got c2d to run on jni.h

Jarrett Billingsley jarrett.billingsley at gmail.com
Mon Mar 16 12:09:41 PDT 2009


On Mon, Mar 16, 2009 at 2:40 PM, Nicholas Jordan <asdf at some.com> wrote:
> == Quote from Jarrett Billingsley (jarrett.billingsley at gmail.com)'s
> article
>> On Mon, Mar 16, 2009 at 1:50 PM, Nicholas Jordan <asdf at some.com>
> wrote:
>> > I put some fields at the front of my first attempt as a note while
>> > prototyping, the compiler did not accept the syntax from the
> sample
>> > code. I had to do it like this:
>> >
>> > invariant {long fail__= -123456;}
>> >
>> > with the brackets surrounding, for this question - what I want
> here is
>> > several values that are never changeable and always represent some
>> > sort of status return from calls - always the same. The above code
> is
>> > not compliant with the code shown in the discussion of what D is
> all
>> > about.
>> The problem is that this conflicts with the D1 syntax for
> class/struct
>> invariants.  This is probably why the immutable keyword has been
>> introduced, and I have a feeling invariant as a form of constness
> will
>> be renamed immutable.
>> Just use:
>> immutable
>> {
>>     long foo = 4;
>>     int bar = 12;
>> }
>> But you said that these are return statuses?  Why not use enums
> instead?
>
> ran htod.exe on jni.h - this looks sane to me. Some concepting issues
> but is a step forward in computer science, so what I do is just list
> that file on the command line to C:\dmd\bin\dmd.exe along with what it
> produces on EZ-Twain dot h along with other things I might write that
> would make more sense to me as a file scope code base, the compiler
> de-facto takes file scope into account?
>
> In other words, in file foo.d I have a function, on which I want to
> feed it a bar from compilation unit recognition - this is sorta
> obvious for a compiler writer, seems to me to even throw some cs
> 301's, but redeclaration in each file scope would lead to the same
> issues we are trying to parallelize.
>
> I have a rich skill base in subtle nuances of parallelizing work, this
> is the time to fix this. If it gets broken later, one may need your
> burger-flippin resume' to be ready to deliver.
>

So I'm guessing you replied to my post by accident, then?


More information about the Digitalmars-d-learn mailing list