Thanks for the clarification. In otherwords, if we go the GAR route, each person would either have to use the pre-built packages and not likely be able to modify them on their own (without having the exact build system) OR Each person would have to use GAR and compile the whole ball of wax on their own build system (right?) I guess that makes sense.
Being a slacker from way back, I like anything that uses Slackware. Look on the main MythTV list to see how many people are having problems/issues with RedHat versus Slackware (or any other distro for that matter).
I don't need to be a purist. I just want to have a full understanding. Now the picture is much clearer.
After I finish messing with Cinelerra, I'll get back over to MythTV and Mini-myth.
Yup, gives Larry an easier time upgrading and allows people who want to customize to build their own system the way they want. Another plus is to be able to have tarballs of extra features like the modules easily available so folks won't have to worry about deps, just untar in the minimyth filesystem and cramfs it up.dhartman wrote:
In otherwords, if we go the GAR route, each person would either have to use the pre-built packages and not likely be able to modify them on their own (without having the exact build system) OR Each person would have to use GAR and compile the whole ball of wax on their own build system (right?)
Should make everyone's life easier. Well...once it all works.
Just a quick question on customisation, how customisable is the GAR system? For instance, I run my own homebrew receiver config, also i include a customised version of Lcdproc in my cram fs system. Can GAR support these kind of changes or do all the required options have to exist in the CVS?
I was under the impression that the GAR system is either opt in on a package of code, or opt out. From your reply it seems maybe more than this is possible. Can you provide a little more detail about what is and isn't possible.
I suppose RTFM would be an easy answer, I must hold my hands up and say I still haven't done this on the GAR system yet. Still a nice little intro wouldn't go a miss if you have time.
You can pretty much make it do *exactly* what you want. If you want to change how a package is compiled, just edit the Makefile. Want to change optimations or CFLAGS, edit gar.conf.mk Want to add to the distro, copy a Makefile and edit it to suit your new program. The meta/minimyth-contents package is just a empty package with deps of all the programs that get built by default so you can change what gets included from there. Another package that does a lot of work is meta/minimyth, it takes the new file system out of the build dir, cleans it up and cramfs it.Nem wrote:I was under the impression that the GAR system is either opt in on a package of code, or opt out. From your reply it seems maybe more than this is possible. Can you provide a little more detail about what is and isn't possible.
One of these days a README will get written for this thing. Check out the gar.mk file, that's where all the commands are and its well commented. The lnx-bbc site has a bunch of info also.
Anyway, take a look if you want, I'm going to take a break for a day or so.
This latest build has Ivor's 1.5 libddmpeg driver, and _totte_'s 1.6 X patch over X 126.96.36.1992, as well as kernel 2.4.23-epia1 with Ivor's 1.4 kernel patches. All completely open source. I will probably upgrade to kernel 2.4.24 soon.
Info on this code is here:
My main problem is I can't really test because my backend now is 0.12, and is being used by an actual in-use minimyth-0.2, so I can't upgrade to 0.13 to play with minimyth-0.3...
Gar is not too hard to learn. And you could always downgrade yours to work with 0.12. It's really easy to do that. Just start playing with it.
It's possible that the patch has somehow changed, I'll check it when I get a chance. Please cut 'n paste the exact error if you can.
Also, there are others that have expressed interest in working on the modules, so unless you want to do it all yourself for the learning experience (which is completely valid) I would try to work with the other interested parties. Note that we will most likely have to come up with an NFS solution for this, so it will probably not be a quick chore. I want to think it through a bit. In the mean time, have a ball.