On Sep 7, 2013, at 9:43 AM, Konrad Hinsen wrote:
> That sounds reasonable, but in the scenario we are discussing
> (accessing data structures implemented in Typed Racket from
> plain Racket), the data structure never passes in that direction,
> except if it's constructed from a plain Racket data
David Van Horn writes:
> Thee is also this:
>
> https://pkg.racket-lang.org/info/ralist
>
> Which is not written in typed racket and has an un-contracted form for
> efficiency.
Installed, looks good. I'll try it in real life soon. Thanks!
Neil Toronto writes:
> > The implementati
At Thu, 05 Sep 2013 09:01:58 -0600,
Neil Toronto wrote:
> The best advice I have is to keep performance-critical loops either all
> typed or all untyped, unless the only things that cross the contract
> boundary are flat first-order datums like flonums, strings and
> vectors. (The math library's fl
On 09/05/2013 08:42 AM, Konrad Hinsen wrote:
Scott Klarenbach writes:
> Check out some of the functional data structures found here:
>
> http://www.ccs.neu.edu/racket/pubs/sfp10-kth.pdf
>
> VLists may be of particular interest in your case.
That's the stuff in the pfds package, right?
On 9/5/13 10:42 AM, Konrad Hinsen wrote:
Neil Toronto writes:
> FWIW, `vector->immutable-vector' is pretty fast. It's usually the least
> significant part of an O(n) operation. Its two biggest problems are that
> it allocates memory and annoys people.
That's a good summary of my first imp
Neil Toronto writes:
> FWIW, `vector->immutable-vector' is pretty fast. It's usually the least
> significant part of an O(n) operation. Its two biggest problems are that
> it allocates memory and annoys people.
That's a good summary of my first impressions ;-)
> If you're working in Typed
troductory talk on Contracts and Functional Contracts,
> with most examples in Racket (Robby Findler)
>
>
> --
>
> Message: 1
> Date: Wed, 04 Sep 2013 14:56:27 -0600
> From: Neil Toronto
> To: users@racket-
On 09/04/2013 10:28 AM, Konrad Hinsen wrote:
For the kind of data I am working with, immutable vectors look like
just the right choice: immutable and O(1) element access. However, I
am discovering that they are a real pain to work with.
Pretty much any vector operation returns a mutable vector:
For the kind of data I am working with, immutable vectors look like
just the right choice: immutable and O(1) element access. However, I
am discovering that they are a real pain to work with.
Pretty much any vector operation returns a mutable vector: vector-map,
vector-drop, vector-split-at, etc.
9 matches
Mail list logo
