2014-01-27 Raphael Mack <address@hidden>
I knew that this argument will come up. And I think it is a good one.
:-)
BUT I think we will not be able to do without incompatibilities if we
want improvement. I even think that we have a good example here. A new
prefix in ANY (and every normal feature name) is highly prone to break
compatibility. And as I do not expect that we currently have much users
in general and even less using the new operator, I think it is
acceptable to break compatibility in this case, after we discussed it
publicly - what we currently are doing. We may even think of a general
procedure how we want to introduce incompatibilities. Should we do some
public voting?
BTW in this case it is not a big deal anyhow, as it is fairly simple to
remove or add the prefix "&" alias in lazy_out...
I like that last idea. Something like a LAZY_STRINGS utility class? To be continued ;-)
About EDC in general:
it depends on ESE_TYPED_VISITOR and some other classes. Cyril, what do
you think about them? Shall we merge them? Rename them? Or generally
rework them?
Definitely rework. After all those years I'm not sure I still understand the reasons of such a convoluted design. It is strange at best :-/
Cyril ADRIAN (from office)
To any NSA and FBI agents reading my email: please consider whether defending the US Constitution against all enemies, foreign or domestic, requires you to follow Snowden's example.
| My latest G+: Happy 2014 to everybody |