[Nut-upsdev] recent and planned changes

Peter Selinger selinger at mathstat.dal.ca
Mon Sep 19 20:21:06 UTC 2005


Arnaud Quette wrote:
> 
> > ...
> > I think it is a bad idea to implement such an option. I cannot see how
> > to do it except as a hack.
> > 
> > I think selecting a device based on the bus ID or bus serial number is
> > a much better idea, so I'll continue working on that.
> > 
> > 
> sure, a stable naming like physical path, bus id, ... would be ideal.
> the deal is simply: we should be still able to distinguish 2 devices with  
> the same model/mfr and no serial. Index was simply to skip detected ups 
> until we reached Index. I've not studied much libusb 1.0, but it might have  
> a better bus/dev path notion than the current 0.1. 

I just checked out libusb 1.0 from their CVS, and it seems that the
differences to libusb 0.1 are very minor. Their interface is still not
reentrant (uses global state), and they have not added anything new to
the way devices are identified and enumerated.

I recommend leaviung physical USB paths as a low priority item on the
todo list, until someone needs it and volunteers to implement it. 

-- Peter



More information about the Nut-upsdev mailing list