GDB - lets make D support awesome

[Martin Nowak]

On 11/18/2013 11:41 AM, Iain Buclaw wrote:
> Hi,
>
> I've created a gdb fork on github and will be giving d-lang support some
> love - with the intention of pushing patches back to gdb development as
> soon as possible.  The bulk of the work will be done over the next
> fortnight, but for some other work (see buglink) will take a couple more
> months to roll out as changes on both gdb and compilers need to be
> aligned up.
>
> https://d.puremagic.com/issues/show_bug.cgi?id=9800
>
Great, I also want to look into this in the next few days.
Mostly improving what dmd emits.

> Note #numero due:
> There are a few bug reports here and there, and a few ideas I've got
> scribbled down around the place.  A new DIP should perhaps be created to
> consolidate all these thoughts together.
>
Just file them in bugzilla. IMHO DIPs are for language enhancements.

> Note #numero tre:
> Any changes to the DWARF specification will need to be raised as a new
> issue with details of change and relevant sections.  Lets go through the
> proper channels this time round rather than inventing tags/attributes on
> the fly.
>
I didn't saw much need to extend any DWARF specs.
It's all there, even arrays et.al. in DWARF4.

The current D support in gdb treats every struct with a ptr and length 
field as array.
I also prefixed the array types (_Array_char) so they can be recognized 
better by gdb, pretty-printing is missing though and I don't know 
whether the element types are correctly tagged.