Andrei and Walter's proposal does not break existing code because it makes folders into modules.
On Tue, Apr 3, 2012 at 8:43 AM, deadalnix <[email protected]> wrote: > Le 02/04/2012 18:00, Jacob Carlborg a écrit : > >> On 2012-04-02 16:31, Don Clugston wrote: >> >>> To be brutally honest, I don't think that's got much to do with the >>> language. It's got to do with Phobos adopting the Big Ball Of Mud design >>> pattern. There's no reason for the existing modules to be so huge. Eg, I >>> created std.internal.math so that the math modules would stay small. >>> Not only are the modules huge, they import everything. >> >> >> I couldn't agree more. >> > > I did noticed that, but this isn't the only problem. > > >>> I'd like to see some attempt to fix the problem within the language >>> right now, before jumping straight into language changes. >> >> >> That's not very hard. It will just break existing code. >> > > Yes, this is the point : refactoring a big module into submodules is hard > because it break a lot of code, which is something we don't want ina > standard lib for instance. > > Not because it isn't possible, but because almost all D code repose on > phobos, so refactoring it into submodules is likely to massively break > existing code.
