[lm-sensors] [patch 2.6.23-pre] s/SENSORS_/HWMON_/ in Kconfig, Makefile

[Mark M. Hoffman]

Hi Jean, Jim:

* Jean Delvare <khali at linux-fr.org> [2007-06-15 20:20:09 +0200]:
> On Thu, 14 Jun 2007 14:56:56 -0600, Jim Cromie wrote:
> > Jean Delvare wrote:
> > > Thinking about it some more, I found 2 problems with your patch:
> > >
> > > * It breaks the ams driver:
> > > $ grep -r CONFIG_SENSORS drivers/hwmon/ams
> > > drivers/hwmon/ams/ams-core.c:#ifdef CONFIG_SENSORS_AMS_I2C
> > > drivers/hwmon/ams/ams-core.c:#ifdef CONFIG_SENSORS_AMS_PMU
> > > drivers/hwmon/ams/Makefile:ams-$(CONFIG_SENSORS_AMS_PMU)                += ams-pmu.o
> > > drivers/hwmon/ams/Makefile:ams-$(CONFIG_SENSORS_AMS_I2C)                += ams-i2c.o
> > > drivers/hwmon/ams/Makefile:obj-$(CONFIG_SENSORS_AMS)            += ams.o
> > > drivers/hwmon/ams/ams.h:#ifdef CONFIG_SENSORS_AMS_I2C
> > >
> > > You need to update these too.
> > >
> > > * It breaks a lot of default configurations:
> > > $ quilt grep CONFIG_SENSORS arch | wc -l
> > > 2729
> > > A number of these are for the drivers/i2c/chips part of your patch.
> > >
> > > So this will be a much bigger patch than you originally posted.
> > 
> > Does the size (and the default config breakage!) change anything from 
> > your ends ?
> > the patch is easy to do ( perl -pi.bak -e 's/SENSORS_/HWMON_/'),

BTW: not quite that easy, assuming that i2c/chips SENSORS_* will change to
something other than HWMON_*.

> > and easy to test (make allmodconfig - I didnt do this before, sorry)
> > but its not clear that its value-in-clarity is worth the churn (seen by 
> > everyone who builds)
> 
> Admittedly, this becomes much bigger than I initially thought, so I'm a
> bit hesitant now. I think it's still worth it for the i2c/chips part,
> because there, we're calling SENSORS devices which aren't and this is
> confusing; and the count is lower. For HWMON vs. SENSORS, it can be
> confusing but at least it's not plain wrong - so I'd say leave it alone.
> 
> But this is just my feeling of the moment. If others feel more strongly
> one direction or the other, I won't insist either way.

I think I agree w/ Jean.  At least, I would prefer that the i2c/chips
conversion happens first.  ;)

Regards,

-- 
Mark M. Hoffman
mhoffman at lightlink.com