On Thu, Dec 07, 2006 at 01:48:43PM -0500, Woody Suwalski wrote: > Physically you are correct. > > However we really want to use /dev/temperature which is a SUPERSET of Linux > /dev/temp. We have not much use of /dev/temp without this extra > functionality. > I see it is rather for cross-incompability between architectures: there is > nothing wrong with trying to run Netwinder user-mode programs on any other > architecture, but I do not want to actually mess with _their_ /dev/temp, but > get out if no propriatory /dev/temperature exists. Simply on Netwinder > with udev /dev/temp happens to provide this extra functionality. BTW. You can > run etch without udev, as I have demonstrated up to today ;-) > > So if "temperature" is not propriatory enough, it may need to be renamed to > nw_temperature or so. But this will break (maybe non-existant) historical > back-compatibility, so I would prefer to leave it...
Thanks for your quick replies. I don't necessarily disagree with you, I just don't have the context to understand. I think the biggest piece I am missing is: Under what circumstances is /dev/temp created with a different major/minor than /dev/temperature? I assert that if /dev/temp and /dev/temperature always have the same major/minor, they always behave the same way and backwards compatability can be maintained by having the application look for both /dev entries, and use whichever one is available. A udev symlink is also an option here - but it seems easier and cleaner to avoid messing with udev and just deal with it in nwutil. Apologies in advance for leaving this discussion midstream, I'll be offline for the next few days. -- dann frazier | HP Open Source and Linux Organization -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
