As is, we have today enough changes in our package that we would have to merge anyhow from Debian, so, an additional build-dep its not a big deal (this means, reason #2 is not valid anymore).
However, I'm still not convinced that we should enable this at compilation time since it could really benefit only a minority of Ubuntu conky users.
Some data gathering should help: could somebody with the right hardware let me know what is the difference in terms of performance (memory and cpu) of using conky with an exec on nvidia-settings, and with the built-in ${nvidia} variable?
To have representative data please give conky a relatively short update interval and collect your samples at a frequency at least double of that.
Another possibility we can consider is to have a conky-nvidia binary package (or even a conky-all with all compile options enabled). However, to go this way we would need a clear request from the Ubuntu conky user base, which so far doesn't clearly transpire from the attention this bug report got.
As is, we have today enough changes in our package that we would have to merge anyhow from Debian, so, an additional build-dep its not a big deal (this means, reason #2 is not valid anymore).
However, I'm still not convinced that we should enable this at compilation time since it could really benefit only a minority of Ubuntu conky users.
Some data gathering should help: could somebody with the right hardware let me know what is the difference in terms of performance (memory and cpu) of using conky with an exec on nvidia-settings, and with the built-in ${nvidia} variable?
To have representative data please give conky a relatively short update interval and collect your samples at a frequency at least double of that.
Another possibility we can consider is to have a conky-nvidia binary package (or even a conky-all with all compile options enabled). However, to go this way we would need a clear request from the Ubuntu conky user base, which so far doesn't clearly transpire from the attention this bug report got.