On 09/03/10 17:12, Larry Evans wrote:
On 09/03/10 15:46, Dave Abrahams wrote: [snip]
Since, as I mentioned, I had trouble understanding how apply worked, and the code seems pretty complicated, at least to me, I was hoping DeBruijn's method would offer simplifications.
From the examples I've seen so far, this would make it easier for bind library writers at the expense of usability. On th other hand, once lambdas start to use protect() I'm usually giving up on them ;-)
OK, but at the time I was having trouble, I didn't know of an alternative. I'll try to remember the particulars of the problem I had to illustrate the point. That also might give Mr. Sankel another use case.
The attached illustrates the difficulty I was having with mpl::apply. Maybe the apply_apply typedef would be a good use-case for De Bruijn's method. I suspect the problem with apply_apply is something akin to the variable capture problem (search for "variable getting captured in: http://en.wikipedia.org/wiki/%CE%91_conversion#.CE.B1-conversion ). That problem is what De Bruijn's method is supposed to solve. In apply_apply the "variables" are the _1 and _2. Since those are used both in the op typedef and in the apply_apply typedef and the apply typedef uses op, I'm guessing that leads to what amounts to the variable capture. Of course Mr. Sankel's code doesn't operate on types, only values; however, I'm guessing it could be transformed to work on types also, just like much of mpl's code is a translation of value functions to type functions. -Larry