On Thu, Jun 25, 2020 at 1:16 AM Kent Gibson <warthog618@xxxxxxxxx> wrote: > > On Wed, Jun 24, 2020 at 04:20:49PM +0200, Bartosz Golaszewski wrote: > > śr., 24 cze 2020 o 16:19 Kent Gibson <warthog618@xxxxxxxxx> napisał(a): > > > > > > On Wed, Jun 24, 2020 at 04:04:09PM +0200, Bartosz Golaszewski wrote: > > > > wt., 23 cze 2020 o 06:02 Kent Gibson <warthog618@xxxxxxxxx> napisał(a): > > > > > > > > > > Rename priv to gcdev to improve readability. > > > > > > > > > > The name "priv" indicates that the object is pointed to by > > > > > file->private_data, not what the object is actually is. > > > > > It is always used to point to a struct gpio_chardev_data so renaming > > > > > it to gcdev seemed as good as anything, and certainly clearer than "priv". > > > > > > > > > > Signed-off-by: Kent Gibson <warthog618@xxxxxxxxx> > > > > > > > > > > > > > Ugh now it's gcdev and gdev everywhere and it doesn't really make it > > > > more readable. Maybe chardev_data or cdev_data? > > > > > > > > > > Agreed, it isn't ideal visually, but is at least more unique than priv. > > > Linus was going for short names recently (e.g. gc for gpiochip), so I was > > > going for something short. > > > > > > And I try avoid names ending in _data or _state or similar where they > > > don't really add anything. > > > > > > Would chardev or gchardev work for you? > > > > > > > Yes, chardev is fine. Even cdev is fine for me: gdev vs gcdev is > > confusing but gdev vs cdev looks better IMO. > > > > OK, I was avoiding cdev to try to make the name more likely to be > globally unique, hence the leading "g". > > To try to keep it short, how about attacking it from the other end and > reducing it to gcd? > That would also be in keeping with the naming convention I use in later > patches, e.g. glv for gpioline_values. > So gcd for gpio_chardev_data. Hmmm, or maybe gcdd? > > Why is it that naming things is always the hardest part ;-)? > I prefer cdev here but it's just a personal preference. If anyone has a better idea I'm happy to switch to it. Bart