Hello Nick,
"BCS" <n...@anon.com> wrote in message
news:a6268ff154ca8ccddf1ef51e...@news.digitalmars.com...
Hello Ellery,
Generally I think D's CT capabilities have a way to go yet before
this would be worth tackling. E.g. how do you build a parse tree if
pointers and classes aren't allowed in CT code?
You use the type system. I can say from experience that (memory
issues aside) it works.
Trivial example?
Building an arbitrary tree:
int find(char[] data)
{
foreach(int j, char c; data)
switch(c)
{
default: break;
case '[', ']': return j;
}
return data.length;
}
int match(char[] data)
{
int i = 0;
foreach(int j, char c; data)
switch(c)
{
default: break;
case '[': i++; break;
case ']': if(--i == 0) return j;
}
return data.length;
}
template Tp(T...){ alias T Tp; }
struct Bar(T...) { }
template Foo(char[] data)
{
static if(data.length == 0)
alias Tp!() Foo;
else static if(data[0] == '[')
alias Tp!(Bar!(Foo!(data[1..match(data)])), Foo!(data[match(data)+1..$]))
Foo;
else
alias Tp!(data[0..find(data)], Foo!(data[find(data)..$])) Foo;
}
import std.stdio;
void main()
{
writef("%s\n", Foo!("So [she was [considering ][in her] own[ mind [(as
well] as] she could, ]for the").stringof);
}
Output:
tuple("So ",(Bar!("she was ",Bar!("considering ") ,Bar!("in her") ," own",Bar!("
mind ",Bar!("(as well") ," as") ," she could, ") ),"for the")
By placing things under different aliases and whatnot, named members can
be done.
Also, if the lisp like constructs puts you off, the construction of the tree
can be done via more or less the same parsing code and some fairly trivial
ctfe+mixin
--
... <IXOYE><
