Nut upsdev - Jan 2009 - [nut-commits] svn commit r1734 - in trunk/scripts: hal hotplug udev

Citeren Arjen de Korte <adkorte-guest at alioth.debian.org>:
> Author: adkorte-guest
> Date: Thu Jan 15 20:02:22 2009
> New Revision: 1734
>
> Log:
> Freshly generated USB helper files
>
> Modified:
>    trunk/scripts/hal/ups-nut-device.fdi.in
>    trunk/scripts/hotplug/libhid.usermap
>    trunk/scripts/udev/nut-usbups.rules.in
Questions is, should we try to keep these (generated) files up-to-date  
in SVN as well, or is it better to svn:ignore them? The chance that  
packagers don't have perl installed (and hence, won't be able to  
generate them) is remote (automake won't run without it).

Why are we generating this during 'make dist' and not just during  
'make' by the way? (ducking for cover)

Best regards, Arjen
-- 
Please keep list traffic on the list

On Thu, Jan 15, 2009 at 3:11 PM, Arjen de Korte <nut+devel at
de-korte.org> wrote:
>
> Citeren Arjen de Korte <adkorte-guest at alioth.debian.org>:
>
> > Author: adkorte-guest
> > Date: Thu Jan 15 20:02:22 2009
> > New Revision: 1734
> >
> > Log:
> > Freshly generated USB helper files
> >
> > Modified:
> >    trunk/scripts/hal/ups-nut-device.fdi.in
> >    trunk/scripts/hotplug/libhid.usermap
> >    trunk/scripts/udev/nut-usbups.rules.in
>
> Questions is, should we try to keep these (generated) files up-to-date
> in SVN as well, or is it better to svn:ignore them? The chance that
> packagers don't have perl installed (and hence, won't be able to
> generate them) is remote (automake won't run without it).
As long as we are not using any special Perl modules, it shouldn't be
an issue for developers (who would use automake anyway) to regenerate
them.

We might want to postpone the svn:ignore thing for a little while so
we make sure we are tracking what versions of the files get released
in 2.4.0 final. (As an aside, I still think we should do a bit of
filtering so that the hotplug files do not automatically match bogus
vendor IDs such as 0000, 0001, and ffff. Or at the very least, those
vendor IDs should be in a separate file, to be installed in an
optional package.)
> Why are we generating this during 'make dist' and not just during
> 'make' by the way? (ducking for cover)
I would tend to agree that it would get tested more frequently if it
were done during regular compilation.

--
- Charles Lepple