if anyone ever finds this thread via google, the simplest way i've found of
checking what is a valid type is to use Val{}. if Val{x} doesn't throw an
exception then x can be used as part of a type (the functions linked to
earlier are in C).
since tuples can contain data types, bits types, and nested tuples, you can
encode pretty much any value in "s-expressions" of tuples, where the first
value is a data type and subsequent values are arguments.
andrew
On Thursday, 9 July 2015 12:15:29 UTC-3, andrew cooke wrote:
>
>
> sure, i understand the reasoning, i just didn't know it was actually
> implemented / verified in any way. rewriting the code now... cheers,
> andrew
>
> On Thursday, 9 July 2015 12:03:57 UTC-3, Yichao Yu wrote:
>>
>> On Thu, Jul 9, 2015 at 10:59 AM, andrew cooke <and...@acooke.org> wrote:
>> >
>> > ah! thank-you. i had no idea about that.
>> >
>> > is there any kind of isbits container? will a tuple work?
>>
>> The type system uses === to compare the parameters (and it has to be
>> this way since it should be simple) so having a reference to a mutable
>> should never work. If you want an array of `isbits` as parameter, an
>> tuple of it should work on 0.4
>>
>>
>> Another way to see why a vector should never work is that e.g. you have
>>
>> a = [1, 2]
>> T1 = A{a}
>> push!(a, 3)
>> T2 = A{an}
>> T3 = A{[1, 2]}
>> T4 = A{[1, 2]}
>>
>> Should any of the T*'s be equal?
>>
>>
>> >
>> > thanks again,
>> >
>> > andrew
>> >
>> >
>> > On Thursday, 9 July 2015 11:56:45 UTC-3, Yichao Yu wrote:
>> >>
>> >> On Thu, Jul 9, 2015 at 10:52 AM, andrew cooke <and...@acooke.org>
>> wrote:
>> >> > I want to do what I wrote, I think! In particular, the type
>> parameter
>> >> > is
>> >> > itself a value, the polynomial x. It's immutable and a subclass of
>> >> > Integer.
>> >> > So I have no idea why the code should not work.
>> >>
>> >> It's not `isbits` though.
>> >>
>> >> >
>> >> > It's unusual to have a complex value like that in a type, I know,
>> but it
>> >> > makes logical sense here - it's the type of values in the quotient
>> ring
>> >> > with
>> >> > that ideal.
>> >> >
>> >> > Andrew
>> >> >
>> >> >
>> >> >
>> >> > On Thursday, 9 July 2015 11:44:42 UTC-3, Tom Breloff wrote:
>> >> >>
>> >> >> I'm not 100% sure I understand what you want, but here's a stab for
>> >> >> line
>> >> >> 16:
>> >> >>
>> >> >> ZField{T1.parameters[1],T2}(x)
>> >> >>
>> >> >>
>> >> >> On Thursday, July 9, 2015 at 10:32:28 AM UTC-4, andrew cooke wrote:
>> >> >>>
>> >> >>>
>> >> >>> Before I raise an issue I wondered if I've made some stupid
>> mistake
>> >> >>> here.
>> >> >>> The code is about as simple as I can make it. The idea behind
>> things
>> >> >>> is
>> >> >>> that you have a field of integers module 2 (GF2). Then over that
>> you
>> >> >>> define
>> >> >>> polynomials. And then you can define a Quotient Ring with the
>> >> >>> polynomials,
>> >> >>> which is analogous to the initial field, and so you can use the
>> same
>> >> >>> type as
>> >> >>> the rogiinal field. But you don't need to understand that!
>> >> >>>
>> >> >>> Here's the code:
>> >> >>>
>> >> >>> immutable ZField{N, I<:Integer} <: Integer
>> >> >>> i::I
>> >> >>> end
>> >> >>>
>> >> >>> immutable ZPoly{I<:Integer} <: Integer
>> >> >>> a::Vector{I}
>> >> >>> end
>> >> >>>
>> >> >>> T1 = ZField{2,Int}
>> >> >>> o = T1(1)
>> >> >>>
>> >> >>> T2 = ZPoly{T1}
>> >> >>> x = T2([o, o])
>> >> >>>
>> >> >>> ZField{x,T2}(x)
>> >> >>>
>> >> >>> And this is the error (from the last line, which is line 16)
>> >> >>>
>> >> >>> andrew@laptop:~> julia-trunk IntModN.jl
>> >> >>> ERROR: LoadError: TypeError: apply_type: in ZField, expected
>> Int64,
>> >> >>> got
>> >> >>> ZPoly{ZField{2,Int64}}
>> >> >>> in include at ./boot.jl:254
>> >> >>> in include_from_node1 at loading.jl:133
>> >> >>> in process_options at ./client.jl:305
>> >> >>> in _start at ./client.jl:405
>> >> >>> while loading /home/andrew/IntModN.jl, in expression starting on
>> line
>> >> >>> 16
>> >> >>>
>> >> >>> In short - I have no idea where the Int64 comes from.
>> >> >>>
>> >> >>> Thanks,
>> >> >>> Andrew
>> >> >>>
>> >> >
>>
>
