Comment 13 for bug 65174

On Sun, 2007-07-22 at 19:24 +0000, Oliver Gerlich wrote:
> Brian: it seems that m-a is not so much immature as "only" badly integrated in Ubuntu...

No, I disagree. I don't think it's any different than it is in Debian.
It's just the way the tool works. It's "sub-optimal" (if you don't want
to use the word "immature").

> Feature-wise you can run "m-a a-i lirc" and get working LIRC modules installed

Indeed. But why should I have to (remember (how) to do this every time
the kernel is upgraded)?

> (it asks a few times in the process for a "y" whether you want to install kernel-headers and bla bla, and then automatically installs kernel-headers and build-essentials).

Yup. More stuff I should not have to (remember (how) to do).

> What's lacking is some GTK GUI integration which hides the "y" pressing :-)

Nope. Completely wrong solution, IMHO.

> and maybe it would be nice if m-a would then be run automatically at every kernel upgrade...

Ahh. Now you are talking about the _maturity_ of DKMS. Instead of
trying to wrangle m-a to do exactly what DKMS actually does, why not
just use DKMS. Probably DKMS could be made to simply call m-a if that's
the way you really wanted to go with it.

> It's not lacking very much, in fact it only seems to lack an end-user compatible GUI.

Again, I disagree. A GUI is the wrong solution. Automation, hidden
from the poor end-user is the right solution.

Or as Mario says in a followup, integration with linux-ubuntu-modules,
so that binary modules come out of the box.

b.

--
My other computer is your Microsoft Windows server.

Brian J. Murrell