Chris, I'm fine with committing this as long as it doesn't break anything which I don't see why it would.
Thanks, Wayne On 1/5/2016 11:11 AM, Chris Pavlina wrote: > Hi, > > layers_id_colors_and_visibility.h defines enum LAYER_ID as unsigned, > then #defines two negative constants to be used with the enum. > Comparisons of these generate the following warning: > > warning: comparison of constant 4294967295 with expression of type 'int' > is always true [-Wtautological-constant-out-of-range-compare] > > > Currently there is no problem (I actually dug into the assembler to make > sure the comparison was being generated). However, this is a time bomb > just waiting for a different compiler version which decides to optimize > out what it sees as a tautological comparison, and the fix is very, very > simple. > > Attached patch does two things: > > 1) Move the negative constants INTO the enum so that the compiler knows > to assign the enum a backing type that supports negatives. > > 2) Remove the unnecessary forced-typing of the enum. There is absolutely > no reason this enum has to be an unsigned-char; I suspect this was an > instance of overzealous and misguided "optimization". This > "optimization" will at best not do anything (as the compiler will decide > to pad out the char to the platform's word length anyway), and at worst > make things even /slower/ (if the compiler actually decides to access it > as a char). > > -- > Chris > > > > _______________________________________________ > Mailing list: https://launchpad.net/~kicad-developers > Post to : [email protected] > Unsubscribe : https://launchpad.net/~kicad-developers > More help : https://help.launchpad.net/ListHelp > _______________________________________________ Mailing list: https://launchpad.net/~kicad-developers Post to : [email protected] Unsubscribe : https://launchpad.net/~kicad-developers More help : https://help.launchpad.net/ListHelp
