On Thursday, 28 February 2013 at 15:31:21 UTC, Andrei
Alexandrescu wrote:
On 2/28/13 5:54 AM, Walter Bright wrote:
On 2/27/2013 11:55 PM, Jonathan M Davis wrote:
Again, please see how lexer.c works. I assure you, there is
no double
copying going on, nor is there a double test for the
terminating 0.
I know what the lexer does, and remember that it _doesn't_
operate on
ranges,
and there are subtle differences between being able to just
use char* and
trying to handle generic ranges.
Hence the need to invent SentinelInputRange.
I don't think the sentinel input range is a blocker for redoing
the parser (with ranges) in D. This discussion has probably run
its course. The right thing to do at this point is port the
lexer and figure what primitives are necessary from its input.
Andrei
An actual sentinel range is trivial to implement, and the
algorithms that can *actually* truly exploit it are rare.
While I'm not against having such ranges in phobos, I'd just be
weary to provide too many traits for them, or trying to have
phobos exploit them either. Ranges have enough primitives as it
is.
Just provide it with a disclaimer than it will only ever be
useful for a select class of algorithms (parser), but that the
optimization opportunity will be ignored by the rest of phobos.
