[cfe-commits] r169219 - in /cfe/trunk: lib/CodeGen/CGDebugInfo.cpp test/CodeGenCXX/debug-info-flex-member.cpp test/CodeGenCXX/debug-info-zero-length-arrays.cpp

[Bill Wendling]

On Dec 4, 2012, at 11:39 AM, Eric Christopher <echristo at gmail.com> wrote:

> On Tue, Dec 4, 2012 at 11:33 AM, Bill Wendling <isanbard at gmail.com> wrote:
> On Dec 4, 2012, at 11:11 AM, Eric Christopher <echristo at gmail.com> wrote:
> 
> > On Tue, Dec 4, 2012 at 10:34 AM, Chris Lattner <clattner at apple.com> wrote:
> >
> > On Dec 4, 2012, at 9:36 AM, Eric Christopher <echristo at gmail.com> wrote:
> >
> >> Doesn't really matter.
> >
> > Small savings of debug info size?
> >
> >
> > They're the same size?
> >
> > Wait, do you mean in the IR? Then yeah, we could probably just replace the upper bound with a count and go from there I think.
> >
> Yeah. In fact, it'll be a bit less of a hack than making the lower bound > upper bound for unbounded subranges. I'll make the change.
> 
> 
> Thank you.
>  
> Eventually, we'll want to support the upper bound being a DIE, so that it can refer to the variable it's defined with -- 'n' in this case:
> 
>         void foo(int n) {
>                 int bork[n];
>                 /* */
>         }
> 
> etc.
> 
> Mmm... you mean expression yes?
> 
Er yeah. :)

-bw