Re: [racket-dev] Adding Jacobi symbols to math/number-theory

Thanks everyone for the suggestions, they will be incorporated. Felleisen: Wie geht's? Your version is very nice, thank you for the improvements. Soegaard: I'll write a pull request soon. Thank you. As for the Legendre symbol, I believe it's already implemented

Re: [racket-dev] Re: Bug in Typed Racket optimizer and snapshots

Antonio, I can confirm that, with this example, the contract generation code we identified as a possible culprit is indeed the problem. For reference, here are a few data points. | # of union members | time in problem code | total time | ++--++

Re: [racket-dev] Adding Jacobi symbols to math/number-theory

May I propose a few small Racket-style improvements? Otherwise follow Jens's suggestions and #lang typed/racket (provide ;; coming to a Typed Racket near you soon #; [jacobi-symbol (-> non-negative-integer? (and/c non-negative-integer? odd?) integer?)] jacobi-symbol) ;; ---

Re: [racket-dev] Adding Jacobi symbols to math/number-theory

This makes a very nice addition. Write a few lines as documentation: https://github.com/racket/math/blob/master/math-doc/math/scribblings/math-number-theory.scrbl Move the tests to: https://github.com/racket/math/blob/master/math-test/math/tests/number-theory-tests.rkt Send a PR at Github.

Re: [racket-dev] Adding Jacobi symbols to math/number-theory

On line 7, the call to raise-argument-error should be: (raise-argument-error 'jacobi "odd?" 1 a n) This also speaks to the need for contracts plus types! Robby On Fri, Sep 11, 2015 at 4:59 AM, eu90h wrote: > Hello, > > I wrote a small procedure computing the Jacobi symbol. I thought it might b

[racket-dev] Adding Jacobi symbols to math/number-theory

Hello, I wrote a small procedure computing the Jacobi symbol. I thought it might be cool/useful to put in the math/number-theory pkg if possible. The procedure is contained in this gist , along with some test vectors. Thanks. -- You recei

Re: [racket-dev] Package versioning

On 10 Sep 2015 10:46 pm, "Robby Findler" wrote: > > Maybe the right thing is to just collect the set of licenses of all > dependencies of a given pkg as an alternative view of some pkg, and > then you could navigate around yourself to make decisions? +1 to that. Help the user get the relevant inf