On Wednesday, November 10, 2010 17:53:08 Andrew Wiley wrote: > Well, my assumption was that Phobos 2 was pretty much complete. If more > functionality is planned, then that's definitely a higher priority because > long term, that's a much better solution than any supplemental library.
Goodness, no. Phobos is nowhere near complete. Whole modules such as std.json, std.xml, and std.stream are planned for deprecation in favor of newer implementations (which haven't been done yet). New modules are on the way - such as std.datetime and std.parallelism. std.container has almost nothing in it - SList and Array. It's supposed to get pretty much every container which would be considered fairly standard, but those implementations haven't been completed yet. std.algorithm continues to get tweaked all the time. Phobos is changing all the time. Phobos is way better than it was, but it's far from done. There's plenty of work left to do on it. Eventually, what's there will likely stabilize to the point that even its ABI won't generally change (except presumably for major releases when necessary or when outright new functionality is added), but it's still in quite a bit of flux. The general API isn't even yet stable (though a lot of it doesn't change often). I fully expect that once Phobos is fully mature, it will rival Tango in every respect. But there's still a lot of work to be done on it. - Jonathan M Davis