[mythtv] About packaging CVS versions

[Dwight Hubbard]

> That's a nifty-sounding script.  It's unfortunate about Isaac's
> decision, but a least those of us with strict RPM-only policies will
> still be able to use the latest MythTV code while it's still crunchy in
> milk, as well as enabling installation in remote locations (where
> there's no tech available, like my mom's house .. because I KNOW she'll
> want one of these when mine's finally built and she sees it)

Well, I modified the script to default to downloading and installing the
.8 stable release for now.  The CVS download and build code is still there
but requires a one line change to re-enable it.

I'm wondering if I add an option to build the CVS version with debugging
enabled would that address Isaac's concerns?  I.E. in the event of a
problem the user can rerun setup and enable debugging during the build.

There's still quite a bit of work to do though, specific problems are:

I recently added some code that is Distribution specific, currently I only
have the code for Redhat 9 which is what I'm building the setup stuff on. 
It should be fairly easy to modify it for Redhat 8.0 and Debian though.

It currently blows away the database and rebuilds it each time.

It doesn't build some of the stuff I don't use, in particular lirc (I have
the ATI USB remote) and MythWeb.

It's probably missing building a couple of the prerequsite packages since
I already have them installed and so don't notice the issue.  I'm nearly
positive it needs additional code to install and configure xmltv.

It currently assumes it's setting up a combined frontend/backend server.

It currently assumes a direct connection to the internet (doesn't work
with proxy servers)

I'm hoping sometime this week to go through the MythTV install docs step
by step and compare them to the script to verify it completes all the
tasks mentioned in the documentation.

The current version is up on http://mythtv.dwightandamy.com