On Thu, Dec 09, 2010 at 03:31:30PM -0500, Jean Delvare wrote: > On Thu, 9 Dec 2010 07:40:46 -0800, Guenter Roeck wrote: > > Actually, I think I'll have to re-do the whole patch anyway. As you pointed out, > > it does not reflect the documented attributes, and there are other problems > > such as how and when to display alarms. Without going into details, > > I don't think we really disagree on anything > > I agree. > > > - the idea was to retain the existing output. > > Up to a certain point, yes. Where the current output was decided > because it made sense, we want to keep it as is. But being identical to > the character isn't a goal per se. In particular, if switching to more > generic code is possible and makes the code smaller and/or more > readable and/or more maintainable, we definitely want to consider that, > even if this changes the output slightly for some chips. > Related question: When making libsensors API changes, do I update the API version to 4.3.2, or stick with 4.3.1 ? Guenter _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors