On Wed, 23 Jan 2019 at 15:39, zyx <z...@gmx.us> wrote: > > I gave you a counter-example, where the committed change will break > things (when the Unknown would be changed to -1 instead). For the > consistency, the None is set to -1, the (other) Unknown-s to 0xff. Why > not make them consistent in a way to declare Unknown/None as the first > value of the enum with the value -1? Would that work for you? If not, > why not? I kind of miss what you are aiming for. Is it anything else > than having all Unknown being 0xff, which can eventually cause waste of > memory in combination with -fshort-enums (while the switch is > potentially dangerous, it can be useful for someone whom bundles PoDoFo > into his/her own project and doesn't expose it into the public)? > >
All you say makes sense but I'm gasping on inconsistencies all around the library and I genuinely tried to follow the most used convention by putting Unknown on 0xff. Actually, I follow this personal "golden rule" for *my* enums: 0 : UNKNOWN or a enumeration value with a semantic that make sense in a default initialization (both fit very well with zero initialization) 1...x : Actual values of the enumeration to be specified voluntarily -1 : NONE or ANY: values with special meaning to be specified voluntarily Of course I am not pushing such change because it would require review of all the enums in the library. For EPdfDataType particular case I would also be fine to have Unknown = -1 but I didn't do it this way as said because I tried to follow what is done in other enums. _______________________________________________ Podofo-users mailing list Podofo-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/podofo-users