[PATCH] D25620: DebugInfo: introduce DIAlignment type

[Adrian Prantl via llvm-commits]

> On Oct 14, 2016, at 10:13 AM, David Blaikie <dblaikie at gmail.com> wrote:
> 
> 
> 
> On Fri, Oct 14, 2016 at 9:59 AM Adrian Prantl <aprantl at apple.com <mailto:aprantl at apple.com>> wrote:
> 
> 
> > On Oct 14, 2016, at 9:48 AM, David Blaikie <dblaikie at gmail.com <mailto:dblaikie at gmail.com>> wrote:
> 
> >
> 
> > Could someone point me to where the discussion for adding this type came out of? I didn't spot it at a cursory glance of the previous/existing threads.
> 
> >
> 
> 
> 
> This came out of the review thread in:
> 
> 
> 
> D25073: [DebugInfo]: preparation to implement DW_AT_alignment
> 
> https://reviews.llvm.org/D25073 <https://reviews.llvm.org/D25073>
> 
> 
> 
> where I argued that we should not be using a full uint64_t for the alignment fields in the DI.* metadata nodes. I don't think the concrete solution of introducing a new alignment type has been discussed here before.
> 
> Any particular benefits of introducing a new type? We don't I think have any wrapper types for otherwise singular numeric values, do we? (except maybe DIFlags?)

I don't have a strong opinion about adding a new type. If we also had a separate type for sizes, we could potentially make the interface more typesafe to avoid accidentally confusing size and alignment, but that's about it.

-- adrian

>  
> 
> 
> 
> -- adrian
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20161014/d4880d4d/attachment.html>