To bring this to the attention of a wider audience, what how should bindings be handled?

I've created a pull request to include SQLite3 and its bindings in Phobos. Some discussion has taken places so please read up:

https://github.com/D-Programming-Language/phobos/pull/46

It will probably be best to continue the discussion here, or maybe there :P

-- 
Liberty means responsibility. That is why most men dread it. ? - George Bernard Shaw

On Mon, 16 May 2011, Jesse Phillips wrote:

> To bring this to the attention of a wider audience, what how should bindings be handled?
> 
> I've created a pull request to include SQLite3 and its bindings in Phobos. Some discussion has taken places so please read up:
> 
> https://github.com/D-Programming-Language/phobos/pull/46
> 
> It will probably be best to continue the discussion here, or maybe there :P
> 

I think here is better.. wider audience, better visibility.

---

Regarding your point about the maintainability of a binding layer being the same as the library it binds to.  That's wrong.  The binding layer only changes as the public interface of the library changes, which it shouldn't be doing except at major release points.  The library itself changes at every bug fix.  Two potentially (and generally) drastically different change rates.

Later,
Brad

Fair enough. I guess there should be a vote and I'll gladly make the changes.

On that note, should bindings just be .di files?

On Mon, May 16, 2011 at 11:36 AM, Brad Roberts <braddr at puremagic.com> wrote:
>
> Regarding your point about the maintainability of a binding layer being the same as the library it binds to. ?That's wrong. ?The binding layer only changes as the public interface of the library changes, which it shouldn't be doing except at major release points. ?The library itself changes at every bug fix. ?Two potentially (and generally) drastically different change rates.
>
> Later,
> Brad

There's (at least) two major categories of d to c bindings:

1) straight extern(c) style bindings.  These can be just .di's.  There's no actual code, just symbol table entries.

2) wrappers that put D wrapper around a C api to make it more D-friendly. These almost certainly can't be just .di's.

Later,
Brad

On Mon, 16 May 2011, Jesse Phillips wrote:

> Fair enough. I guess there should be a vote and I'll gladly make the changes.
> 
> On that note, should bindings just be .di files?
> 
> On Mon, May 16, 2011 at 11:36 AM, Brad Roberts <braddr at puremagic.com> wrote:
> >
> > Regarding your point about the maintainability of a binding layer being the same as the library it binds to. ?That's wrong. ?The binding layer only changes as the public interface of the library changes, which it shouldn't be doing except at major release points. ?The library itself changes at every bug fix. ?Two potentially (and generally) drastically different change rates.
> >
> > Later,
> > Brad
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
> 

As Brad is the only one providing input, I'll work on a pull request with just the bindings this weekend.

I think that's the right way to go, too. Thanks!

Andrei

On 5/19/11 9:29 AM, Jesse Phillips wrote:
> As Brad is the only one providing input, I'll work on a pull request
> with just the bindings this weekend.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

On 5/19/2011 7:29 AM, Jesse Phillips wrote:
> As Brad is the only one providing input, I'll work on a pull request with just the bindings this weekend.

To be clear, I haven't reviewed the actual binding parts of it at all.  I just started and saw a hundred thousand lines of code and stopped.

I don't see were bindings need much review.  They're basically just a mechanical translation of a header file.  Wrappers, on the other hand, deserve a thorough review.

On Thu, May 19, 2011 at 12:42 PM, Brad Roberts <braddr at puremagic.com> wrote:

> On 5/19/2011 7:29 AM, Jesse Phillips wrote:
> > As Brad is the only one providing input, I'll work on a pull request with just the bindings this weekend.
>
> To be clear, I haven't reviewed the actual binding parts of it at all.  I
> just started and saw a hundred thousand lines
> of code and stopped.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/phobos/attachments/20110519/d34434f6/attachment.html>

Yeah, no wrapper here. The binding isn't my personal work either, just trying to package it up.

Thank you Brad for the feedback, and Andrei for agreeing :)

On Thu, May 19, 2011 at 9:46 AM, David Simcha <dsimcha at gmail.com> wrote:
> I don't see were bindings need much review.? They're basically just a mechanical translation of a header file.? Wrappers, on the other hand, deserve a thorough review.
>
> On Thu, May 19, 2011 at 12:42 PM, Brad Roberts <braddr at puremagic.com> wrote:
>>
>> On 5/19/2011 7:29 AM, Jesse Phillips wrote:
>> > As Brad is the only one providing input, I'll work on a pull request with just the bindings this weekend.
>>
>> To be clear, I haven't reviewed the actual binding parts of it at all. ?I
>> just started and saw a hundred thousand lines
>> of code and stopped.