From: William S F. <ws...@fu...> - 2014-01-23 19:55:33
|
On 23/01/14 19:20, Vladimir Kalinin wrote: >> I'd rather have the parser add 'feature:flatnested' to any >> class/template that really is nested if >> and only if NestedClassSupport == NCS_None. Sorry that was pseudo code, and could be a function call. > > I don't quite like the idea of introducing yet another global variable > in the parser, instead I was thinking to check for the > nestedClassesSupported() result in the same place where 'flatnested' > feature is handled and ignore/flatten the nested classes accordingly. Whatever you see fit, but I think it is better to just follow the same pattern that is done elsewhere. One day all the odd global variables can go into a settings class as part of moving all the source to C++. Following the current globals pattern will make it easy to port it later on. William |