[Nut-upsdev] New default for SNMP and USB (and HAL) compilation

[Peter Selinger]

Indeed, this was a bug. Thanks for finding it. Yes, it should have
been nut_have_libusb and not nut_with_libusb. I fixed this now. 
I tested a lot of the changes I made, but there are so many
different possible control paths that I missed this obvious one.

-- Peter

Charles Lepple wrote:
> 
> On 12/28/06, Charles Lepple <clepple at gmail.com> wrote:
> > On 12/28/06, Peter Selinger <selinger at mathstat.dal.ca> wrote:
> >
> > > I have solved this slightly differently, by using EXTRA_LIBRARIES
> > > instead of noinst_LIBRARIES. Your solution will not work if the user
> > > selects something like
> > > --without-hal --with-drivers=hald-addon-usbhid-ups
> >
> > Further exposing my ignorance on automake variables, I can't see why
> > this simple sanity check doesn't work.
> >
> > * On Ubuntu 5.10, with libusb headers and hiddev.h, but no HAL headers.
> >
> > * checked out a clean copy of the trunk (rev 666)
> >
> > * ./configure && make
> >
> > The "Feature summary" matches what I have:
> >
> > Feature summary:
> > enable SSL development code: yes
> > build CGI programs: no
> > build upsclient library: no
> > build serial drivers: yes
> > build SNMP drivers: yes
> > build USB drivers: yes
> > enable HAL support: no
> >
> > ... but no USB drivers are built.
> >
> > If it says 'build USB drivers: yes', do you still have to pass in
> > '--with-drivers=...' and explicitly name the USB drivers?
> 
> Not sure if this helps any, but I found this in config.log:
> 
> HAVE_HIDDEV_FALSE=''
> HAVE_HIDDEV_TRUE='#'
> HAVE_LIBUSB_FALSE=''
> HAVE_LIBUSB_TRUE='#'
> 
> so there might be a typo here:
> 
> configure.in:353:AM_CONDITIONAL(HAVE_LIBUSB, test "${nut_with_libusb}" = "yes")
> 
> I can't seem to find where ${nut_with_libusb} is set, although there
> is a ${nut_have_libusb} in m4/nut_check_libusb.m4, and ${nut_with_usb}
> in configure.in.
> 
> -- 
> - Charles Lepple
>