Custom carpc distro

**monkeyBox** · 11-06-2006, 06:51 PM

Hi all, I was thinking about starting a new project, similar to minimyth, that will aim to be a carpc-specific linux distro that would have a relatively small footprint. It will also be fine-tuned to support the features available in the VIA EPIA series motherboards. The minimyth project already has most of the difficult stuff in place (ie, getting all of the epia-specific drivers working, getting X to work, etc..). The idea would be to take minimyth and strip out the mythtv-specific stuff, and add in a carpc frontend w/ all the necessary linux apps installed (mplayer/xine, roadster, etc..). Among other things, I could add-in support for various touchscreens, support for "plug-n-play" devices such as iPod and other music players, etc... Unfortunately, right now I have no idea what carpc frontend to use. All the ones I've seen seem fairly "blah" compared to the different windows offerings. I don't feel like building my own, but who knows, it may come down to that :P

Anyways, I just wanted to post here to see if there would be any interest in such a thing.

**BarryWoods** · 11-06-2006, 08:42 PM

Wasn't this already suggested?? It would be cool, but you'd have to specify the hardware supported.

**tripzero** · 11-06-2006, 09:46 PM

Go for it man!! There are already several of us working on front-ends. A car specific distro for those would be good. Also making suggestions on how to improve the already written and in-progress front-ends would probably help. I'm sure the authors are willing to listen to requests and suggestions.

**the_crowbar** · 11-07-2006, 08:02 AM

I say go for it as well. I would be willing to help in whatever way I can. Right now I run Gentoo but cross compile it from my main desktop. A fully loaded Gentoo setup takes ~550MB once the portage tree and source code directories are removed. I don't see why a EPIA specific distro should not easily fit on a 256MB flash drive.

There is no reason why we could not make an image that could easily fit on a 256MB flash drive. If it was as easy as writing out the image and setting the BIOS to boot from USB I could see many people trying out Linux in their car. I have been playing around with Gentoo on a flash drive for testing in my car.

I think SilverWolf 2.0 is coming along. Nanonymous, that Kev000 is writing, also looks promising. I am not much of a coder, but I do have quite a bit of experience in Linux admin and Perl/PHP scripting.

Cheers,
James

**monkeyBox** · 11-07-2006, 08:22 AM

Originally Posted by BarryWoods

Wasn't this already suggested?? It would be cool, but you'd have to specify the hardware supported.

The hardware supported (as mentioned in the original post) would be the EPIA-M series motherboards (namely right now the M10k, as that's the only one I have), and the embedded audio/video on that board.

**monkeyBox** · 11-07-2006, 08:34 AM

It's good to see that some people are interested in it. My plan is to kick-start the project in the hopes that others will take it and run with it, as I don't have all the time in the world to spend in front of a computer. My initial to-do list is:

1) Copy over the minimyth GAR repository to SVN
2) Get the GAR build system working on my local box
3) Strip out any mythtv-specific stuff
4) Decide on a frontend
5) Get the frontend built in the GAR system
6) Start a website for the project...

**reborg** · 11-07-2006, 08:49 AM

I'd offer to get involved too, but unfortunately I don't have time at this point.

I can however offer to donate web hosting and if you want to house a CVS repo there that is fine too. There is also a (as yet unused ) Vbulletin on my host with which you could run a support/info/request forum.

**blk** · 11-07-2006, 12:13 PM

i like the idea and i might get involved too but i think we should spend more time into planning than just rushing into one thing - it may be a dead end.. what i had in mind was an arch-linux as setup which is quite slick, but since i haven't ever used it i can say much about up/down-sides from a practical point of view... i also had gentoo running but didn't really see a big performance improvement (especially when considering the time spent compiling)

i could setup a wiki where we could propose and discuss different aspects/wishes.
what would definately be cool is a modular design where you can choose between different frontends/features/etc. depending on what hardware you might have.. i have a touchscreen,an epia MII-10k and a lot of experience using linux (mainly debian based distros (debian, ubuntu) but also some rpm-based (suse) and gentoo) i am a bit of a programmer but usually i do Perl,PHP or shell-script

as a frontend i'm currently using myth-svn with a working cle266 acceleration. maybe we could also investigate using framebuffers since we wouldn't need an X-server for this (although i don't know about touchscreen support there...with GPM?) what about a reduced-Xserver?
an alternative frontend possibility is the window-manager (WM) based one (there's a about this somewhere here) which i found quite cool too.

boot speed is also a very important aspect and currently i get mythfrontend up in about 20s from POST, im using initng on ubuntu (for testing, i intended to switch to arch-linux once i knew what i wanted)

..someone for a wiki?

[update]
having seed miniMyth: it does sound like a good starting point..but i would keep in sync with them and not completely branch off.. or sync back from time to time (like ubuntu does with debian after nearly every release)

**tripzero** · 11-07-2006, 03:38 PM

If we can get some good organization here I'm also willing to help out where I can. My c++ skills are at your disposal.

We can plan out the features using launchpads feature blueprint system: https://features.launchpad.net/

**monkeyBox** · 11-07-2006, 06:42 PM

Minimyth has built in functionality/scripts for enabling/disabling certain options. Furthermore, with the GAR build system (as I understand) you can enable/disable certain features at build-time. For example, if we wanted to offer a build of the distro that's just intended for use with character LCD's we could do that quite easily.

We would definately want to keep watch on any improvements to minimyth that could improve this distro, but it will still be mostly a branch of minimyth due to the change in general purpose of the distro.