Am Mittwoch, 14. September 2016 15:43:14 UTC+2 schrieb Markus Zimmermann: > > On Wednesday, September 14, 2016 at 3:37:10 PM UTC+2, Alan Donovan wrote: >> >> On 14 September 2016 at 09:32, Markus Zimmermann <zim...@gmail.com> >> wrote: >>> >>> Do you think this might be worth bringing up in golang-dev? >>> >> >> No. We can't make this change to the language without breaking existing >> programs. >> > > That is unfortunately. I will just write a linter check for that. Thanks > for the talk! >
I think it is perfectly fine to have constants which are unused inside a package: A package exporting just a bunch of physical constants (think of c, alpha, h, ...) but not using any of them seems useful and not a sign of an error or even a hint for current or future problems. Regarding UNexported constants: I could imagine myself defining some physical constants from other constants, e.g. const RK = 25812.8074555 const rk = h/(e*e) with the exported RK the official measured value and rk the value as computed (from other constants). Now rk serves as a nice documentation and test can compare RK and rk and do some sanity checking (for this rk could live in a _test.go file, I know and would be used.) V. -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.