On Wednesday, 29 June 2016 at 10:02:08 UTC, Dicebot wrote:
Please don't. There is a dire need to fix the implementation
first, currently DIP25 is almost unusable for anyone who
doesn't know exact limits of the feature - compiler diagnostics
are pretty much non-existent. Right now it is miles away from
what I'd call a success.
See this d.learn discussion
(http://forum.dlang.org/thread/[email protected]) for example of diagnostics problems I am talking about.
There are also plenty of open issues in bugzilla about current
implementation :
https://issues.dlang.org/buglist.cgi?quicksearch=dip25&list_id=209219
I don't think DIP25 is salvageable. The bugs here can be fixed,
but that won't mean much. The approach is unprincipled. If there
is something to be learn from the @safe story is that it doesn't
work as a general approach.
Additionally, the fact that is isn't used much tells us that it
is probably too limited to be useful enough to pay for itself.
