemo.org/showthread.php?t=89416]
[2] http://talk.maemo.org/showthread.php?t=91709

The Following 5 Users Say Thank You to sulu For This Useful Post:
AapoRantalainen, Estel, klinglerware, reinob, wicket