[dragonegg PATCH] Fix debug info for std::nullptr_t

Eric Christopher echristo at gmail.com
Thu Jun 27 00:22:41 PDT 2013


Agreed.
On Jun 27, 2013 12:06 AM, "Peter Collingbourne" <peter at pcc.me.uk> wrote:

> On Thu, Jun 27, 2013 at 08:41:12AM +0200, Duncan Sands wrote:
> > Hi Peter,
> >
> > On 26/06/13 22:34, Peter Collingbourne wrote:
> >> On Wed, Jun 26, 2013 at 07:29:36PM +0200, Duncan Sands wrote:
> >>> Hi Peter,
> >>>
> >>> On 26/06/13 01:59, Peter Collingbourne wrote:
> >>>> On Tue, Jun 25, 2013 at 04:57:38PM -0700, Peter Collingbourne wrote:
> >>>>> Hi,
> >>>>>
> >>>>> This fixes debug info emission for std::nullptr_t (previously it
> would crash).
> >>>>>
> >>>>> OK to commit?
> >>>>
> >>>> With patch this time.
> >>>
> >>> thanks.  How about getting the name from the type rather than assuming
> it is
> >>> nullptr_t?
> >>
> >> I don't think it's possible for the debug info for nullptr_t to
> >> be of any form other than a DW_TAG_unspecified_type with a name of
> >> "nullptr_t" -- at least there doesn't seem to be a way to get Clang
> >> to spit out something else.  (For that matter, it isn't clear to me
> >> why DIBuilder::createNullPtrType lets you specify the name.)  Eric?
> >
> > if other languages start to produce nullptr GCC nodes (and why not -
> maybe this
> > concept is useful for them too) I doubt that they will call it
> nullptr_t, so it
> > seems best to me to just take the GCC name.
>
> Sounds reasonable.  This patch follows GCC's logic when deciding the
> name of a nullptr type.  As a bonus, it also adds debug support for
> LANG_TYPE nodes.
>
> Thanks,
> --
> Peter
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130627/6fdf3ab9/attachment.html>


More information about the llvm-commits mailing list