On 11/23/2011 07:59 PM, Peter Alexander wrote:
On 23/11/11 6:05 PM, Jonathan M Davis wrote:
On Wednesday, November 23, 2011 18:41:46 Timon Gehr wrote:
On 11/23/2011 12:53 AM, Jonathan M Davis wrote:
This basic idea was discussed a couple of months back, and it was
determined that it was overly complicated for very little gain.
I strongly disagree.
Well, you're free to disagree, but that's essentially what the discussion
resulted. And glancing at the thread ( http://www.mail-
archive.com/[email protected]/msg65324.html ), it looks like
you
posted a couple of times in it, so I'd expect you to be at least somewhat
aware of what was discussed in it. It's nice in theory, but there are
a number
of complications for what most people consider to be a minor gain. And
for the
folks who really don't like transitive const, it doesn't go anywhere
near far
enough, since they want full-on caching with const, not just lazy
evaluation,
which just isn't going to happen.
- Jonathan M Dav
Just to be clear, I don't think anyone has any problem with the
transitivity of const, but rather the fact that D only offers bitwise
const and not logical const, and also that D protocols enforce bitwise
const when only logical const is needed.
Exactly. This is how pure functional programming looks like in D at the
moment:
http://pastebin.com/C6vf9DQQ
I didn't spot any 'immutable' or 'const'.
Bitwise const is needed for safe concurrency, but only logical const is
needed for pure functional programming.
And lazy evaluation already suffices.