Re: hardware detection using libpci
David Whedon wrote:
> I made a new version of ddetect with a few changes, even if we don't
> go the libdetect way it was instructive to figure out how to make
> multiple .udebs conviennently.
> I'm now generating 2 udebs which each provide the virtual package
> We want udebs to be small. Therefore they shouldn't contain anything more
> than the bare minimum to accomplish their function. This is against
> policy (no manpage, copyright, etc.). For this example
> the .udebs contain very little:
> $ dpkg -c snddetect_0.0.5_i386.udeb
> drwxr-xr-x root/root 0 2000-11-14 20:38:59 ./
> drwxr-xr-x root/root 0 2000-11-14 20:38:58 ./usr/
> drwxr-xr-x root/root 0 2000-11-14 20:38:59 ./usr/bin/
> -rwxr-xr-x root/root 26460 2000-11-14 20:38:59 ./usr/bin/snddetect
> Documentation, copyrights etc. will be in another udeb.
> The .udebs can be found at:
> Comments, critisisms are welcome,
snddetect didnt didnt my isa awe64, but ethdetect detect my rtl8139.
I can see that your approach to hardware detection is best at least in
the short term. I would still like to be able to avoid depending on
/proc (which libdetect uses), but thats probably not very important to